INTERPRETING REPORTS........................................................................................................

2
Pageviews................................................................................................................................................ 2
Time Metrics.......................................................................................................................................... 3
Traffic Sources....................................................................................................................................... 3
Campaign Attribution.............................................................................................................................................4
Content................................................................................................................................................... 5
FUNDAMENTALS........................................................................................................................6
Account Administration........................................................................................................................ 6
Campaign Tracking & Adords !ntegration.....................................................................................6
Data Discrepancies.................................................................................................................................................8
Manual Campaign URL Tagging...........................................................................................................................9
Ana"#sis $ocus%Ad ords...................................................................................................................&&
Reports ................................................................................................................................................................12
'oa"s..................................................................................................................................................... &(
URL Destination Goal..........................................................................................................................................1
Time on !ite.........................................................................................................................................................1"
#ages #er $isit.....................................................................................................................................................1"
Re%erse Goal #at& Report....................................................................................................................................18
$i"ters.................................................................................................................................................... &)
Using 'ilters ( #ro)iles.......................................................................................................................................2*
*egu"ar +,pression.............................................................................................................................. 22
Cookies................................................................................................................................................. 2-
+commerce Tracking.......................................................................................................................... 32
!ecure #ages.........................................................................................................................................................+4
.omains & Su/domains...................................................................................................................... 3(
Trac,ing Across Multiple !ub-omains................................................................................................................+4
Trac,ing Across Multiple !ub-omains................................................................................................................+
Trac,ing Across Multiple Domains .it& Multiple !ub-omains.........................................................................+/
IN-DEPTH ANALYSIS................................................................................................................37
Advanced Segmentation...................................................................................................................... 3-
Ana"#tics !nte""igence...........................................................................................................................30
!nterna" Site Searc1............................................................................................................................. 30
+vent Tracking and 2irtua" Pageviews..............................................................................................(2
$irtual #age%ie.s................................................................................................................................................4+
1
0%ent Trac,ing.....................................................................................................................................................4+
Additiona" Customi3ations.................................................................................................................. (6
1o. To C&ange !ession Timeout $alue.............................................................................................................4"
1o. to C&ange Campaign 02piration.................................................................................................................4"
1o. to C&ange Campaign #rece-ence................................................................................................................4"
1o. to A-- !earc& 0ngines3...............................................................................................................................4/
1o. To Treat Certain 4e5.or-s As Direct.........................................................................................................4/
1o. to Treat Certain Re)erring !ites As Direct..................................................................................................4/
*These notes are complements of the Goole Analytics IQ Lessons. You can find these notes
and slides via this link: http:!!!."oo"le.comintlenanalyticsi#.html$
INTERPRETING REPORTS
Pageviews
A pageview is counted every time a pa"e on your !e%site loads.
& 'o( for e)ample( if someone comes to your site and vie!s pa"e A( then pa"e *( then
+a"e A a"ain( and then leaves your site ,, the total pa"evie!s for the visit is -.
& If a visitor hits reload after reachin" the pa"e( this !ill %e counted as an additional
pa"evie!.
& If a user navi"ates to a different pa"e and then returns to the ori"inal pa"e( an additional
pa"evie! !ill also %e recorded.
& .ound in the Audience and /ontent 'ection
A unique pageview represents the num%er of visits durin" !hich that pa"e !as vie!ed,,
!hether one or more times.
& In other !ords( if a visitor vie!s pa"e A three times durin" one visit( Goo"le Analytics
!ill count this as three pa"evie!s and one uni#ue pa"evie!.
& .ound in the content section
&
A visit ,, or session ,, is a period of interaction %et!een a !e% %ro!ser and a !e%site.
& /losin" the %ro!ser or stayin" inactive for more than -0 minutes ends the visit.
& .or e)ample( a visitor is %ro!sin" the Goo"le 'tore( a site that uses Goo"le Analytics.
1e "ets to the second pa"e( and then "ets a phone call. 1e talks on the phone for -1
minutes( durin" !hich he does not click any!here else on the site. That2s - pa"evie!s
and 3 visits.
& The 45e! vs. 6eturnin"7 report classifies each visit as comin" from either a ne! visitor
or a returnin" visitor. 'o !hen someone visits your site for the first time( the visit is
cate"ori8ed as 49isit from a ne! visitor.7 If the person has %ro!sed your !e%site %efore(
the visit is cate"ori8ed as 49isit from a returnin" visitor.7
& .ound in the Audience and Traffic 'ources section
3
A visitor is uni#uely identified %y a Goo"le Analytics visitor cookie !hich assi"ns a random
visitor I: to the user( and com%ines it !ith the timestamp of the visitor2s first visit. The
com%ination of the random visitor I: and the timestamp esta%lish a ;ni#ue I: for that visitor.
& Generally, the Visitors metric will be smaller than the Visits metric which in turn will be
smaller than the Pageviews metric.
& 'o( if visitor A comes to your site < times durin" the selected date ran"e and visitor *
comes to your site =ust once( you !ill have 3 9isitors. 6emem%er( a visitor is uni#uely
identified %y a Goo"le Analytics visitor cookie.
Time Metrics
& .or e)ample( a visitor sa! pa"e A( then pa"e *( and then left the site.
& The Time on +a"e for pa"e A is calculated %y su%tractin" the pa"e A timestamp from
the pa"e * timestamp.
& The Time on +a"e for pa"e * is 0 seconds( %ecause there is no su%se#uent timestamp
that Goo"le Analytics can use to calculate the actual Time on +a"e.
Average Time on Page( %ounces are e)cluded from the calculation. In other !ords( any Time
on +a"e of 0 is e)cluded from the calculation.
Average Time on Site, %ounces remain a part of the calculation.
& To calculate Avera"e Time on 'ite( Goo"le Analytics divides the total time for all visits
%y the num%er of visits.
Flas !ase" Sites
& 'ome sites make e)tensive use of .lash or other interactive technolo"ies.
& >ften( these kinds of sites don2t load ne! pa"es fre#uently and all the user interaction
takes place on a sin"le pa"e.
& As a result( it2s common for sites like this to have hi"h %ounce rates and lo! avera"e
times on site.
& If you have such a site( you may !ish to set up your trackin" so that virtual pa"evie!s or
events are "enerated as the user performs various activities.
9isit :uration cate"ori8es visits accordin" to the amount of time spent on the site durin" the
visit. .ound in the ?n"a"ement section of the 9isitors Ta%
Tra##ic Sources
You can compare your traffic sources a"ainst each other to find out !hich sources send you the
hi"hest #uality traffic.
-
$irect Tra##ic% %ookmarks or typin" ;6L directly into %ro!ser.
Searc Engine tra##ic represents visitors !ho click on a search results link in Goo"le( Yahoo(
or any other search en"ine.
& >r"anic: free search results
& +aid search en"ine traffic is pay per click or cost per click traffic that you purchase from
a search en"ine ,, for e)ample on Goo"le Ad@ords.
Re#erring Sites include any sites that send traffic to you. These could %e %anner ads or links
featured on %lo"s( affiliates( or any site that links to your site.
Lookin" at the hi"hest traffic drivers is a start( %ut it doesn2t tell you !hether the traffic !as
#ualified. In other !ords( did the traffic help you achieve the "oals you2ve set for your site$
One easy indicator of quality is Bounce Rate ,, the percenta"e of visits in !hich the person left
!ithout vie!in" any other pa"es.
In the slide( althou"h %lo""er.com sent the most traffic( it has an AAB %ounce rate. A %ounce
rate this hi"h su""ests that the site isn2t relevant to !hat the visitor is lookin" for
By clicking the comare to site average! icon and selecting a comarison metric, you can see
which sources outerform and undererform the site average.
'o here( for e)ample( if !e select *ounce 6ate as our comparison metric. !e can see that the
t!o most popular sources of traffic underperform the site avera"e.
>ne note a%out %ounce rate( if your site is a %lo"( %ounce rate may not %e relevant. @ith %lo"s(
it2s common for people to look at a sin"le pa"e and then leave.
Lookin" at key!ords is a very useful for understandin" !hat visitors !ere e)pectin" to find on
your site.
"eywords with a high bounce rate tell you where you failed to meet that e#ectation.
You can find out !hich landin" pa"es are responsi%le for the poor performance and send the
key!ord traffic to the most effective landin" pa"e.
Be sure to also check the bounce rates for organic, non$aid keywords. %his information can
offer insights into how to best focus your search engine otimi&ation efforts.
&ampaign Attri'ution
*y default( Goo"le Analytics attri%utes a conversion or sale to the campai"n that most recently
preceded the conversion or sale.
.or e)ample( if a visitor clicks on an Ad@ords ad C/ampai"n 1 in the first sessionD and then
later returns via a referral to purchase somethin" C6eferrer 1 in the second sessionD( the referral
!ill "et credit for the sale.
E
1o!ever( if instead the visitor returns directly( then the Ad@ords ad C/ampai"n 1D !ill still "et
credit for the sale.
To prevent a specific referral or campai"n from overridin" a prior campai"n( simply append
4utmFnooverrideG17 to all referrin" campai"n links as sho!n in the slide. This ensures that the
conversion is al!ays attri%uted to the ori"inal referrer Cor first campai"n the user clicked onD.
Therefore( in the e)ample a%ove( the ori"inal campai"n !ill continue to "et credit for the
conversion.
If a visitor returns via a link !ithout the utmFnooverride( as in the third e)ample( that campai"n
!ill "et credit for the sale since it over!rites all previous referrin" campai"ns.
&ontent
The &ontent $rill"own Report "roups pa"es accordin" to directory. You can click on a
directory to see the pa"es in the directory.
The (an"ing Pages Report lists all of the pa"es throu"h !hich people entered your site.
& You can use this report to monitor the num%er of %ounces and the %ounce rate for each
landin" pa"e.
& Bounce rate is good indicator of landing age relevance and effectiveness.
'ou can lower bounce rates by tailoring each landing age to its associated ads and referral
links.
The Navigation Summar) can help you understand ho! people move throu"h your site.
It sho!s ho! people arrived at a specific pa"e and !here they !ent after!ards. 6eports
pa"evie!s.
& The report is availa%le from the +a"es report.
& Percent Entrances sho!s ho! fre#uently the pa"e !as a landin" pa"e.
& Percent Previous Pages sho!s ho! fre#uently visitors came to the pa"e after vie!in"
another pa"e on the site.
& Percent E*its sho!s ho! fre#uently visits ended on this pa"e.
& Percent Ne*t Pages sho!s ho! fre#uently visitors continued on to another pa"e on the
site.
& The list of pa"es that !ere vie!ed immediately %efore the pa"e or pa"es is sho!n in the
left column( under +revious +a"e +ath.
& The list of pa"es that !ere vie!ed immediately after the pa"e or pa"es is sho!n in the
ri"ht column( under :estination +a"e.
The Entrance Pats Report Cunder Landin" +a"es 6eportD is a po!erful tool for analy8in"
navi"ation paths. It reports vistor data.
<
& .or e)ample( let2s say that you !ant to find out !hether people clicked the +urchase
%utton on your landin" pa"e and actually completed the purchase.
& To find out( "o to the Landin" +a"es report and click ?ntrance +aths.
F+N$AMENTA(S
Account A"ministration
4A"ministrators7 have access to all reports and they can also modify settin"s. 'o(
Administrators can create profiles( filters( and "oals( and they can add users.
;sers only have read access to your reports and they can2t modify analytics settin"s. Also(
4;sers7 can %e restricted to vie!in" only specific profiles.
A pro#ile is a set of rules that defines the data you see for a !e% property. .or e)ample( you
mi"ht have !e% property e)ample.com for !hich you have three profiles.
>ne of the profiles mi"ht sho! all the data for all the traffic that comes to e)ample.com.
Another profile mi"ht use filters to only sho! the data for traffic to a certain su%directory.
'till another profile mi"ht use a different set of filters to sho! only another su%set of data.
To see a list of the profiles that %elon" to a specific !e% property( navi"ate to that !e% property
from the Account Administration screen.
>nce you are on the screen for the !e% property( click the +rofiles ta%. >n the +rofiles ta%(
you2ll see a +rofile selector menu that lists all the profiles.
+rofiles are very fle)i%le ,, they are %asically =ust a set of rules that define !hat data is to %e
included in the reports.
& The other profile mi"ht %e filtered so that it contains only traffic from Ad@ords visitors.
& In addition( you mi"ht !ant to "ive certain users access only to the filtered profile. This
has the effect of only allo!in" these users to see Ad@ords traffic to e)ample.com.
*ut( if you are not an administrator( you2ll only see the Assets ta% under the Adminstrator
settin"s. That2s %ecause you need to %e an admnistrator to add ne! profiles or to edit a profile2s
"oals( users( filters( and settin"s.
1o!ever( you don2t need to %e an administrator to add or edit assets. This includes advanced
se"ments( annotations( and custom alerts.
The +rofile 'ettin"s su%,ta% is !here you ena%le e,commerce and site search reports( set your
preferred time 8one( and other settin"s.
&ampaign Trac,ing - A".or"s Integration
H
Goo"le Analytics allo!s you to track and analy8e all of your marketin" campai"ns ,, includin"
paid search campai"ns( %anner ads( emails and other pro"rams.
There are t!o !ays to track ad campai"ns.

.or Ad@ords campai"ns( you should ena%le key!ord autota""in". This allo!s Goo"le
Analytics to automatically populate your reports !ith detailed Ad@ords campai"n information.
In order to ena%le autota""in"( you2ll need to link your Ad@ords and Goo"le Analytics accounts.
The second !ay to track campai"ns is to manually ta" links. 'o( for e)ample( you could ta" the
links in an email messa"e !ith campai"n,identifyin" information. You may also choose to
manually ta" Ad@ords links if you do not !ish to ena%le autota""in".
The ta"s are campai"n varia%les that you append to your ;6L.
*y linkin" Goo"le Analytics to your Ad@ords account( you can "et advanced reportin" that
measures performance and 6>I for your Ad@ords campai"ns. It also makes it easier for you to
evaluate your Ad @ords performance on a key!ord %y key!ord %asis.
Your cost data ,, the information a%out clicks and key!ord spendin" ,, !ill %e applied once you
link your accounts. If you donIt !ant cost data imported into a particular profile( you can edit the
profile settin"s and de,select the cost data option ,, after youIve completed the linkin" process.
(utotagging your links is imortant because it hels (nalytics differentiate the traffic coming
from Google aid listings and traffic coming from Google organic listings.
If autota""in" is not ena%led( your Analytics reports !ill sho! that the clicks from the sponsored
listin"s and the or"anic listin"s are %oth comin" from the same source: "oo"le or"anic.
*y default( Analytics considers them %oth to %e from Goo"le or"anic search results.
'o( ena%lin" autota""in" allo!s you to see !hich referrals to your site came from your paid
Goo"le campai"ns and !hich ones came from Goo"le or"anic search results.
Autota""in" !orks %y addin" a uni#ue id( or ",c,l,i,d( to the end of your destination ;6Ls.
This uni#ue id allo!s Analytics to track and display click details in your reports.
It is important to note that -rd party redirects and encoded ;6Ls can prevent autota""in" from
!orkin" properly.
You should test these cases %y addin" a uni#ue parameter to the end of your ;6L ,,, for
e)ample you could add $testGtest.
Test to make sure that the parameter is carried throu"h to your destination pa"e and that the
link doesn2t %reak.
5otice that the first #uery parameter is al!ays preceded !ith a #uestion mark. 'u%se#uent
values are separated usin" ampersands.
J
$ata $iscrepancies
You may notice differences %et!een the data in your Goo"le Analytics and Ad@ords reports.
There are several reasons for these differences.
.irst( Ad@ords tracks clicks( !hile Analytics tracks visits.
'econd( some visitors !ho click on your Ad@ords ads may have Kava'cript( cookies( or ima"es
turned off. As a result( Analytics !onIt report these visits( %ut Ad@ords !ill report the click.
You2ll also see differences %et!een Analytics and Ad@ords if the Goo"le Analytics Trackin"
/ode on your landin" pa"e doesn2t e)ecute. In this case( Ad@ords !ill report the click %ut
Analytics !ill not record the visit.
Invalid clicks may also cause reportin" differences %ecause !hile Goo"le Ad@ords
automatically filters invalid clicks from your reports( Goo"le Analytics !ill still report the visits.
?)amples include clicks or impressions "enerated %y a pu%lisher clickin" on his o!n ads( a
pu%lisher encoura"in" clicks on his ads( automated clickin" tools or traffic sources( ro%ots( or
other deceptive soft!are.
.inally( keep in mind that Ad@ords data is uploaded once a day to Analytics so the results for
each may %e temporarily out of sync.
Lake sure that your landin" pa"es contain the Goo"le Analytics Trackin" /ode. If they don2t(
campai"n information !ill not %e passed to Analytics( %ut clicks !ill re"ister in Ad@ords.
Lake sure that you have autota""in" ena%led. >ther!ise( visits !ill %e marked as Goo"le
>r"anic instead of Goo"le /+/. @hile !e stron"ly recommend that you use autota""in"
instead of manual ta""in"( if you do manually ta" your destination ;6Ls( you must make sure
that all of them are ta""ed( other!ise data discrepancies !ill occur.
*e a!are that campai"n data can %e lost if your site uses redirects. As a result( Analytics !on2t
sho! the visits as comin" from Ad@ords( %ut your Ad@ords report !ill still report the clicks.
Auto,redirectin" is the techni#ue of automatically sendin" a site visitor to another pa"e once
she has landed on a pa"e. The other pa"e is often on the same !e%site( %ut it can %e on a
different site alto"ether. In fact( auto,redirectin" is used !hen a !e%site has %een set up for the
sole purpose of rankin" hi"hly in the search en"ines. @hen visitors arrive at a pa"e on the site(
from a listin" in the search en"ine results( they are automatically redirected to the main site.

>n,site auto,redirectin" is common !hen a pa"e( !ithin the site( has %een created specifically
to rank hi"hly in the search en"ines( %ut has %een so hi"hly optimi8ed that it is no "ood for
people to actually see. Auto,redirectin" takes visitors to the IproperI pa"e !ithin the site.

A
*oth of those uses are not !anted %y the search en"ines( and they sometimes penalise a pa"e
or site for doin" it. .ortunately( the en"ines are not a%le to automatically detect most of the auto,
redirectin" methods.

The en"ines are not a"ainst auto,redirectin" !hen it is done for they !hat consider to %e a valid
reason. ?.". !hen you have posted a messa"e in a forum( you often "et a confirmation pa"e
!hich !ill !ait a fe! seconds and then auto,redirect you to the messa"es.
Manual &ampaign +R( Tagging
Goo"le Analytics automatically tracks all of the referrals and search #ueries that send traffic to
your !e%site.
1o!ever( if you are runnin" paid advertisin" campai"ns( you should add ta"s to the destination
;6Ls of your ads. Addin" a ta" allo!s you to attach information a%out the campai"n that !ill
sho! up in your Analytics reports.
Althou"h it2s possi%le to manually ta" your Ad@ords ads( you should ena%le auto,ta""in"
instead.

If you manually ta" your Ad@ords ads( the Ad@ords reports !ill only sho! you information %y
/ampai"n and Mey!ord. If you ena%le auto,ta""in"( you2ll %e a%le to see much more detail. The
Ad@ords reports !ill sho! you results %y ad "roup( matched search #uery( placement domain
and many other Ad@ords attri%utes.
There are five varia%les you can use !hen ta""in" ;6Ls. %o tag a )*+, you add a question
mark to the end of the )*+( follo!ed %y your ta" or #uery strin"( as sho!n in the slide.
The varia%les and values are listed as pairs separated %y an e#uals si"n. ?ach varia%le,value
pair is separated %y an ampersand.
Let2s look at each varia%le.
& utm/source to identify the specific !e%site or pu%lication that is sendin" the traffic.
& utm/me"ium to identify the kind of advertisin" medium ,, for e)ample( cpc for cost per
click( or email for an email ne!sletter. 1o! the messa"e is delivered to the consumer.
& utm/campaign to identify the name of the campai"n ,, for e)ample( this could %e the
product name or it mi"ht %e a slo"an. ;sed to "roup several marketin" mediums
to"ether.
& utm/content specifies the version of the ad. .or e)ample( if you have t!o call,to,action
links !ithin the same email messa"e( you can use utmFcontent to differentiate them so
that you can tell !hich version is most effective.
& utm/term specifies the search term %ou"ht.
You should al!ays use these three varia%les !hen ta""in" a link. You can use them in any
order you !ant.
N
?)ample: http:mysite.com$
utmFsourceGyoursiteOutmFmediumG%annerOutmFcampai"nGsummerpromot
Let2s look at a destination ;6L from an Ad@ords ad.
In the first e)ample( no ta" has %een provided and autota""in" is disa%led. In this case( you
!on2t see this traffic in your Ad@ords reports.
& ?)ample: http:!!!.mysite.com
The second e)ample sho!s ho! to manually ta" an Ad@ords link. This traffic !ill sho! up in
your Ad@ords reports( %ut information !ill %e limited to campai"n and key!ord.
You must specify cpc as your medium and "oo"le as your source in order to see this traffic in
your Ad@ords reports. You should also specify cpc as your medium !hen ta""in" paid search
campai"ns from other search en"ines.
& ?)ample: http:mysite.com$
utmFsourceG"oo"leOutmFmediumGcpcOutmFcampai"nGsummerpromoOutmFtermGmyk
ey!ord
The third e)ample sho!s !hat an Ad@ords autota""ed ;6L mi"ht look like once Ad@ords has
appended the ",c,l,i,d varia%le to the end of the ;6L
& ?)ample: http:mysite.com$"clidG13-)y8
You can select any of these varia%les as a dimension in most reports.
.or e)ample( to see all of the sources in /alifornia from !hich you received traffic( you could "o
to the Lap >verlay report( drill do!n to /alifornia( and select 'ource as a dimension.
You can use the ;6L *uilder in the Goo"le Analytics 1elp /enter to construct your ;6Ls.
You enter in the destination ;6L and the values for each campai"n varia%le. You should al!ays
use source( medium and campai"n name.
The ;6L *uilder can %e found via the link displayed here on the slide( or you can search for
4;6L *uilder7 in the Analytics 1elp /enter.
The ;6L %uilder can only construct one ;6L at a time( so you pro%a%ly !on2t !ant to use it to
construct every ;6L for every campai"n.
If you have a lar"e num%er of ;6Ls to ta"( you can use spreadsheets to automate the process.
Generate a sample ;6L in the ;6L *uilder and create a simple spreadsheet formula.
'preadsheets can make it much easier to "enerate thousands of ta""ed ;6Ls.
@hen someone clicks on a ta""ed link Goo"le Analytics( usin" the urchin.=s Kava'cript( stores
all the values in a cookie on the visitor2s machine.
'tick to these %est practices !hen ta""in" your advertisin" campai"ns.
10
& If you use Ad@ords( %e sure to ena%le auto,ta""in". >ther!ise( you2ll miss out on
important information that can help you optimi8e your Ad@ords campai"ns.
& 'econd( for each campai"n( use the ;6L *uilder to create a template ;6L. Then( copy
and paste from the template to create the rest of the ;6Ls for the campai"n.
& Third( use consistent names and spellin"s for all your campai"n values so that they are
recorded consistently !ithin your Analytics reports
& .inally( use only the campai"n varia%les you need. You should al!ays use source(
medium( and campai"n name( %ut term and content are optional.
:ifference %! campai"ns and referrals is that campai"ns are ta""ed.
9isitor campai"n information is stored in a cookie on the visitor2s machine. This cookie stores
the referral information for the visitor2s session. This cookie tracks or"anic referrals( ta""ed
campai"n links( un,ta""ed referral links and direct visits.
?ach time a visitor visits your site the Goo"le Analytics Trackin" code updates this cookie !ith
the appropriate campai"n information. @hen the cookie is updated GA discards the previous
campai"n information. As a result GA only tracks the current campai"n information( not previous
campai"n information.
& ,irect traffic is always overwritten by referrals, organic and tagged camaigns
& 5e! campai"n( referral or or"anic link that %rin"s a visitor to the site al!ays overrides the
e)istin" campai"n cookie
1ere2s an e)ample. A visitor visits your site from a ne!sletter !ith ta""ed links. They look
around and decide to leave. @hen they leave your site the campai"n trackin" cookie !ill persist
and indicate that they ori"inated from the ne!sletter.
The same visitor decides to come %ack the ne)t day and types your ;6L into the %ro!ser. The
campai"n cookie !ill still indicate that the visitor arrived via your ne!sletter %ecause the second
visit !as a direct visit( and direct traffic does not over!rite e)istin" campai"n information.
Anal)sis Focus0A" .or"s
9isits: Pof visits
Impressions: num%er of times your ads !ere displayed.
/licks: 'ho!s the num%er of clicks for !hich you paid and !hich your ads received.
If there are fe!er visits than clicks(
'ome visitors may have clicked a!ay from the !e%site
Landin" pa"e stopped loadin" %efore the trackin" code !as e)ecuted
Landin" pa"es may %e missin" the GA trackin" code
9isitors may not have Kavascript( ima"es and cookies ena%led in their %ro!sers
11
>ften you !ill have fe!er clicks then visits. The reason is that some visitors clicked on the ad
and then later durin" a different session returned directly to the site throu"h a %ookmark. The
referral information from the ori"inal visit !as retained so some clicks resulted in multiple visits.
Impressions( clicks( costs and /T6 Cclick throu"h rateD all relate to ho! many times your ads
!ere displayed and ho! fre#uently people clicked on them. These metrics can help you
understand ho! visi%le and compellin" your ads are to searchers on these key!ords. .or
e)ample( if you !ant a hi"her click throu"h rate( you mi"ht consider %iddin" for a hi"her position
or re!ritin" your ads so that its more relevant to the searcher. If you are "ettin" all 8eroes in the
cost column( make sure you2ve linked your Ad@ords account and that your ena%led auto
ta""in".
6evenue per click( 6>I( and mar"in can help you assess key!ord profita%ility. .or e)ample(
6>I is useful %c it provides a sin"le metric comparison of ho! much you spent vs. ho! much
you made.
An ROI of 0B means that you earned in revenue the same amount of money that you spent. An
6>I of 100B means that you spent Q)2 and received Q3)2 in revenue. If your 6+/ num%ers are
all 8ero and your 6>I num%ers are all minus 100B its %c you have 8ero revenue so make sure
that you have set "oal values or that you2ve ena%led e,commerce trackin".
*efore you delete or pause ne"ative 6>I key!ords( consider !hether you have enou"h data
yet to make a decision. In particular( !atch out for short date ran"es. It2s "enerally not a "ood
idea to make key!ord chan"es on the %asis of a fe! days !orth of data. /onsider return
customers( those that find the site via an Ad@ords ad and then return later to %uy a"ain. You2ll
miss repeat conversions if you set too short of a date ran"e. Also( it may take days or lon"er for
many visitors to %ecome customers so set a date ran"e that is at least as lon" as your e)pected
sales cycle.
Reports
(d-ords .amaigns *eort
>r"ani8es your data accordin" to campai"n( ad "roup( and key!ord. /lick a campai"n and you
drill do!n to the ad "roups. /lick one of the ad "roups and you drill do!n to the key!ords in the
ad "roup. The same metrics are availa%le at every level of the report %ut you can also or"ani8e
your data accordin" to a variety of Ad@ords "imensions( =ust select the Ad@ords dimensions
you !ant from the drop do!n menu. This is po!erful %c you can drill do!n to a su%set of your
data and then %reak it out accordin" to any Ad@ords dimension. In this case !e have already
drilled do!n to an ad "roup !ithin a campai"n. 5o! %y selectin" match type !e can see ho!
different match types perform for this ad "roup. >ne thin" to %e a!are of if you manually ta" the
destination ;6Ls for your key!ords you !ill only %e a%le to see your data %y campai"n and
13
key!ord %ut as lon" as you have ena%led autota""in" you can see your data any of the
Ad@ords dimensions. Let2s look at another report.
,ay Parts *eort
To find out which times of day your campaigns are most effective( navi"ate to the :ay +arts
report. 5o! "raph visits a"ainst transactions and vie! the data hourly. 1ere transactions
increase relative to the visits at noon( 3+L( E+L( A+L and 11+L so !e mi"ht !ant to increase
our %ids at those times. You can also use the ta%le to drill do!n to a specific hour and then a
specific day for that hour and then look at !hich key!ords drove transactions for that hour and
day. 1ere !e can see that the key!ord QGoo"le 'tore2 drive all the revenue for noon on
Thursdays.
,estination )*+ *eort
The destination ;6L report sho!s each destination ;6L on your site that received Ad@ords
traffic. /lick one of the destination ;6Ls then select ad content for the Ad@ords dimensions
drop do!n. 5o! you can see which ad headlines performed best for this particular
destination URL.
Overview *eort
The overvie! report provides shortcut links to commonly needed analyses. .or e)ample( you
can find out the actual search #ueries that drove visits to your site.
Placements *eort
To find out wic placements wor, 'est #or )ou( you can click the link on the overvie! or
navi"ate to the placement reports and you can drill do!n into your placements to find the
specific ;6Ls that !ork %est.
"eyword Positions *eort
You can use the key!ords position report to find out how ad position affected your keywords
performance. 'elect a key!ord on the left side and then select a metric like 4avera"e time on
site7. The 4'ide 17 position for this key!ord referred visitors !ho spent an av" of almost <
minutes on the site. The 4'ide 37 position referrred visitors !ho spent an avera"e of 3.< minutes
on the site.
%V (ds *eort
.inally you can upload your T9 ad( a video file( to your Ad@ords account and start a campai"n
on nation!ide T9. You specify the time of day and !eek( audience demo"raphic and type of
1-
pro"ram you !ould like to tar"et. Then you can track it usin" the T9 campai"ns report. You can
drill do!n into the specific T9 campai"ns and see the impressions delivered( P of ad airin"s(
cost( and /+L alon"side your metrics like visits( time on site and conversions. .or e)ample(
this screenshot sho!s !e%site visits plotted a"ainst impressions delivered( the num%er of active
T9s tuned to your commercial. Lookin" at your !e% traffic metrics alon"side your T9 campai"n
metrics can help you optimi8e your T9 campai"n.
If youIre ta""in" paid /+/ campai"ns( you should also use utmFterm to specify the key!ord.
And( you can differentiate versions of a link ,, for e)ample( if you have t!o call,to,action links
!ithin the same email messa"e( you can use utmFcontent to differentiate them so that you can
tell !hich version is most effective.
Goals
:efinin" site "oals and trackin" "oal conversions is one of the %est !ays to assess ho! !ell
your site meets its %usiness o%=ectives. You should al!ays try to define at least one "oal for a
!e%site.
'o !hat is a "oal$ In Goo"le Analytics( a "oal represents an activity or a level of interaction !ith
your !e%site that2s important to the success of your %usiness.
'ome e)amples of "oals are an account si"nup( a re#uest for a sales call( or even that the
visitor spent a certain amount of time on the !e%site.
There are four types of "oals in Goo"le Analytics.
A +R( $estination goal is a pa"e that visitors see once they have completed an activity. .or
an account si"n,up( this mi"ht %e the 4Thank You for si"nin" up7 pa"e. .or a purchase( this
mi"ht %e the receipt pa"e. A ;6L :estination "oal tri""ers a conversion !hen a visitor vie!s
the pa"e youIve specified.
A Time on Site "oal is a time threshold that you define. @hen a visitor spends more or less
time on your site than the threshold you specify( a conversion is tri""ered.
A Pages per 1isit "oal allo!s you to define a pa"es vie!ed threshold. @hen a visitor vie!s
more pa"es ,,or fe!er pa"es ,,than the threshold youIve set( a conversion is tri""ered.
An Event goal allo!s you to attach a conversion to an event that you have defined. @e2ll learn
a%out events in a su%se#uent lesson.
You can see total conversions and conversion rates for each of your "oals in your reports.
1E
.or each ;6L :estination "oal that you define( you can also define a funnel. A funnel is the set
of steps( or pa"es( that you e)pect visitors to visit on their !ay to complete the conversion. A
sales checkout process is a "ood e)ample of a funnel. And the pa"e !here the visitor enters
credit card information is an e)ample of one of the funnel steps. 'o( the "oal pa"e si"nals the
end of the activity ,, such as a 4thank you7 or 4confirmation7 pa"e ,, and the funnel steps are the
pa"es that visitors encounter on their !ay to the "oal. :efinin" a funnel is valua%le %ecause it
allo!s you to see !here visitors enter and e)it the conversion process. .or e)ample( if you
notice that many of your visitors never "o further than the 4?nter shippin" information7 pa"e( you
mi"ht focus on redesi"nin" that pa"e so that it2s simpler. Mno!in" !hich steps in the process
lose !ould,%e customers allo!s you to eliminate %ottlenecks and create a more efficient
conversion path.
To set up a "oal( first "o the Account Administration pa"e. /lick the account and !e% property
for !hich you !ant to confi"ure a "oal.
'elect the profile to !hich you !ant to add the "oal.
Then( click the "oals ta% and click the plus,Goal link in one of the Goal sets.
You can create up to E sets of < "oals each.
+R( $estination Goal
To define a ;6L :estination Goal( select ;6L :estination as the "oal type. 5e)t( enter the ;6L
of the "oal pa"e. You don2t have to enter the entire ;6L. You can simply enter the re#uest ;6I
, that2s !hat comes after the domain or hostname.
'o( if the complete ;6L is !!!."oo"lestore.comconfirmation.php( you only need to enter
confirmation.php.
Lake sure that the ;6L you enter corresponds to a pa"e that the visitor !ill only see once they
complete the conversion activity. 'o( pick somethin" like the Thank You pa"e or a confirmation
pa"e for your "oal.
You can also enter a name for the Goal ,, here !e2ve entered 4/ompleted >rder7. This name
!ill appear in your conversion reports.
:efinin" a funnel is optional. To define your funnel steps( you add the ;6Ls of the pa"es
leadin" up to the "oal ;6L. Kust as !ith "oals( you don2t have to enter the entire ;6L of a
funnel step ,, =ust the re#uest ;6I is fine.
+rovide a name for each step in the funnel ,, here !e2ve entered 4'elect "ift card 4 for 'tep 1.
The names you enter !ill appear in your reports.
Matc T)pe
The match type defines ho! Goo"le Analytics identifies a "oal or funnel step. You have three
choices for the Latch Type option.
1<
23ea" Matc4 is the default. It indicates that the ;6L of the pa"e visited must match
!hat you enter for the Goal ;6L( %ut if there is any additional data at the end of their
;6L then the "oal !ill still %e counted. .or e)ample( some !e%sites append a product
I: or a visitor I: or some other parameter to the end of the ;6L. 1ead Latch !ill i"nore
these.
1ere2s another e)ample( illustrated on this slide: If you want every page in a
subdirectory to be counted as a goal, then you could enter the
su%directory as the "oal and select 1ead Latch.
4E*act Matc7 means that the ;6L of the pa"e visited must e)actly match !hat you
enter for the Goal ;6L. In contrast to 1ead Latch( !hich can %e used to match every
pa"e in a su%directory( ?)act Latch can only %e used to match one sin"le pa"e. Also
notice that ?)act Latch does not match the second pa"evie!( 4offer1si"nup.html$
#ueryGhats7 %ecause of the e)tra #uery parameter at the end.
4Regular E*pression Matc7 "ives you the most fle)i%ility. .or e)ample( if you !ant to
count any si"n,up pa"e as a "oal( and si"n,up pa"es can occur in various
su%directories( you can create a re"ular e)pression that !ill match any si"n,up pa"e in
any su%directory. 6e"ular ?)pressions !ill %e covered in a later module.
@hen you use 6e"ular ?)pression Latch( the value you enter as the "oal ;6L
as !ell as each of the funnel steps !ill %e read as a 6e"ular ?)pression.
6emem%er that re"ardless of !hich option you choose( Goo"le Analytics is only matchin"
6e#uest ;6Is. In other !ords( the domain name is i"nored.
/heck 4/ase 'ensitive7 if you !ant the ;6Ls you entered into your "oal and funnel to e)actly
match the capitali8ation of visited ;6Ls.
Time on Site
To define a Time on 'ite "oal( select Time on 'ite as the "oal type. 5e)t( select RGreater thanR
or RLess thanR and enter an amount of time( for e)ample 1< minutes. @eIll discuss "oal value
shortly.
Pages Per 1isit
To define a +a"es per 9isit "oal( select +a"es per 9isit as the "oal type. 5e)t( select RGreater
thanR( R?#ual toR( or RLess thanR and enter a num%er of pa"es.
%hreshold goals are useful for measuring site engagement, whereas )*+ ,estination goals are
best for measuring how frequently a secific activity has been comleted. If your o%=ective is for
visitors to vie! as much content as possi%le( you mi"ht set a +a"es per 9isit "oal. >r( if you
have a customer support site and your o%=ective is for visitors to "et the information they need in
as short a time as possi%le( you mi"ht set a Time on 'ite "oal !ith a RLess thanR condition.
1H
The 4Goal 9alue7 field allo!s you to specify a monetary value for "oal. You should only do this
for non,ecommerce "oals.

*y settin" a "oal value( you make it possi%le for Goo"le Analytics to calculate metrics like
avera"e per,visit,value and 6>I. %hese metrics will hel you measure the monetary value of a
non$ecommerce site.
Kust think a%out ho! much each "oal conversion is !orth to your %usiness. 'o( for e)ample( if
your sales team can close sales on 10B of the people !ho re#uest to %e contacted via your
site( and your avera"e transaction is S<00( you mi"ht assi"n S<0 or 10B of S<00 to your
R/ontact LeR "oal.
A"ain( to avoid inflatin" revenue results( you should only provide values for non,ecommerce
"oals.
There is an important difference %et!een "oal conversions and e,commerce transactions.
( goal conversion can only haen once during a visit, but an e$commerce transaction can
occur multile times during a visit.
Let2s say that you set one of your "oals to %e a +:. do!nload and you define it such that any
+:. do!nload is a valid "oal conversion. And let2s say that the "oal is !orth S<.
In this case( if a visitor comes to your site and do!nloads < +:. files durin" a sin"le session(
you2ll only "et one conversion !orth S<. 1o!ever( if you !ere to track each of these do!nloads
as a S< e,commerce transaction( you !ould see < transactions and S3< in e,commerce
revenue.
You2ll learn ho! to set up ecommerce trackin" and ho! to track +:. do!nloads in later
modules.
If you are usin" a filter that manipulates the 6e#uest ;6I( make sure that your ;6L :estination
"oal is defined so that it reflects the chan"ed 6e#uest ;6I field. .or e)ample( in the slide( !e
have a profile that defines thankyou.html as a ;6L :estination "oal. *ut !e have another
profile !ith a filter that appends the hostname to the 6e#uest ;6I. 'o( for this profile( !e need
to chan"e the "oal definition accordin"ly.
If you define a funnel for a "oal( Goo"le Analytics populates the .unnel 9isuali8ation report(
sho!n here in the slide.
>n the left( you can see ho! visitors enter your funnel. >n the ri"ht( you can see !here they
leave the funnel and !here they "o.
The middle sho!s you ho! visitors pro"ress throu"h the funnel ,, ho! many of them continue
on to each step.
1J
In this e)ample( !e can see that there !ere N(3A- entrances at the top of the funnel and 1AJ
completed orders( at the %ottom of the funnel.
%his reort is very useful for identifying the ages from which visitors abandon your conversion
funnel.
Reverse Goal Pat Report
1ere2s another report in the Goals section. It2s the Reverse Goal Pat report5 You can see this
data even if you haven2t defined a funnel. It lists the navi"ation paths that visitors took to arrive
at a "oal pa"e and sho!s you the num%er of conversions that resulted from each path.
In this e)ample( !e can see that NJ of the conversions resulted from the first navi"ation path
that2s sho!n.
This is a "reat report for identifyin" funnels that you hadn2t considered %efore and it can "ive
you "reat ideas for desi"nin" a more effective site.
.ilters
Goo"le Analytics filters provide you !ith an e)tremely fle)i%le !ay of definin" !hat data is
included in your reports and ho! it appears.
You can use them to customi8e your reports so that data that you deem useful is hi"hli"hted in
interestin" !ays. .ilters can also help you clean up your data so that it is easier to read.
There are t!o types of filters in Goo"le Analytics T predefined filters and custom filters.
.ilters process your ra! traffic data %ased on the filter specifications. The filtered data is then
sent to the respective profile.
>nce data has %een passed throu"h a filter( Goo"le cannot re,process the ra! data.
That2s !hy !e al!ays recommend that you maintain one unfiltered profile so that you al!ays
have access to all of your data.
To set up a "oal( first "o the Account Administration pa"e. /lick your desired account. You can
use the .ilters ta% to create ne! filters( edit their settin"s( and apply them to profiles. To create
a ne! filter you !ill need to complete several fields( includin" the filter name and type. If you
elect to create a custom filter( you !ill need to complete several additional fields.
Goo"le Analytics provides three commonly used predefined filters.
The first filter called 4E*clu"e tra##ic #rom "omains7 e)cludes traffic from the domain that you
specify in the :omainfield. If you apply this filter( Goo"le Analytics !ill apply a reverse lookup
!ith each visitor2s I+ address to determine if the visitor is comin" in from a domain that should
%e filtered out. :omains usually represent the I'+ of your visitor althou"h lar"er companies
"enerally have their I+ addresses mapped to their domain name.
1A
The second filter( 4E*clu"e tra##ic #rom IP a""resses7( removes traffic from addresses entered
into the I+ address field. This filter is "enerally used to e)clude your internal company traffic.
The third filter( 4Inclu"e tra##ic su'"irectories7( causes your profile to only report traffic to a
specified directory on your site. This is typically used on a profile that is created to track one part
of a !e%site.
As a %est practice( !e recommend that you create a filter to e)clude your internal company
traffic from your reports.
To do this you can use the predefined filter 4?)clude traffic from I+ addresses7. You !ill need to
enter your I+ address or ran"e of addresses into the QI+ address7 field.
In addition to the pre,defined filters that Analytics offers( you can also create custom filters.
/ustom filters offer you "reater control over !hat data appears in your profiles.
To create a custom filter( select 4/ustom filter7. Additional fields !ill appear !hen you choose
this option.
?ach custom filter has three main parts.
6 Filter T)pes: There are si) filter types availa%le and each one serves a specific purpose.
@e2ll look at these in a minute.
6 Filter Fiel": There are numerous fields you can use to create your filter. ?)amples of
some commonly used fields are the 46e#uest ;6I7 and 49isitor /ountry7 fields. The
complete list of fields can %e found throu"h the link sho!n here or you can search for
4filter fields7 in the Analytics 1elp /enter.
6 Filter Pattern% This is the te)t strin" that is used to attempt to match pa"evie! data. The
pattern that you provide is applied to the field and( if it matches any part of the field( it
returns a positive result and causes an action to occur. You2ll need to use +>'IU
6e"ular ?)pressions to create the filter pattern. Learn more in the module on 6e"ular
?)pressions.
1ere2s a chart that descri%es the filter types.
& E*clu"e an" Inclu"e Filters are the most common types. They allo! you to se"ment
your data in many different !ays. They2re fre#uently used to filter out or filter in traffic
from a particular state or country.
& (owercase an" +ppercase Filters do not re#uire a filter pattern( only a filter field. It
converts the field into all lo!ercase and uppercase letters. Lo!ercase and ;ppercase
filters are very useful for consolidatin" line items in a report. Let2s say( for e)ample( that
you see multiple entries in your reports for a key!ord or a ;6L( and the only difference
%et!een the multiple entries is that sometimes the ;6L or key!ord appears !ith a
different com%ination of uppercase and lo!ercase letters. You can use the Lo!ercase
and ;ppercase filters to consolidate these multiple entries into a sin"le entry.
1N
& Searc an" Replace Filters replace one piece of data !ith another. They are often
used to replace lon" ;6L strin"s !ith a shorter strin" that is easier to read and identify
in your reports.
& A"vance" Filters remove unnecessary data( replace one field !ith another( or com%ine
elements from multiple filter fields. .or e)ample( a %est practice !hen trackin" multiple
su%domains in a sin"le profile is to append the su%domain name to the pa"e names.
You can do this %y creatin" an advanced filter that appends 1ostname to 6e#uest ;6I.
1ere2s an e)ample of ho! you mi"ht use a 'earch and 6eplace filter. Let2s say that your
!e%site uses cate"ory I:s as an or"ani8ational structure. 'o( in your +a"es report( you2d see a
list of 6e#uest ;6Is that indicate the different pa"es on your site.
The pa"e 4cate"ory.asp$catidG<7 is actually the Goo"le 'tore @eara%les pa"e. You could
make the +a"es report more meanin"ful %y replacin" 4catidG<7 !ith a descriptive !ord( like
4@eara%les7.
1ere2s !hat the 'earch and 6eplace filter mi"ht look like. This particular filter !ould over!rite
the entire 6e#uest ;6I !ith 4@eara%les.7
This is a simplified e)ample to "ive you an idea of ho! you can use filters.
+sing Filters - Pro#iles
>nce you2ve defined a filter( you can apply it to a sin"le profile or across several profiles.
'o( for e)ample( in the slide( the "raphic sho!s a sin"le !e% property !ith t!o profiles.
.ilter 1 has %een applied to %oth profiles.
.ilter 3 has %een applied only to +rofile 3.
*y settin" up multiple profiles and applyin" filters creatively to each of them( you have a "reat
deal of reportin" and analysis fle)i%ility.
You can also use profiles and filters to"ether to create customi8ed data vie!s.
Let2s say that you !ant to have t!o different vie!s of your data ,, one vie! includes only traffic
to a su%domain and the other vie! only includes customers from a specific "eo"raphic re"ion.
To do this( you2d set up +rofile 3 and +rofile - as sho!n here in the chart.
>r( for e)ample( you mi"ht !ant to set up a profile that only includes Goo"le Ad@ords traffic.
@e2ll look at ho! to do this in the ne)t slide. 6emem%er( you al!ays !ant to maintain a profile
that contains all of your data. That2s +rofile 1 in the chart.
To set up a profile that includes only Goo"le Ad@ords traffic( you need to apply the t!o /ustom
Include filters sho!n in the slide.
In filter one( you2ll filter on campai"n source for a pattern of "oo"le.
In filter t!o( you2ll filter on campai"n medium for a pattern of cpc.
You can apply these t!o filters in any order.
30
Let2s look at ho! you can use profiles and filters to track su%domains.
If your su%domains are totally separate %usinesses( and you have no need for reports that
include cumulative traffic to %oth( then you could simply create a uni#ue !e% property for each
su%domain.
Goo"le Analytics creates a uni#ue !e% property I: for each !e% property you set up.
The !e% property I: comprises the letters 4;7 4A7( follo!ed %y the account I:( follo!ed %y
another num%er that distin"uishes the !e% property from other !e% properties in the account.
In the slide e)ample( !e% property 1 is distin"uished %y a dash 1. @e% property 3 is
distin"uished %y a dash 3.
'o( you2d install the 4dash 17 version of your trackin" code on your 'u%domain A pa"es( and the
4dash 37 version of your trackin" code on your 'u%domain * pa"es.
*ut !hat if you !ant to analy8e the traffic a""re"ated across %oth su%domains$ In this case(
you could set up - duplicate profiles under a sin"le !e% property.
Then( you2d apply an Include filter to t!o of the profiles.
+rofile 1 includes all traffic to %oth su%domains.
+rofile 3 only includes traffic to su%domain A.
+rofile - only includes traffic to su%domain *.
In this scenario( you2d install identical trackin" code on every pa"e of the site re"ardless of
su%domain.
@hen settin" up profiles and filters for your Analytics account( you should al!ays create one
unfiltered profile that can %e a %ack,up in case your filters do not function as planned or you
need more data than you ori"inally thou"ht.
6emem%er( once your ra! data has passed throu"h filters( Goo"le cannot "o %ack and
reprocess the data. 'o( maintainin" an unfiltered profile provides you !ith a %ackup.
You can apply multiple include and e)clude filters to a sin"le profile( %ut keep in mind that !hen
more than one filter is applied( the filters will be e#ecuted in the same order that they are listed
in your Profile /ettings.
In other !ords( the output from one filter is then used as the input for the ne)t filter.
The e)ample sho!n here illustrates that if you !ant to include only users from /alifornia and
Te)as( you cannot create t!o separate include filters %ecause they !ill cancel each other out.
The solution is to create one filter that uses a re"ular e)pression to indicate that the 9isitor
6e"ion should %e /alifornia or Te)as.
If you drive traffic from Ad@ords to multiple sites( each of !hich is tracked in a separate
Analytics profile( you2ll need to apply a filter to each site2s profile.
*ecause( !hen you apply cost data from an Ad@ords account( data from the entire account is
applied to each profile , Goo"le Analytics doesn2t automatically match campai"ns to specific
profiles.
31
To illustrate !hat !ould happen if you don2t apply a filter( let2s ima"ine that you have t!o sites
and you spend S<0 to drive traffic to each of them.
@ithout a filter( the /licks ta% on each profile !ould include S100 !orth of cost data instead of
=ust the S<0 you spent for that site.
'o( for each profile that should include a su%set of your Ad@ords data( you2ll need to create a
custom include filter.
/reate a custom filter and select the Include filter type.
.or the filter field( select 4/ampai"n Tar"et ;6L7. This field only applies to Goo"le Ad@ords
data.
;se a re"ular e)pression to create the filter pattern %ased on the Ad@ords destination ;6L that
is applica%le to this profile.
>nce you2ve saved this filter( only Ad@ords data for this profile !ill %e displayed in the reports.
Regular E*pression
A re"ular e)pression is a set of characters and metacharacters that are used to match te)t in a
specified pattern.
You can use re"ular e)pressions to confi"ure fle)i%le "oals and po!erful filters.
.or e)ample( if you !ant to create a filter that filters out a ran"e of I+ addresses( you2ll need to
enter a strin" that descri%es the ran"e of the I+ addresses that you !ant e)cluded from your
traffic.
Let2s start off %y lookin" at each metacharacter.
Letacharacters are characters that have special meanin"s in re"ular e)pressions.
$ot: !ildcard to match any sin"le character.
& ?)ample: 4Act .( 'cene -7 !ould match 4Act 1( 'cene -7 and 4Act 3( 'cene -7
& The operative !ord here is 4sin"le7( as the re"e) !ould 5>T match Act 10( 'cene -.
The dot only allo!s one character( and the num%er ten contains t!o characters ,, a 1
and a 0.
& You could use t!o dots to !rite a re"ular e)pression that !ould match 4Act 10( 'cene
-7.
& To make your re"e) more fle)i%le( and match ?IT1?6 4Act 1( 'cene -7 or 4Act 10(
'cene -7( you could use a #uantifier like the V si"n.
33
!ac,slases allo! you to use special characters( such as the dot( as thou"h they !ere literal
characters. ?nter the %ackslash immediately %efore each metacharacter you !ould like to
escape.
& 4;.'. 1oliday7 !ritten this !ay !ith periods after the ; and the ' !ould match a num%er
of unintended strin"s( includin" ;+'. 1oliday( ;.'% 1oliday( and ;-'" 1oliday.
& 6emem%er that the dot is a special character that matches !ith any sin"le character( so
if you !ant to treat a dot like a re"ular dot( you have to escape it !ith the %ackslash.
& ;W.'W. 1oliday matches only the ;.'. 1oliday.
& You2ll use %ackslashes a lot( %ecause dots are used so fre#uently in precisely the strin"s
you are tryin" to match( like ;6Ls and I+ addresses.
& .or e)ample( if you are creatin" a filter to e)clude an I+ address( remem%er to escape
the dots.
& .or 1N3.1HA.1.1 !ould need to %e !ritten as 1N3W.1HAW.1W.1
;se square 'rac,ets to enclose all of the characters you !ant as match possi%ilities. 'o( in the
slide( you2re tryin" to match the strin" ;.'. 1oliday( re"ardless of !hether the ; and the ' are
capitali8ed.
& Xu;YW.Xs'YW. 1oliday matches u.s. 1oliday and ;.'. 1oliday
& 1o!ever( the e)pression !on2t match ;.'. 1oliday unless periods are used after %oth
the ; and the '. The e)pression also re#uires that the 1 is capitali8ed.
& There is a re"e) you can !rite to match all of these variations. The #uestion mark used
here is another 4#uantifier7( like the QV2 si"n mentioned earlier.
& Xu;YW.$Xs'YW.$Xh1Yoliday !ill match ;' 1oliday or u.s. holiday.
& You can either individually list all the characters you !ant to match( as !e did in the first
e)ample( or you can specify a ran"e.
Z ;se a )pen inside a character set to specify a ran"e. 'o instead of typin"
s#uare %racket 0 1 3 - E < H J A N( you can type s#uare %racket 0,N.
& And( you can ne"ate a match usin" a caret after the openin" s#uare %racket.
Z Typin" s#uare %racket caret 8ero dash nine !ill e)clude all num%ers from
matchin". X[0,NY does not match 0,N. X0,NY matches 0,N.
%he use of the caret shown here is secific to character sets, and the negating behaviour
occurs only when the caret is used after the oening square bracket in a character set.
5o! let2s talk a%out usin" #uantifiers to indicate repetition.
In earlier e)amples( !e2ve used the plus si"n and the #uestion mark.
& The question mar, re#uires either 8ero or one of the precedin" character. In the
e)pression 4-,1,$7 ( the precedin" character is a 1. 'o( %oth - and -,1 !ould match.
There !ill only %e t!o possi%le options.
& The plus sign re#uires at least one of the precedin" character. 'o( 4-,1,V7 !ouldn2t
match =ust a -. It !ould match -,1( -,1,1( -,1,1,1( and so on.
3-
& The asteris, re#uires 8ero or more of the precedin" character. In the e)pression( 4-,1,
*7( the precedin" character is a 1. 'o it !ould match -( -,1,( -,1,1( -,1,1,1 and so forth.
You can also '+?/I.Y repetition usin" a minimum and ma)imum num%er inside curly %rackets.
& -,1,\3] matches -,1,1. It does not match -( -,1( or -,1,1,1.
& -,1,\1(-] matches -,1( -,1,1( -,1,1,1. It does not match - or -,1,1,1,1.
6ecall that a dot matches any sin"le character. @hat !ould you use to match a !ildcard of
indeterminate len"th$
$ot star !ill match a strin" of any si8e. :ot star is an easy !ay to say 4match anythin"(7 and is
commonly used in Goo"le Analytics "oals and filters.
It is handy to use the parentheses and the pipe sym%ol Calso kno!n as the >6 sym%olD
to"ether.
*asically( you can =ust list the strin"s you !ant to match( separatin" each strin" !ith a pipe
sym%ol ,, and enclosin" the !hole list in parentheses.
1ere( !e2ve listed four variations of 4;'7 that !e2ll accept as a match for ;' 1oliday.
& C;W.'W. ^ uW.sW. ^ ;' ^ usD 1oliday matches 4;' 1oliday7 4;.'. 1oliday7 4u.s. 1oliday7 and
4us 1oliday7
& This !ill not match 4;.' 1oliday7 %c it !as not one of our options.
& ;se CuW.$sW.$ ^ ;W.$'W.$D !ill match all of the a%ove.
If it2s not in the list( it !on2t "et matched. That2s !hy 4;' 1oliday7 !on2t "et matched if one of
the periods is missin".
In our list( !e2ve accounted for %oth periods missin"( %ut not for =ust one period missin".
;sin" #uestion marks( the second re"e) in the slide !ill match all of the a%ove.
Ancors
& 7 'tart of a 'trin"
& 8 ?nd of a 'trin"
The caret si"nals the %e"innin" of an e)pression. In order to match( the strin" must *?GI5 !ith
!hat the re"e) specifies..
The dollar si"n says( if there are any more characters after the ?5: of this strin"( then it2s not a
match.
'o( [;' means start !ith ;'. 4;' 1oliday7 matches( %ut 45e)t Londay is a ;' 1oliday7 does
not match.
1olidayS means end !ith 1oliday. ;' 1oliday still matches( %ut 4;' 1oliday 'chedule7 does
not match.
Anchors can %e useful !hen specifyin" an I+ address. Take a look at these e)amples.
3E
& 1N3W.1HAW.1W.1S matches 41N3.1HA.1.17 %ut not 41N3.1HA.1.1<7
& *J3W.1HAW.1W.1 matches 4J3.1HA.1.17 %ut not 41J3.1HA.1.17
'ome character classes are used so commonly that there is a shorthand you can use instead of
!ritin" out the ran"es !ithin s#uare %rackets.
Let2s look at the e)ample of a simplified re"e) that could match an address:
& !ac,slas " 9:"; means match any one di"it 8ero throu"h nine.
& ;se curly %rackets and a minimum and ma)imum num%er to specify ho! many di"its to
match.
& *ackslash d follo!ed %y 1 comma < in curly %rackets means that the address must
contain at least one di"it( and at most five di"its.
& !ac,slas s 9:s; means that the num%er should %e follo!ed %y one space(
& !ac,slas w 9:w<; means match any alphanumeric character and the star means
include as many alphanumeric characters as you !ant Cdoes not include spacesD
?)ample: Wd\1(<]WsW!*
4-E< ?m%arcadero7 matches( %ut =ust 4?m%arcadero7 does not( %ecause this re"e) re#uires the
strin" to start !ith a num%er.
If you !ant to make the num%er optional( "roup the first part of the re"e) !ith parentheses,,
includin" the space,,and follo! it !ith the #uestion mark.
?)ample: CWd\1(<]WsD$W!*
5ote that an address like 41H00 Amphitheatre +ark!ay7 !ould not match either( %ecause the
re"e) does not account for the space %et!een Amphitheatre and +ark!ay.
The slide sho!s one !ay you could account for this.
?)ample: Wd\1(<]WsW!*WsW!*
6e"e) 6evie!
Let2s revie!.
X Y: Latches /haracters
^ : ?ither or
C D: Groups 'tatements
In the e)ample on the slide( !e2ve created an e)pression that !ill match the strin"s Goo"le or
Yahoo( re"ardless of !hether or not Goo"le and Yahoo are capitali8ed.
?)ample: CX"GYoo"le ^ XyYYahooD
1ere( !e2ve created an e)pression that !ill match ;6Ls for internet and theatrical movie
trailers. The first part of the e)pression indicates that the ;6L can %e"in !ith anythin". Then the
e)pression specifies that the ;6L must end !ith inde).php$dlGvideotrailers and then either
internet or theatrical. The S si"n ensures that any ;6Ls that are any lon"er than this !on2t "et
included in the match.
3<
?)ample: .*inde)W.phpW$dlGvideotrailersCinternet^theatricalDS
X Y: Latches /haracters
^ : ?ither or
C D: Groups 'tatements
. : Latches Any 'in"le /haracter
* : Latches _ero or Lore /haracters
W :?scape
S: ?nd of a 'trin"
You2ll find lots of applications for re"ular e)pressions in Goo"le Analytics.
'ome common e)amples are:
` filterin" out internal traffic %y specifyin" a set of I+ addresses. The follo!in" !ould %e entere
,30<W.1J3W.3-X3-E<Y
` settin" up a "oal that needs to match multiple ;6Ls. The follo!in" !ould %e entered in the
Goal ;6L .ield !hen creatin" a ne! "oal
.*inde)W.phpW$dlGvideotrailersCinternet^theatricalDS
` trackin" e#uivalent pa"es in a funnel
do!nloadscasestudy.*
` and usin" the filter %o) that appears on your reports to find specific entries in a ta%le. This
!ould %e entered under the .ilter +attern field.
CX"GYoo"le ^ XyYYahooD
1ere2s ho! you mi"ht use re"ular e)pressions to "roup pa"es or funnel steps on your site.
;sin" a re"ular e)pression allo!s you to track them as one funnel step rather than trackin"
each pa"e or action individually. If your site has multiple pa"es or steps that e#uivalent in value
Ce.". !hitepaper or case study do!nloadsD you can use 6e"?) to "roup them. 6e"?) allo! you
to create one funnel as opposed to trackin" each pa"e or individually.
Learn ho! "oals and funnels !ork in the module on "oals.
And( here2s an e)ample of usin" re"ular e)pressions !ithin your reports. @e2re usin" the
'earch filter to display all the ro!s in the ta%le that contain Goo"le or Yahoo.
Goo"le Analytics provides a tool that makes it easier to "enerate a re"ular e)pression that
matches a ran"e of I+ addresses.
It2s called the 6e"ular ?)pression Generator and you can find it at the ;6L sho!n in the slide.
>r( you can search for 6e"ular ?)pression Generator in the Goo"le Analytics 1elp /enter.
You2ll find a num%er of useful applications for re"e) as you use Goo"le Analytics.
*ut( it2s important that you think throu"h all the implications of each e)pression that you use
!hen you set up a filter or a "oal.
3H
It2s easy to make a mistake and not "et the data or the result you2re lookin" for.
'et up a duplicate profile to test your re"e) statements. After enou"h data has %een collected(
check your results and make sure they2re !hat you e)pect.
6emem%er to al!ays maintain a %ackup profile that includes all your data.
There are lots of re"e) resources on the !e%. To "et started( =ust search for re"e).
;sin" filters to includee)clude traffic to specific sections or pa"es !ill not mimic your advanced
filters. The reason for the differences is that your filters are pa"evie! %ased and the advanced
se"ments are visit %ased. .or e)ample( !hen you e)clude traffic to a specific directory you are
removin" any pa"evie!s of =ust those pa"es. If a visitor that visited the filtered directory also
visited other pa"es on your site durn" their visit( then their visit !ill still %e counted as !ell as
the pa"evie!s of those other sections. 1o!ever( !hen you use advanced filters to e)clude
pa"es( you are not =ust e)cludin" those pa"evie!s( %ut you are e)cludin" ALL the pa"evie!s of
every pa"e for visits that sa! your e)cluded pa"es durin" their visit. This is the reason that the
visit and pa"evie! num%er usin" advanced se"ments is lo!er than your profilefilter method.
&oo,ies
'ome !e% sites store information a%out you or your computer in a small te)t file called a cookie.
The cookie is stored on your hard drive.
'ites that run Goo"le Analytics issue first party cookies that allo! the site to uni#uely( %ut
anonymously( identify individual visitors.
'o( !hen a visitor returns to a site that runs Goo"le Analytics( the site is a%le to remem%er that
the visitor has %een to the site %efore and Goo"le Analytics !ill only count that visitor once in
uni#ue visitor calculations.
There are t!o types of cookies. First0part) coo,ies are set %y the domain %ein" visited. >nly
the !e% site that created a first,party cookie can read it. This is the kind of cookie used for
Goo"le Analytics trackin".
Tir"0part) coo,ies are set %y third party sites ,, %asically sites other than the site %ein"
visited.
;sers can choose !hether to allo! some( none( or all types of cookies to %e set on their
computers.
1o!ever( if a user does not allo! cookies at all( they may not %e a%le to vie! some @e% sites
or take advanta"e of customi8ation features.
/ookies can %e set !ith or !ithout an e)piration date. This detail is important in order to
understand ho! Goo"le Analytics tracks visits and uni#ue visitors.
3J
+ersistent cookies have an e)piration date( and remain on your computer even !hen you close
your %ro!ser or shut do!n. >n return visits( persistent cookies can %e read %y the !e% site that
created them.
Temporary cookies do not have an e)piration date( as they are only stored for the duration of
your current %ro!ser session. As soon as you #uit your %ro!ser( temporary cookies are
destroyed.
@hile it2s impossi%le to determine the e)act num%er of !e% visitors !ho have cookies ena%led
or disa%led( availa%le statistics su""est that the vast ma=ority of visitors ena%le cookies.
Lany kinds of sites re#uire that visitors have cookies ena%led.
.or e)ample( you need to have cookies ena%led in order to lo"in to many online shoppin" carts
and to use !e% mail.
.irst party cookies( !hich are the kind used for Goo"le Analytics( are allo!ed %y a ma=ority of
visitors.
/ookie trackin" makes it possi%le to correlate shoppin" cart transactions !ith search campai"n
information( and perform other visitor analysis.
6emem%er ,, !e%sites only have access to the information that you provide. @e%sites can2t "et
your email address or access to any information on your computer unless you provide it. And
since Goo"le Analytics only uses first party cookies( Goo"le Analytics cookies can only %e read
%y the !e%site that created them.
Goo"le Analytics sets the five first,party cookies sho!n in the slide.
1. Futma: 9isitor Identifier C?)pires after t!o yearsD
3. Futm%: 'ession Identifier C?)pires after -0 minutes of inactivityD
-. Futmc: 'ession Identifier CTemporary cookiea destroyed !hen you #uit the %ro!serD
E. Futm8: /ampai"n 9alues C?)pires after si) monthsD
<. Futmv: 9isitor 'e"mentation C?)pires after t!o yearsD
The FFutmv cookie is optional( and !ill only %e set if the Fset9arCD method is called. You !ill
learn a%out Fset9arCD in the module on /ustom 9isitor 'e"mentation.
All of the Goo"le Analytics cookies are persistent e)cept for one. The FFutmc cookie is a
temporary cookie that is destroyed !hen the visitor #uits the %ro!ser.
?ach of the other Goo"le Analytics cookies has an e)piration date set in the future( meanin"
that the cookie !ill persist on the user2s computer until it e)pires( or until the user deletes it from
their computer.
1ere2s an e)ample of the cookies set %y the Goo"le 'tore. You can see that FFutma( FFutm%(
FFutmc( and FFutm8 have %een set. @e2ll learn more a%out each cookie shortly.
3A
.irst( let2s try a %rief e)periment. @hich of the sites that you2ve visited are usin" Goo"le
Analytics$
To find out( open your %ro!ser2s cookie !indo!. You2ll usually find it under your %ro!ser2s
4>ptions7 or 4+references7.
5o!( in the cookies !indo!( search for underscore underscore u,t,m. You should see all the
different Goo"le Analytics cookies set %y all the sites that you2ve visited that use Goo"le
Analytics.
All cookies are %ro!ser,specific. 'o( if you2ve already %een to a site( %ut you open a different
%ro!ser to visit that site a"ain( another set of Goo"le Analytics cookies !ill %e set.
5o!( %efore !e continue( search for the Goo"le 'tore cookies %y typin" the domain name
4"oo"lestore.com7 into the /ookies search %o).
If you2ve never visited the Goo"le 'tore( "o to "oo"lestore.com no! so that cookies are
created.
'elect the Goo"le 'tore FFutma cookie. In the cookie information( note the 4/ontent7 and
e)piration date for the cookie.
Futma , 9isitor Identifier
utmaG1J11HNEE3 3<<0-H3-A 111<1JHHJ< 111<1JHHJ< 11<1JHHJ< 1
The first num%er in the content of every Goo"le Analytics cookie is called the 4"omain as.7 It
represents the domain that you visited and that set these cookies. Goo"le Analytics applies an
al"orithm to the domain and outputs a uni#ue numeric code that represents the domain. ?ach
Goo"le Analytics cookie set %y the domain !ill %e"in !ith this num%er.
The ne)t num%er is a ran"om unique I$.
The three su%se#uent num%ers are timestamps. They represent the time of the initial visit( the
%e"innin" of your previous session( and the %e"innin" of your current session. %he timestams
reresent the number of seconds since 0anuary 1, 1234.
5otice that the last three timestamps are the same. @hat does this tell you$
The last num%er( te session counter, can "ive you the ans!er. The last num%er tells you the
num%er of times you have visited this site. This num%er !ill increment each time you visit the
site. The session counter here is 417( and the last three timestamps are all the same %ecause
this is your first visit to the site.
%he random unique 5, combined with the first timestam make u the visitor 5, that Google
(nalytics uses to identify unique visitors to the site. These details allo! Goo"le Analytics to
calculate the num%er of uni#ue visitors and num%er of visits.
3N
Look at your Goo"le 'tore FFutma cookie.
1o! many times have you visited the Goo"le 'tore$ If you think you2ve visited more times than
is indicated %y the cookie( remem%er that the cookie only includes the num%er of times you
visited from this computer usin" this %ro!ser.
Also( if you have cleared your cookies at some point( it is only countin" from the last time you
cleared your cookies.
@hen does this cookie e)pire$ You should see that the date is t!o years from last the time you
visited.
The FFutm% and FFutmc cookies to"ether identify a session.
The content of the FFutmc cookie is simply the domain hash.
The content of the FFutm% cookie !ill also %e the domain hash plus( if the site is usin" "a.=s(
some additional values.
The key difference %et!een the t!o cookies is that FFutm% is a persistent cookie !ith an
e)piration date that is set -0 minutes after it is created. @hile FFutmc is a temporary cookie that
is destroyed as soon as the visitor #uits the %ro!ser.
Let2s revie! !hat you kno! a%out a session( or visit( in Goo"le Analytics. .irst note that the
terms 4session7 and 4visit7 are used interchan"ea%ly. A session is defined %y -0 minutes of
inactivity or if a visitor #uits the %ro!ser.
?ach time the visitor navi"ates to a ne! pa"e and the Kava'cript in the Goo"le Analytics
Trackin" /ode is e)ecuted( the FFutm% cookie is refreshed and set to e)pire in -0 minutes.
This is ho! a session can %e 3 hours lon". As lon" as the visitor remains active on the site( the
session remains active.
*ut if the visitor stays on a pa"e for more than -0 minutes( the FFutm% cookie !ill %e destroyed.
The ne)t time the visitor loads a pa"e( Goo"le Analytics !on2t find aFFutm% cookie. Instead( a
ne! FFutm% cookie is created and( from the standpoint of trackin"( this is a ne! session.
'o( !hy is the FFutmc cookie needed$ Let2s say a visitor #uits and starts the %ro!ser and
comes %ack ri"ht a!ay to the same site. 'ince the FFutmc cookie !as destroyed( Goo"le
Analytics !ill kno! that this is a ne! session.
'o( to summari8e( !hen the visitor loads a pa"e( the Kava'cript in the Goo"le Analytics
Trackin" /ode checks for %oth the FFutm% and FFutmc cookies. If either one is missin"( it notes
this as a ne! session( and creates !hichever cookie,, FFutm%( FFutmc( or %oth,, !as missin".
5ote that it is possi%le to ad=ust this %ehavior. @ith a small customi8ation to the Goo"le
Analytics Trackin" code( you can make the session timeout len"th anythin" you !ant. You2ll
learn a%out this in the /ode /ustomi8ations module.
-0
The //utm= coo,ie stores the campai"n trackin" values that are passed via ta""ed campai"n
;6Ls.
utm8G1J11HNEE3 110AA<AJ1H.3.-.utmcsrG"oo"le^utmccnGCor"anicD^utmcmdGor"anic^
utmctrGtoys
3Gsession num%er
-Gcampai"n num%er
'o( for e)ample( if a visitor comes to your site on a link ta""ed !ith campai"n varia%les
utmFsource( utmFmedium( and utmFcampai"n( the values for these varia%les !ill %e stored in
the FFutm8 cookie.
+recedin" the campai"n trackin" values( you !ill see four num%ers stored in the FFutm8 cookie.
The first num%er is the domain hash( as !ith the other Goo"le Analytics cookies.
The second num%er is a timestamp.
The third and fourth num%ers are the 4session num%er7 and 4campai"n num%er7( respectively.
The 4session num'er7 increments for every session durin" !hich the campai"n cookie "ets
over!ritten.
The 4campaign num'er7 increments every time you arrive at the site via a different campai"n
or or"anic search( even if it is !ithin the same session.
The FFutm8 cookie has a si) month timeout( meanin" that a visit !ill %e attri%uted to a particular
campai"n for up to si) months( or until the FFutm8 cookie is over!ritten !ith another value.
You can modify the si) month timeout and you can chan"e the rules !hich "overn !hen the
FFutm8 cookie value is over!ritten. You2ll learn ho! in the /ode /ustomi8ations module.
The FFutm8 data sho!n here !ould sho! up in your All Traffic 'ources report as comin" from
the source medium 4"oo"le or"anic7.
5o!( in your %ro!ser2s cookie !indo!( select the FFutm8 cookie from your visit to
"oo"lestore.com. Assumin" that it !as a direct visit
The slide sho!s ho! the values in the FFutm8 cookie map to campai"n varia%les.
.or e)ample( the utmcsr value in the FFutm8 cookie is the source( or the value that !as
assi"ned to utmFsource in the ta""ed link.
/utm=0&ampaign 1aria'les
& utmcsr inFutm8 is the 'ource CutmFsourceD
& utmccn inFutm8 is the /ampai"n CutmFcampai"nD
& utmcmd in Futm8 is the Ledium CutmFmediumD
& utmctr in Futm8 is the Mey!ord CutmFtermD
& utmcct in Futm8 is the Ad /ontent CutmFcontentD
-1
'o( if you reached 4somesite.com7 via a ta""ed ;6L that looks like this( then the FFutm8 cookie
!ould look like this.
;6L: http : !!! .somesite .com $
utm F source G ne!sletter O utm F campai"n G urchin <O utm F medium G cpc O utm F term G !e% V analytics O
utm F content G %anner F ad
/ookie:
utm8G1J11HNEE3.110AA<AJ1H.-.3.utmscrGne!sletter^utmccnGurchin<^utmcmdGcpc^
utmctrG!e%Vanalytics^utmcctG%annerFad
The F/utmv coo,ie is for custom visitor se"mentation. You2ll only see this cookie if the site
calls the Fset9arCD method. This cookie contains the domain hash( and one other value: the
value you assi"n usin" Fset9arCD.
.or e)ample( suppose all site visitors !ho lo" in "et set to 4Lem%er7( !hile those !ho do not lo"
in remain unassi"ned. The Goo"le Analytics account o!ner !ould then %e a%le to compare
4Lem%ers7 to those !ho are 4Cnot setD7 and see !hether( for e)ample( Lem%ers convert more
often or spend more money on the site.
The FFutmv is a persistent cookie that e)pires after 3 years.
Try searchin" your %ro!ser cookies for 4utmv7. Any sites that appear !ill %e those that use the
Goo"le Analytics custom se"mentation feature.
6efer to the module on /ustom 9isitor 'e"mentation to learn more a%out Fset9arCD and the
FFutmv cookie.
Ecommerce Trac,ing
If your site sells products or services online( you can use Goo"le Analytics e,commerce
reportin" to track sales activity and performance. The ?commerce reports sho! you your site2s
transactions( revenue( and many other commerce,related metrics.
'ome e)amples of the kind of information you can "et from the e,commerce reports include:
` the products that !ere purchased from your online store
` a list of transactions
` the num%er of times people visited your site %efore purchasin"
6$commerce metrics are also available on the 6commerce tab which aears in many reorts.
.or e)ample( on the ?commerce ta% of the Ad@ords /ampai"ns report( you can see ho! much
revenue is associated !ith your Ad@ords campai"ns. You can also see revenue associated
!ith specific referrals( the num%er of transactions associated !ith or"anic search and avera"e
per visit value across all traffic sources.
In order to use e,commerce reportin"( you2ll need to do three thin"s.
-3
1. ?na%le e,commerce reportin" !ithin your Analytics !e%site profile.
/lick the Account Administration icon. 5avi"ate to the desired account and !e%
property.
'elect the desired profile and click the +rofile 'ettin"s ta%.
'elect 4Yes7 ne)t to ?,commerce @e%site and save your chan"es.
3. Add or make sure that you2ve added the Goo"le Analytics Trackin" /ode to your receipt
pa"e or 4Transaction /omplete7 pa"e.
-. .inally( you2ll need to add some additional e,commerce trackin" code to your receipt pa"e
so that you can capture the details of each transaction.
FaddTransCD /all
The call to FaddTransCD tells Goo"le Analytics that a transaction has occurred.
The ar"uments to FaddTransCD provide details a%out the transaction
FaddTransCD takes the follo!in" ar"uments
Z >rder I:*( Affiliation( Total*( Ta)( /ity( 'tate( /ountry C*Gre#uired fieldsD
You can type sin"le,#uote sin"le,#uote to leave an optional field %lank( %ut note that
>rder I: and Total are re#uired.
Your code !ill need to dynamically retrieve the values from your merchant soft!are to
populate these fields.
FaddItemCD /all
After the call to FaddTransCD( there must %e at least one call to the FaddItemCD method.
This call provides Goo"le Analytics !ith details a%out the specific item purchased.
.or each item that the visitor purchases( call addItemCD
FadditemCD takes the follo!in" ar"uments:
Z >rder I:*( 'M;/ode*( +roduct 5ame( /ate"ory( +rice*( Quantity*
.or each item that a visitor purchases( call FaddItemCD. If more than one item is
purchased( you2ll call FaddItemCD multiple times.
As !ith FaddTransCD( you can leave some of the fields %lank( %ut note that >rder I:(
'M; or /ode( +rice and Quantity are re#uired ar"uments.
;se the same >rder I: that you used in the call to addTransCD.
If you2re not sure ho! to !rite this code( contact your merchant soft!are provider.
FtrackTransCD /all
.inally( there is a call to the trackTransCD method !hich sends all the data to Goo"le
Analytics.
'ends %oth the transaction data and the item data to the GA server
/all after the Goo"le Analytics Trackin" /ode calls Ftrack+a"e9ie!CD and after you2ve
called FaddTransCD and FaddItemCD
--
*emember that all of the e$commerce code must aear after the Google (nalytics %racking
.ode calls 7trackPageview89
Secure Pages
The standard Goo"le Analytics Trackin" /ode automatically detects !hen an https protocol is
%ein" used. 'o you !on2t need to add any special trackin" code for secure pa"es. Generally(
you2ll %e placin" ecommerce trackin" code on a secure shoppin" cart pa"e.
.or many e,commerce !e%sites( the checkout process occurs on a separate domain or
su%domain. .or e)ample( if you send customers from !!!.mystore.com to cart.mystore.com(
you2re sendin" them to a su%domain. If either of these scenarios applies to your site( you2ll need
to add some code to some of your pa"es so that you can track activity across domains and
su%domains. The specific methods you2ll use are listed on the slide and you can learn ho! to
use them in the module on trackin" domains and su%domains.
$omains - Su'"omains
Trac,ing Across Multiple Su'"omains
'o far in this course( !e2ve focused on trackin" !ithin a sin"le domain. *efore !e learn ho! to
track across multiple domains( let2s understand !hy !e mi"ht !ant to do this.
A "omain is a hostname that represents a numeric I+ address on the internet. It allo!s us to
easily identify a !e%site %y a name instead of havin" to use a lon" strin" of num%ers. .or
e)ample( Goo"le.com and YouTu%e.com are %oth domains o!ned %y Goo"le.
You may sometimes need to track activity across multiple domains. A common e)ample of this
is !hen you send visitors from your site to a separate shoppin" cart site to complete their
purchases. 1o!ever( since Goo"le Analytics uses e)clusively first party cookies( it can2t
automatically track !hether those visitors actually complete a purchase or not( %ecause the
purchase is takin" place on another site.
+hrased more "enerally( if a session spans multiple domains( it !ould not %e possi%le to track
the session as a sin"le visit attri%uted to one visitor. 'o( you2ll need a !ay of sharin" the cookie
information %et!een the t!o domains.
*y callin" the FlinkCD method( you can send this cookie information across domains.
%his allows Google (nalytics to track a user across multile domains by sending cookies via
)*+ arameters.
To track across domains( you2ll need to follo! t!o steps.
Trac,ing Across $omains%
-E
Add a few lines to the Google Analytics racking !ode on all pages of each site"
/all Fset:omain5ameCD !ith an ar"ument of 4none7 and call FsetAllo!LinkerCD !ith an ar"ument
of 4true7.
& F"a#.pushCXQFset:omain5ame2(2none2YDa
& F"a#.pushCXQFsetAllo!Linker2(trueYDa
Use the #link$% method in all links between domains"
In this e)ample( !e2re updatin" all links from Goo"le.com to YouTu%e.com and vice versa. @e
update each link to call the FlinkCD method as sho!n here.
& ba hrefG7http : !!! .youtu%e .com 7 onclickG7F"a#.pushCXQFlink2(2http:!!! .youtu%e .com 2YDa
returnGfalsea7cGo to our sister site YouTu%ebac
5o!( !hen a user clicks on a link that takes them to the other domain( the session information
is preserved and the user is identified as %ein" the same visitor across %oth domains.
Trans#ering 1ia Forms
If you use a form to transfer your visitors from one domain to another( you !ill need to use the
Flink*y+ostCD method instead of the FlinkCD method.
This situation occurs most often !ith third party shoppin" carts. It can also %e used to send
cookie to other domains in popups or in i.rames.
To use forms to transfer from one domain to another( you must modify all the appropriate forms
!ith the code sho!n here.
& bform actionG7http : !!! .shoppin"cartsite .com myservice form+rocessor .php 7 nameG7f7
methodG7post7 onsu%mitG7"a#.pushCXQFlink*y+ost2( thisYDa7c
& d.
& bformc
The Flink*y+ostCD method !ill chan"e the form action %y addin" #uery,strin" parameters to the
value in the action attri%ute !hen the visitor su%mits the form.
Trac,ing Across Multiple Su'"omains
You may also sometimes need to track across multiple su%domains. A su%domain is part of a
lar"er domain and fre#uently each su%domain contains the pa"es for a specific department or
offerin". .or e)ample( Goo"le.com has several su%domains such as ne!s."oo"le.com(
sites."oo"le.com( and maps."oo"le.com.
'ince Goo"le Analytics uses first,party cookies( cookies set on a su%domain can not
automatically %e read on the main domain( and vice versa. As !ith multiple domains( you need
to e)plicitly share the cookie information %et!een su%domains or you2ll lose session information.
If you don2t share cookie information %et!een your su%domains( it may appear as thou"h your
o!n site is a referrer since only one domain is reco"ni8ed as the main domain.
Trac, across multiple su'"omains
/all Fset:omain5ameCD and specify your parent domain name as the ar"ument. This !ill allo!
the Goo"le Analytics Trackin" /ode to use the same cookies across the su%domains.
-<
.or e)ample( to track across Goo"le2s various su%domains( you !ould call Fset:omain5ameCD
!ith an ar"ument of 4dot "oo"le dot com7 .
& F"a#.pushCXQFset:omain5ame2(2."oo"le.com2YDa
A side effect of usin" this method is that your reports may not differentiate %et!een visits to
identically named pa"es !ithin the various su%domains.
'o( for e)ample( visits to maps."oo"le.comhome.html and mail."oo"le.comhome.html !ould
%e interpreted as visits to a sin"le pa"e. To correct this( you2ll need to set up an advanced filter.
@e2ll e)plain this in a minute.
There are a fe! %est practices for settin" up your Analytics account to track across multiple
su%domains.
.irst( create separate profiles for each su%domain. This !ay( you2ll %e a%le to see reports for
each su%domain.
'et up duplicate profiles , one master profile( plus one profile for each su%domain. In this
e)ample( !e2re lookin" at t!o su%domains.
Your master profile has no filters( and each of the other t!o has an Include filter.
+rofile 1 includes all traffic to %oth su%domains.
+rofile 3 includes only traffic to su%domain A.
+rofile - includes only traffic to su%domain *.
'econd( if you track across several su%domains !ithin one profile( your reports may not
differentiate %et!een visits to identically named pa"es !ithin the various su%domains. %his is
because the reorts only show the *equest )*5 ,, !hich( in this e)ample( is home.html.
The hostname ,, maps."oo"le.com ,, is stored in the 1ostname data field in Goo"le Analytics.
'o( once you2ve called Fset:omain5ameCD to set your primary domain name( visits to
maps."oo"le.comhome.html and mail."oo"le.comhome.html !ould %e interpreted as the same
pa"e,,7home.html7.
To correct this( you can set up an advanced filter to include the su%domain in your reports. 'et
up your filter as sho!n in the slide.
& .ilter Type: /ustom .iltercAdvanced
& .ield A: 1ostname
& ?)tract A: C.*D
& .ield *: 6e#uest ;6I
& ?)tract *: C.*D
& >utput to: 6e#uest ;6I
& /onstructor: SA1S*1
5ote that the constructor must match e)actly !hat is sho!n in the slide( startin" !ith the
for!ard slash.
-H
The filter !orks %y appendin" the 1ostname to the 6e#uest ;6I. As a result( you2ll %e a%le to
distin"uish %et!een identically named pa"es on your su%domains. 6esults appear !ith the
su%domain attached.
Trac,ing Across Multiple $omains wit Multiple Su'"omains
If you !ant to track across %oth multiple domains and su%domains( you2ll need to ensure that
the Analytics cookies are set across the su%domains and that the cookies are %ein" passed
%et!een the parent domains.
There are t!o steps.
.or the first step( add the lines of code sho!n in %lue to Goo"le Analytics Trackin" /ode on
every pa"e of :omain 1 and each of its su%domains.
Lake sure that FsetAllo!LinkerCD has an ar"ument of true and FsetAllo!1ashCD has an
ar"ument of false.
& F"a#.pushCXQFset:omain5ame2(2.domain1.com2YDa
& F"a#.pushCXQFsetAllo!Linker2(trueYDa
& F"a#.pushCXQFsetAllo!1ash2(falseYDa
Then( to each pa"e of :omain 3 and each of its su%domains( add the same code ,, %ut !ith a
different ar"ument to Fset:omain5ameCD.
& F"a#.pushCXQFset:omain5ame2(2.domain3.com2YDa
& F"a#.pushCXQFsetAllo!Linker2(trueYDa
& F"a#.pushCXQFsetAllo!1ash2(falseYDa
.or step 3( call FlinkCD or Flink*y+ostCD in all links and forms that cross %et!een the t!o parent
domains.
.or e)ample( the code sho!n in the slide sho!s ho! you2d do this to track across Goo"le.com
and YouTu%e.com.
& ba hrefG7http : !!! .youtu%e .com 7 onclickG7F"a#.pushCXQFlink2(2http:!!! .youtu%e .com 2YDa
returnGfalsea7cGo to our sister site YouTu%ebac
5ote that you don2t need to use FlinkCD or Flink*y+ostCD in links %et!een su%domains !ithin the
same domain.
A"ain( you should create separate profiles in your account for each primary domain andor each
su%domain.
You can easily do this %y usin" an Include filter %ased on the hostname field.
IN0$EPT3 ANA(>SIS
A"vance" Segmentation
-J
@ith Advanced 'e"ments( you can #uickly isolate and analy8e su%sets of your traffic. You can
create an advanced se"ment that only includes visits that meet a specific set of criteria. 'o( for
e)ample you can create an advanced se"ment that only includes visits from a certain
"eo"raphic re"ion or visits durin" !hich more than S100 !as spent.
$i##erences 'etween Filtere" Pro#iles an" A"vance" Segments
& Advanced se"ments can %e applied to historical data( %ut a filtered profile !ill only filter
traffic "oin" for!ard.
& @hen you create an advanced se"ment( that se"ment is availa%le across all of your
accounts and profiles. *ut( a filtered profile is only useful for a specific !e% property.
& You can compare up to four advanced se"ments side %y side in your reports. In
contrast( filtered profiles can only %e vie!ed one at a time.
& It is much easier to create an advanced se"ment than it is to create a filtered profile.
& If you !ant to permanently affect the data that a profile sho!s( you should use a filtered
profile. 'o if you !ant a profile that only sho!s /+/ data( you should set up a filtered
profile to do this.
& And if you !ant to restrict user access to only a su%set of data( the %est !ay to do this is
to set up a filtered profile and restrict the usersI access to only that profile.
To apply an advanced se"ment( simply /lick Advanced 'e"ments and select the se"ments you
!ant. The :efault 'e"ments are predefined( so you don2t have to do anythin" to use them
e)cept to select them.
>nce you2ve applied one or more advanced se"ments( you can see the data for the se"ments
throu"hout all of your reports. You can also chan"e your date ran"e and see the se"ments
applied to historical data. The se"ments remain applied until you deselect them or you lo"off.
?)ample: Let2s create an advanced se"ment that only includes visits durin" !hich more than
S100 !as spent.
& *e"in %y clickin" the Advanced 'e"ments pulldo!n.
& 5e)t( click /reate a ne! advanced se"ment.
& 5o! you2ll see a screen that looks like this.
& ;sin" this screen( you can com%ine one or more lo"ical statements to define a se"ment.
& To include only visits durin" more than S100 !as spent( first look for the metric
6evenue.
& It2s usually easiest to type !hat you are lookin" for into the search %o)( %ut you can also
%ro!se the complete list of metrics and dimensions.
& 'elect the condition Greater than and specify 100.
& /lick +revie! 'e"ment and you can see the percenta"e of total visits that are included
in the se"ment.
-A
& You can add as many conditions to the se"ment as you like. @hen you2ve finished( click
'ave 'e"ment.
& The se"ment !ill no! appear in the /ustom 'e"ments area of the Advanced 'e"ments
pulldo!n.
Anal)tics Intelligence
Goo"le Analytics Intelli"ence monitors your !e%site2s traffic to detect any si"nificant chan"es
and creates an alert !hen somethin" important happens on your site. It can help you discover
insi"hts that mi"ht other!ise "o unnoticed.
There are t!o kinds of alerts. Automatic in "reen and custom in %lue. Automatic alerts happen
on their o!n !ithout any input from you. /ustom alerts are tri""ered %ased on conditions you
!ant Goo"le Analytics to !atch for. @herever you see a %ar( that means there !as at least one
alert on that date. Let2s look at 030<. There is an alert for a %i" spike in visits for Tech/runch.
>rdinarily !e !ould e)pect %et!een 0 and 330 visits. *ut on this day there !ere over 1N00
visits. If !e "o over to the Tech/runch site( it turns out that Tech/runch did a story on these
ne! Goo"le scarves that are %ein" sold in the Goo"le 'tore. This referral traffic !ould have
%een easy to miss %ut the alert fla""ed it. /lick this %utton and !e can look at =ust the traffic
from Tech/runch . This !as clearly a one time spike. >rdinarily !e only "et a handful of visits.
The importance slider lets you vie! =ust the most si"nificant alerts or %y settin" it to lo! all the
alerts that GA created. The slider doesnIt2 affect the ho! GA creates alerts. It2s =ust a !ay for
you to vie! fe!er or more alerts in your reports.
If you2ve linked your GA account !ith your Ad@ords account( you !ill also "et automatic alerts
from Ad@ords. .or e)ample( GA may si"nal an alert if the clickthrou"h rate for a campai"n
increases une)pectedly. @ith a custom alert you tell GA !hat to !atch for and GA alerts you
!hen it happens.
You create alerts throu"h the settin"s 4co"7 and selectin" the appropriate profile and selectin"
/ustom Alerts. There you can select the parameters for the alert. ?)amples of alerts include
!hen the %ounce rate for a campai"n increases over a specific level or !hen the !eekly
revenue reaches a certain level.
Internal Site Searc
Goo"le Analytics provides internal site search reports that allo! you to see ho! people search
once they2ve arrived at your site.
'o !hy analy8e ho! people search your site$
-N
>n %oth lar"e and small sites( visitors fre#uently use search %o)es as a form of navi"ation. *y
lookin" at !hat people search for( you can identify missin" or hidden content on your site(
improve search results for key phrases( and even "et ideas for ne! key!ords to use in
marketin" campai"ns.
In order to set up 'ite 'earch Trackin" for your !e%site( you2ll need to confi"ure your +rofile
settin"s. /lick the account administration icon at the top ri"ht of any screen in Analytics.
Then navi"ate to the account( !e% property( and profile for !hich you !ant to ena%le 'ite
'earch reports.
In the 'ite 'earch 'ettin"s section( select the I:o Track 'ite 'earchI radio %utton.
In the IQuery +arameterI field( enter the letter( !ord or !ords that desi"nate an internal #uery
parameter. To find out !hat the #uery parameter is( perform a search on your site. 5ormally
!hen a user searches on your site( their #uery can %e found in the ;6L. .or e)ample( if you
search on Goo"le.com( you !ill see your search #uery preceded %y I#GI. Therefore( Goo"leIs
#uery parameter !ould %e I#.I In the e)ample a%ove( the #uery parameter is I#(I and the #uery
!as IGoo"le Analytics2. Your parameter mi"ht %e different ,, it could %e the !ord Rterm7 or
4search7. >r it mi"ht %e =ust a letter( like RsR or 7pR. If you have a particularly lar"e site( some
sections of your site may use different #uery parameters. You may provide up to five
parameters( separatin" each parameter %y a comma.
5e)t( select !hether or not you !ant Goo"le Analytics to strip out the #uery parameter from
your ;6L.
'trippin" out the #uery parameter has the same effect as e)cludin" ;6L Query +arameters
under +rofile 'ettin"s General Information.
'o( if you choose to strip the #uery parameters( you donIt have to also e)clude them from your
main settin"s. 5ote that Goo"le Analytics !ill only strip out the #uery parameters you listed( and
not any other parameters in the same ;6L.
If you use I/ate"oriesQ on your site , such as the a%ility to use drop,do!n menus to narro! a
search , you can include cate"ories in your search analytics.
.irst( select the 4'ite search cate"ories7 check%o).
Then( enter your I/ate"ory +arameterI in the field provided. ?nter only the letters that desi"nate
an internal #uery cate"ory such as Icat( #c(I. The same principle that you used to identify the
#uery parameter can %e used to identify the cate"ory parameter. >r( you can contact your
!e%master to identify the #uery and cate"ory parameters for your site.
E0
:ecide if you !ant to strip out the cate"ory parameters that you =ust provided. If you select the
check%o)( only the parameters you provided !ill %e stripped out.
As !ith the #uery parameter settin"( this has the same effect as e)cludin" ;6L Query
+arameters in the General Information section.
'o( if you choose to strip the cate"ory parameters here( you donIt have to e)clude them a"ain
from your main settin"s.
/lick 2ApplyI to finish.
To find the 'ite 'earch reports( click 'ite 'earch under /ontent.
Site Searc +sage Report compares visitors !ho used site search to those !ho did not.
1ere !e can see that 1NB of all visits to this site included a search.
Kust a%ove the pie,chart( you2ll notice t!o dropdo!n menus. if you select Goal /onversion 6ate
in the left,most dropdo!n( you can see ho! visits that included search compare to visits that did
not include search !ith respect to conversions.
And( you can click the ecommerce ta% to see ho! revenue and other ecommerce metrics differ
for visits !ith and !ithout site search.
Searc Terms Report only includes visits durin" !hich a search !as performed. .rom the
screenshot on the slide( you can see that there !ere JH(--1 total uni#ue searches. The search
terms are listed in the ta%le. You can see ho! each term compares in terms of num%er of
searches( percenta"e of search refinements( and other metrics. Lookin" at the search terms
that people use to search once they are on your site can "ive you ideas for key!ords that mi"ht
also help drive traffic to your site.
You can look at this traffic %y another dimension. .or e)ample( if you !anted to see !hich cities
these visitors came from( you could select /ity from the :imension dropdo!n.
Start Pages lists all of the pa"es from !hich visitors searched. To find 'tart +a"es( click
+a"es under 'ite 'earch. Then( select 4'tart +a"e7 as the vie!in" option a%ove the ta%le. /lick
on a pa"e in the ta%le to learn more a%out the searches that occurred from that pa"e. A detail
report !ill appear !hich lists all of the search terms that !ere used from that pa"e. You can use
this report to find out !hat visitors are searchin" for from your landin" pa"es and you can use
the information to improve the pa"e content. .or e)ample( if many visitors search on Rshippin"
optionsR from your shoppin" cart pa"e( you may !ant to display shippin" information directly on
the pa"e.
E1
$estination Pages tells you !hich pa"es are most commonly found throu"h the search on
your site. To find :estination +a"es( click +a"es under 'ite 'earch. Then( select 4:estination
+a"e7 as the vie!in" option a%ove the ta%le. The ta%le sho!s popular destination pa"es. /lick
on a pa"e in the ta%le to see the specific search terms that led to the pa"e.
You can see !hich cate"ories your visitors selected !hen searchin" your site.
Go to the 'earch Terms report and click 4'ite 'earch /ate"ory7 as the vie!in" option.
This information helps you understand ho! visitors use your search en"ine( !hich product
areas and cate"ories are most popular( and ho! successfully visitors find !hat they are lookin"
for in each cate"ory.
Your 'ite 'earch reports !ill "enerally sho! a different num%er of conversions than !hat is
sho!n in all of your other reports.
%his is because goal conversions in the /ite /earch reorts are based on visits that include at
least one search on your website whereas the goal conversions shown in all other reorts are
based on all visits.
*ecause 'ite 'earch reports only include conversions from visits that included a search( you
can see ho! effectively searches on your site drive conversions.
If you are confused a%out the difference %et!een 4search term7 and 4key!ord7( it2s helpful to
remem%er that Goo"le Analytics reports use 4search term7 !hen referrin" to internal site
searches and 4key!ord7 !hen referrin" to e)ternal searches.
Event Trac,ing an" 1irtual Pageviews
Lany !e%sites use technolo"ies such as .lash and A=a) to interact !ith visitors. .or e)ample(
some !e%sites em%ed video players( "ames( and other interactive e)periences on site pa"es.
1o!ever( the %asic !e% analytics model of trackin" pa"evie!s doesn2t capture these kinds of
interactions. This is %ecause !hen a visitor interacts !ith a video player( for e)ample( no
pa"evie! is "enerated.
'ome other e)amples of interactions that don2t "enerate pa"evie!s are flash !e%sites( video
players( dynamically "enerated pa"es Cc"i( asp( phpD( Kavascript O A=a),%ased activities( file
do!nloads( and clicks on links that take the visitor to another site Ce)ternal linksD.
'o ho! do you track these kinds of activities$ There are t!o !ays: virtual pa"evie!s and ?vent
Trackin".
E3
1irtual Pageviews
You can create a virtual pa"evie! to represent practically any kind of activity or interaction you
!ant. You simply call Ftrack+a"evie!CD and provide any name you !ant as the ar"ument. It2s
4virtual7 %ecause you2re tellin" Goo"le Analytics to re"ister a pa"evie! even thou"h no ne!
pa"e has actually %een loaded. You2ll see these virtual pa"evie!s alon"side ordinary
pa"evie!s in the +a"es and /ontent :rilldo!n reports.
'o you !ould call Ftrack+a"evie!CD each time an event occurs that you !ant to track:
& F"a#.pushCXQFtrack+a"evie!2YDa
+rovide a 4filename7 ar"ument that identifies the event. .or e)ample:
& F"a#.pushCXQFtrack+a"evie!2(2eventsplayvideo2YDa
If you look at the Goo"le Analytics Trackin" /ode( you2ll notice that it calls Ftrack+a"evie!CD.
This lets Goo"le Analytics kno! that the %ro!ser has loaded a pa"e. @hen you call
Ftrack+a"evie!CD( ho!ever( you2ll !ant to provide an ar"ument that specifies a virtual
pa"ename for the event you2re trackin". Additional Ftrack+a"evie!CD calls should %e made after
re"ular GA Trackin" /ode.
1ere are some more e)amples.
In the first e)ample( !e2re trackin" a do!nload.
& ba hrefG7http : !!! .e)ample .com 7
on/lickG7F"a#.pushCXQFtrack+a"evie!2(2do!nloade)ample.pdf2YDa7c
In the second e)ample( !e2re trackin" a .lash event.
& on CreleaseD \
& Track !ith no action
& "et ;6LC7F"a#.pushCXQFtrack+a"evie!2(2do!nloade)ample.pdf2YDa7Da
& ]
In each of these cases( !e2re simply callin" Ftrack+a"evie!CD to re"ister a virtual pa"evie!.
It2s a "ood idea to adopt a clear namin" convention for your virtual pa"evie!s. You mi"ht( for
e)ample( "roup virtual pa"evie!s into cate"ories %y "ivin" them a virtual su%directory Ce)ample:
do!nload%rochure.pdf or flashmovie1playD.
Also( since virtual pa"evie!s appear alon" !ith standard pa"evie!s in reports( you may !ish to
create a duplicate profile !here you filter out the virtual pa"evie!s.
To make this easy( you mi"ht or"ani8e all of your virtual pa"evie!s into a 4virtual7 directory.
CvirtualD.
Event Trac,ing
E-
The other !ay to track non,pa"evie! interactions is to use ?vent Trackin". >ne advanta"e of
usin" ?vent Trackin" is that you !on2t "enerate an e)tra pa"evie! each time an interaction
occurs.
Another advanta"e is that you can easily or"ani8e your events into cate"ories( actions( and
la%els. And you can even provide values for each event you track.
All of your events sho! up in the ?vents reports !ithin the /ontent section.
Kust call the Ftrack?ventCD method each time you !ant to re"ister an event. The slide sho!s the
full specification of Ftrack?ventCD ,, !hich you can also find documented on the Goo"le /ode
site ,, and ho! you !ould actually call it( assumin" that you are usin" Asynchronous Trackin".
& Ftrack?ventCcate"ory( action(optionalFla%el(optionalFvalueD
You can call it as follo!s:
& F"a#.pushCXQFtrack?vent2(2name2(2la%el2(value YDa
@e2ll discuss the ar"uments to Ftrack?ventCD in a minute.
1ere2s an e)ample of ho! you2d call Ftrack?ventCD from a .lash video player. In this e)ample(
Ftrack?vent !ill "et called each time the visitor releases the +lay %utton on the video player.
Ftrack?vent !ill re"ister an event !ith a cate"ory name of 49ideos7( an Action name of 4+lay7(
and a La%el of 4Lovie :rama7.
& on6eleaseC%uttonD\
& "et;6LC4=avascript:F"a#.pushCXQFtrack?vent2( Q9ideos2( Q+lay2( QLovie :rama2YDa7D]
Let2s look at each of the ar"uments to Ftrack?vent.
&ategor) 9?st Argument;
& The name you supply for the "roup of o%=ects you !ant to track e.". Q9ideos2.
& /ate"ory is a name that you supply as a means to "roup o%=ects ,, !hich are usually
user interface elements that you !ant to track.
& 'o( for e)ample( if you have "ames and videos on your site( you2d pro%a%ly !ant to have
a 4Games7 cate"ory and 49ideos7 cate"ory.
& /lick 4?vent /ate"ory7 in the Top ?vents report to see all the user interface elements
!ith !hich your visitors interacted.
Action C3nd Ar"umentD
& A strin" that is uni#uely paired !ith each cate"ory( and commonly used to define the
type of user interaction for the !e% o%=ect( e.". 4+lay7 or 4'top7 %utton clicks.
& Action is the name you !ant to "ive to the type of interaction you2re trackin".
& 'o( for e)ample( for 9ideos( you2d pro%a%ly !ant to track ho! many times your visitors
pressed +lay.
& /lick 4Actions7 in the Top ?vents report to see the interactions that occurred.
La%el C>ptional -rd Aru"mentD
& .urther desi"nation "iven to the event e.". 4Titanic7 or 4.indin" 5emo7.
EE
& The La%el ar"ument is optional. A La%el allo!s you to provide additional information for
for the event you are trackin".
& .or e)ample( if you are trackin" video plays( you mi"ht use the La%el ar"ument to
specify the name of the movie that !as played.
& >r( for file do!nloads( you mi"ht use it for the name of the file %ein" do!nloaded.
& /lick 4La%els7 in the Top ?vents report to see the the La%els of of the events that
occurred.
9alue C>ptional Eth Aru"mentD
& 5umerical value assi"ned to the event e.". load time.
& 9alue is the fourth( and optional( ar"ument to Ftrack?ventCD.
& ;nlike the other ar"uments !hich are all strin"s( 9alue is an inte"er. You can use it to
assi"n a numeric value to a tracked pa"e o%=ect.
& You2ll then %e a%le to see a sum total of these values in the ?vent 9alue column of your
?vents reports.
& You2ll also %e a%le to see an avera"e of these values in the Av". 9alue column of your
?vents reports.
& 'o( you mi"ht( for e)ample( specify a dollar value !hen a specific play%ack marker is
reached on your video player. To call Ftrack?ventCD !ithout a value( simply omit the
ar"ument.
& >ther e)amples: pass the time in seconds for a video to loada
The strin"s that you provide for the first - ar"uments( /ate"ory( Action( and La%el( "overn ho!
the events !ill %e or"ani8ed in your reports.
'o( you2ll !ant to think carefully a%out ho! you !ant to structure your events.
In your reports( you2ll notice that %oth Total ?vents and ;ni#ue ?vents are counted.
Total ?vents is simply the total num%er of times an event occurs ,, really it2s =ust the num%er of
times Ftrack?vent !as called.
*ut( for ;ni#ue ?vents( each particular event is only counted once per visit.
'o( if durin" a sin"le visit( a visitor presses +lay < times on the same movie( Total ?vents !ill %e
incremented %y <.
*ut ;ni#ue ?vents !ill only %e incremented %y 1( %ecause for ;ni#ue ?vents( a particular event
is only counted once per visit.
As !e mentioned earlier( the ar"uments you provide !hen you call Ftrack?vent !ill "overn ho!
events are or"ani8ed in your reports.
'o( %efore you add the calls to Ftrack?vent to your site( consider these %est practices.
E<
.irst( determine in advance all of the kinds of events you2ll !ant to track.
Try to create a hierarchy of /ate"ories( Actions( and La%els that !ill "ro! !ith your needs.
@ork !ith your report users to make sure that the hierarchy makes sense.
And use a clear and consistent namin" convention for your /ate"ories( Actions( and La%els.
;sin" track?ventCD allo!s you to analy8e event %ased interactions in much "reater detail than is
possi%le usin" virtual pa"evie!s.
.or e)ample( instead of =ust seein" ho! many times a movie !as played on your site( you can
analy8e ho! people use your video player( and see ho! different events correlate !ith site
usa"e and ecommerce metrics.
Also( %y trackin" events separately from pa"evie!s( you !on2t inflate your pa"evie! count.
A""itional &ustomi=ations
3ow To &ange Session Timeout 1alue
In Goo"le Analytics( a visiteor sessioneis defined %y -0 minutes of inactivity( or !hen a user
#uits the %ro!ser. You can chan"e the -0 minute default %y callin" set'ession/ookieTimeout
as sho!n in the slide. 'imply specify a ne! timeout value in milliseconds as the ar"ument to
Fset'ession/ookieTimeoutCD.
& F"a#.pushCXQset'ession/ookieTimeout2(-H00000YDachan"e session to 1 hour
3ow to &ange &ampaign E*piration
*y default( a conversion can %e attri%uted to a campai"n that is up to H months old. *ut( if your
%usiness has a lon"er or shorter marketin" campai"n timeframe( you can chan"e this value.
Kust call Fset/ampai"n/ookieTimeoutCD and specify your ne! campai"n len"th in milliseconds.
.or e)ample( let2s say that you !ant to set a campai"n len"th of -0 days.
To fi"ure out the num%er of milliseconds that is( type 4-0 days in milliseconds7 into Goo"le
'earch.
The search en"ine !ill "ive you the ans!er !hich you can plu" into
Fset/ampai"n/ookieTimeoutCD.
& F"a#.pushCXQFset/ampai"n/ookieTimeout2(3<N3000000YDa
3ow to &ange &ampaign Prece"ence
Goo"le Analytics attri%utes conversions to the campai"n that most recently referred the visitor.
EH
.or e)ample( let2s say that someone discovers your site %y clickin" one of your Ad@ords ads.
Then( they come %ack to your site %y clickin" a %anner ad that you2ve ta""ed !ith campai"n
varia%les. This time( they convert to one of your "oals.
*y default( the %anner ad !ill "et the credit for the conversion( not the Ad@ords ad that
ori"inally referred them.
To chan"e this %ehavior( you can ta" all of your campai"n links !ith utmFnooverrideG1.
If you do this consistently !ith all of your campai"ns( Goo"le Analytics !ill attri%ute conversions
to the first referrin" campai"n( instead of the most recent one.
5ote that the utmFnooverride settin" can %e used in con=unction !ith autota""in".
& http : !!! .mysite .com $
utm F source G "oo"le O utm F medium G cpc O utm F campai"n G %acktoschool O utm F nooverride G
1
3ow to A"" Searc Engines%
Goo"le Analytics automatically tracks referrals from over -0 search en"ines.
*ut( if you !ant to add a search en"ine( you can do it %y callin" Fadd>r"anicCD in your Goo"le
Analytics Trackin" /ode.
.irst( perform a search in the search en"ine and look at the ;6L of the search results pa"e.
In the ;6L( look for the key!ord you searched ,, it should %e preceded %y a letter and an e#ual
si"n. This letter is the #uery varia%le for the search en"ine.
In the e)ample( the #uery varia%le is 4p7.
Add a call to Fadd>r"anic in your Goo"le Analytics Trackin" /ode. The first ar"ument is the
name of the search en"ine. The second ar"ument is the #uery varia%le.
& F"a#.pushCXQFadd>r"anic2(2supersearch2(2p2YDa
3ow To Treat &ertain @e)wor"s As $irect
You may !ish to treat traffic that results from certain search key!ords as :irect.
.or e)ample( if someone searches for the e)act name of your site( you mi"ht !ant to treat that
visit as a :irect visit instead of a search.
To do this( simply add a call to FaddI"nored>r"anicCD in your Goo"le Analytics Trackin" /ode.
'pecify the key!ord as the ar"ument.
& "a#.pushCXQFaddI"nored>r"anic2(2"oo"leanalytics2YDa
3ow to Treat &ertain Re#erring Sites As $irect
You can also treat referrals from certain sites as :irect traffic instead of as referrals.
.or each site that you !ant to e)clude as a referral and treat as :irect( add a call to
FaddI"nored6efCD in your Goo"le Analytics Trackin" /ode.
'pecify the name of the site as the ar"ument.
& "a#.pushCXQFaddI"nored6ef2.2!!!.sister,site.com2YDa
EJ
EA