Analyzing events on Exit Pages

[elenae1993]

Hello,

Currently Adobe Analytics provides insight into Exit Pages - type, page views, exit rate etc. However, these are all page metrics. Is there a way to also analyse the events that took place on the Exit pages?

Using a fallout funnel I could set up, for example, Page Type = Product Page > Then add File download exists > but what would be the exit metric? Because if I try adding Exits as a metric it doesn't let me and if I try adding Exit Page = Product Page then it wouldn't make sense?

We thought of potentially setting up a prop on Actions but was wondering if there was another way.

Thank you!

[Andrew_Chepurny]

Hey there Elena,

You can certainly tie in an exit event into an eVar so you can set it to only count on the last hit of the visit.  This way you can tie it to whatever product view you want.  Or customize it as such.  This would be the only way I can think of; having to use an eVar that is.

Hope this helps!

Andrew

[PratheepArunRaj]

Dear Elena,

Guess that would also help, give a try.

Exit Page = Your Product Page and File Downloads exits (In visit container or visitor container).

You can also add a sequence of your wish.

Thank You!

Arun

[Urs_Boller · Accepted solution]

I think this is really an advanced topic. however, I try to give you a solution.

first, "exit page" is not really what you would expect until you understand how data is processed. basically, the "exit page" is a dimension which records the last "page" throughout a visit. what might be clear, gets difficult if you start using other metrics than visit or break down by other dimensions.... why? the "exit page" is valid throughout the whole visit!

best is to give you an example with a user journey

page 1: Landingpage

page 2: product page + download (eg. product description)

page 3: checkout page + purchase event

if you start to analyse this journey, you would get "exit page = checkout page". but if using "download instances" as metric you would see a 1!

the reason is that the "exit page" is valud for all those pages/hits in the whole visit. and there is one hit "download" where the "exit page" (of the visit!) was the "checkout page".

I believe it is important to understand that "exit XY" (as well as "entry XY") are dimensions valid for the whole visit and breaking down by other dimensions (or using other events) could not show what you want.

solution: sequential segment

but lets get back to your question: you want the exit pages and actions users have taken on this specific page. what you need is a segment that really has the "hits" (data) of the last page and the corresponding actions. the trick is to use a fancy "sequential segment" to filter out all hits before the last page. it looks as follow:

what you have within the outer "hit-segment" is a visit container (you can add one on the right using the gear icon). this inner container has a "sequential segment" which looks for "day" then "page view" with "sequence before" and "exclude visits". (change  sequence and exclude behind the gear icon).

a little bit complicated to explain in a post, but this segments returns all "data" of the last page view and the actions after (downloads, exits, ...)

using the example above the data you examine is just the "checkout page" and the "purchase event". so start analysing what users have done on the last page.

for example use your segment to show what have been the pages (using the "pagname" dimension and "pageviews" as metric) or exit link clicks (using "exit link name" and "exit link clicks"). there are a lot more you can add to the table (metrics) or using breakdowns, all valid since you are looking at only the last page.

I hope that helps answering the question you have. let me know if you need more explanations.

[elenae1993]

This is a really helpful explanation - thank you! Just one question - why the 'Day exists'?

[Urs_Boller]

hi elenae1993​

basically you need to define a sequence of "any hit" then "page view". because there is no "any hit" condition, you need to select a dimension which really exists one every single server call. and since the OOB dimension "day" matches this criteria I decided to choose this one. but you could as well take any other dimension you have on every single server call....