Many Site­Cat­a­lyst users are famil­iar with the Time Spent on Page met­ric in the Pages report (which shows the aver­age amount of time that vis­i­tors spend on the var­i­ous pages of your site) and with the Time Spent per Visit report, which dis­plays over­all visit times as well as the amount of time that users spend prior to com­plet­ing any suc­cess event. But what if you’d like to see “time spent” for some­thing other than a page?

A user asked me just such a ques­tion a cou­ple of weeks ago via Twit­ter. He had a Flash appli­ca­tion and wanted to see the aver­age amount of time spent inter­act­ing with the app. It’s a KPI that makes tons of sense when opti­miz­ing flow and struc­ture of an app, site sec­tion, or some other por­tion of the user expe­ri­ence that doesn’t fit neatly into the def­i­n­i­tion of a “page.” For­tu­nately, this sort of mea­sure­ment is entirely pos­si­ble and fairly easy to implement.

Site­Cat­a­lyst pathing allows you to see how users moved from one value of a traf­fic vari­able to another. When you view the Time Spent on Page met­ric in the Pages report for “home page,” what you’re really see­ing is the aver­age time elapsed from one page name (“home page”) to another (such as “prod­uct page”). The fact that Time Spent cal­cu­la­tions are based on changes in vari­able val­ues is impor­tant, because it means that with a bit of plan­ning, you can cap­ture Time Spent data for var­i­ous site sec­tions and appli­ca­tions sim­ply by con­trol­ling how traf­fic vari­ables change from one page view or user action to the next.

First, you’re going to need one Cus­tom Traf­fic (s.prop) vari­able, or per­haps the Site Sec­tions (s.channel) vari­able if you’re look­ing time spent on a cer­tain site sec­tion. Regard­less of how you choose to do it, make sure to get pathing enabled on this vari­able. Your Account Man­ager can help set this up.

Once you’ve got your vari­able, set it on every page view (and, where applic­a­ble, cus­tom links) regard­less of whether the page or link falls inside or out­side of the area/app in ques­tion. If you’re using s.channel, come up with a good strat­egy for group­ing pages and areas of the site into sec­tions. In a more spe­cific case, such as time spent on a Flash app, you can set the vari­able to any value (e.g., the page name) out­side of the app, and the app name or some other identifier/differentiator inside of the app. This iden­ti­fier should stay the same as the user moves through the site sec­tion or app in ques­tion! Because of the way that pathing works, you’ll want to use the same value for every page view or user action that should be “grouped” in terms of time spent. To see time spent on a Flash app, you would use the same value in this vari­able as the user moves through the app.

Exam­ple:

Land­ing Page: s.prop1=“landing page“
Sec­ond Page of Visit: s.prop1=“some other, unre­lated page”

…now the user enters the Flash app and clicks around a cou­ple of times…

On app load: s.prop1=“flash app — super fun game v2.0″
First click: s.prop1=“flash app — super fun game v2.0″
Sec­ond click: s.prop1=“flash app — super fun game v2.0″

…notice that the value of prop1 stayed the same within the app…

Third Page of Visit, now out­side of app: s.prop1=“a third unre­lated page”

Because time spent met­rics describe the amount of time until a variable’s value changed, an imple­men­ta­tion like the one shown above would cause all of the time spent in the Flash app to be reported together, as a sin­gle “time spent” total, because the value of the vari­able never changed within the app (or site section).

The result­ing report might look some­thing like this (pre­sum­ing, again, that pathing has been enabled on the vari­able in question):

Time Spent

So, by care­fully imple­ment­ing the same vari­able with the same value as users move into, through, and out of key site sec­tions and apps on your site, you can eas­ily deter­mine the time spent inter­act­ing with key areas of your site so that you can gauge suc­cess and bet­ter deter­mine ways to improve your site.

BONUS: This can also show you how peo­ple move from site sec­tion to site sec­tion, or how they move into and out of your site sec­tions and apps.

As always, please leave a com­ment with any ques­tions, thoughts, or sug­ges­tions that you may have! I’m also avail­able Twit­ter or LinkedIn, or by e-mailing omni­ture care [at] adobe dot com.

6 comments
Bella Benedict
Bella Benedict

Great post. Waiting for you to continue the topic. Bella Benedict escort service warsaw indiana

Bella Swenson
Bella Swenson

It is very interesting for me to read that post. Thanx for it. I like such themes and everything connected to this matter. I would like to read a bit more on that blog soon. Bella Swenson escort privat schweiz hausbesuch

Rob Blakeley
Rob Blakeley

I believe it is worth noting that one needs to make sure everything passes a value for the chosen variable. For example, say you have "sponsored" pages in Prop1. No not all of your pages are sponsored so some pages may not have a sponsored value. When you move from sponsored page #3 to some non-sponsored pages where there is no value set, the time in those non-sponsored pages can be counted in the sponsored time spent. This is because the last sponsored value did not change when you looked at the non-sponsored pages. However, if you pass a value such as "not-coded" on non-sponsored pages the value will change and the time spent in non-sponsored pages will not be counted. Time spent also appears to be variable based, so that a value passed in prop2 will not affect the time spent calculation for prop1.

nic
nic

Hi Ben, there is one - probably important - addition to this: A "change" of an s.prop value is not happening, if the next click does not populate the s.prop at all. When it comes to Time Spent, the s.prop behaves like a visit-based eVar. Example: 1. s.pageName="pathtest1" and s.channel="pathstart" 2. s.pageName="pathtest2" and s.channel="" 3. s.pageName="pathtest3" and s.channel="pathend" Pages-Report with "Time spent": * pathtest1: 2 Minutes * pathtest2: 2 Minutes * pathtest3: 0 Minutes And a look at the Site Section report with "Time spent": * pathstart: 4 Minutes * pathend: 0 Minutes So it is dangerous to use Time Spent, if you can't guarantee, your s.prop is set all. so long nic

Joy
Joy

Great post Ben! "Time on..." metrics can be confusing, but this post and the knowledge base helped me too. I had a follow-up question about time on page: Let's say we have a custom traffic variable (prop1), that shows site sub-sections (electronics, housewares, apparel, etc). The sub-sections are made up of pages (naturally). If I run a pages report, filter by "housewares" and add up all the time on page data, will that total time on page be the same as if i ran my prop1 (filter by "housewares") report? Thanks!

Ben Gaines
Ben Gaines

Joy, assuming that Site Sections (s.channel) is always populated on page views and does not change to a different value on custom link calls, and that you're sure to include all of the pages that belong to the site section, then yes. They should at least be very close (depending on rounding and bucketing). There may be exceptions that I haven't seen/considered, but fundamentally I think you're understanding the way that this works.