Many site owners spend a lot of time creating content that is supposed to drive conversions. But what’s the best way to measure the performance of this content with Google Analytics? How can we measure a specific piece of content, be it a page or a piece of creative, and it effect on conversions?
Google Analytics has a metric called $Index to help measure the “value” of site pages. But the problem with $Index is that it is an average, and averages can be skewed very easily. $Index is about the performance of a page, not the content on a page. Also, many people want to know how many times a piece of content directly led to a conversion. We just can’t get that with $Index.
We could view this type of analysis as a navigation analysis. Google Analytics has the All Navigation report and the Initial Navigation report, but these reports track things that happen in under 3 clicks and not everyone converts in 3 clicks.
Rather than tackle this problem using navigational analysis, let’s consider it a content challenge. What we want to do is see if a specific piece of content ultimately lead to conversion.
Given this approach we could use the Site Overlay report, which is supposed to show the performance of each link on a page. But, in my experience, the Site Overlay report is buggy at best (and I’m being nice).
We need is a way to link a piece of content, i.e. a pageview, to a conversion. There’s a very simple way to do this using the Funnel Visualization report.
The Concept
Each funnel has a ‘required step’ setting. When enabled, this setting requires that the visitor views the first step in the funnel prior to conversion. If the visitor does not see the first step then the Funnel Visualization report will not count a conversion. The conversion will still be recorded in all other reports, but not the Funnel Visualization report.
What few people know is that it does not matter when the ‘required step’ is viewed, as long as it is viewed prior to the conversion.
We’re going to use this setting to associate a conversion with the content we want to evaluate.
Example
Let’s say I want to track how many people view the About Me page on this blog before subscribing to my RSS feed. I can create a goal and funnel that links the About Me pageview to the RSS subscription goal.
Step 1: Set up the goal
The first step is to create the goal. Just set up the goal like any other normal goal. Identify the goal URL, give the goal a name and a value (if you so desire).
Step 2: Identify the “required step”
Now let’s turn to the funnel. Remember, step 1 in the funnel, the ‘required step’, is really the piece of content (i.e. the pageview) we want to evaluate in terms of conversions. Simply add the page URL to the Step 1 URL field, give the step a name, and check the “required step” checkbox.
That’s it! There’s nothing else to do. The funnel visualization report, for this specific funnel, will only show a conversion if the visitor views the About Me page at some point prior to conversion. GA doesn’t care when the visitor sees the page, as long as they see the page prior to conversion.
The Data
Here’s a sample Funnel report. We can see that 4 conversions occurred after viewing the about me page. Remember, it does not matter when the About Me page was viewed, as long as it was viewed prior to conversion.
Now, if we compare the number of conversions in the Funnel Visualization report, to the overall number of conversions for this goal, we notice there is a difference.
The difference is the number of visits that did not include the About Me page prior to subscribing to the RSS feed. There were 8 total RSS Subscription conversions, but only 4 of those conversions viewed the About Me page prior to converting. Now we know how effective the About Me page was at driving RSS subscriptions.
Taking it Further
What about associating a set of pages with a conversion activity? No problem, just use a regular expression to define your required step. Here’s the same example, but I’ve tweaked to to track visits that include the About Me page or the All Posts page.
And remember, the pageview you specify for your required step does not need to be a real “pageview.” It can be a virtual pageview, generated with the pageTracker._trackPageview() method. In fact, that’s what I’m doing on the blog. I generate a virtual pageview every time someone clicks on the RSS icon.
This technique is very useful if you want to measure how well a specific piece of content on a page is performing. Generate a pageview when a visitor clicks on the content and use it as step 1 in the funnel.
Think this is a good idea? Got one that’s better? Leave a comment!
Hi,
Nice Post :)
My question is how to how to create a funnel when the
first step is “visitors who entered on Homepage”
Regards
Gunjan
Hi Justin, when you say “the pageview you specify for your required step does not need to be a real “pageview.”It can be a virtual pageview, generated with the pageTracker._trackPageview() method.” do you mean that GA automatically generates the pageview itself, or do you have to apply a setting somewhere in GA ?
Gunjan The technique described above works great for linking the homepage to conversion. But, it is very difficult to identify when the hompage was the entry page to the site. Google Analytics does not make the session landing page available via any of the filters, so we can’t tell which page was the landing page.
If you wanted to get creative you could write some custom JS that tracks the session landing page and then creates a “virtual” pageview if the homepage is the landing page.
Jerome If you would like to capture a visitor’s action, like clicking on a banner or clicking on a specific piece of content, you can create a “virtual” pageview when the click happens by adding trackPageview() to the onClick event. You can then use the “virtual” pageview as the first step in your funnel.
I’ve got an old post that covers tracking clicks in Google Analytics.
I hope that clears up any confusion.
Thanks to both of you for the questions.
Justin
Justin – Great post. I use this trick a lot. its great. but I got the impression that the order of viewing the pages is not important as long as it is in the same session. is it not true?
Gunjan – you can view the traffic sources report, segment it by landing page and press the goal conversion tab to see the conversion rate for every entry page.
Hi Zvika,
Glad you like the post. It’s a very handy trick.
You are correct. It does not matter when the visitor sees the required step during the session. It could be at the beginning, the middle or the end. As long as it happens.
Thanks for the comment.
Justin
Hi Justin…
Good to have you back with great tips!
:)
1. You say:
“Google Analytics has the All Navigation report and the Initial Navigation report”
??? Am i missing something?
I’m not familiar with these reports…
at least not under that name…
2. The “Exit” box in the funnel visualization report is actually turning into a Black box (:) when the funnel is configured with more than 3 steps. it would be great if you could explain how to read it right…
Thanks,
Amit.
Hi Justin,
Great post. I’m curious about the interpretation of the results. Using your example, I’ve been interpreting the results as “4 people who subscribed came directly from the About Me page.” That is, they have to view About Me immediately before subscribing (the reason why there are 0 new entries to the funnel for each subsequent step). Would that be correct?
Blam: The funnel visualization report is showing that there were 10 unique pageviews for the About Me page. This means that there were 10 visits that included the about me page.
4 of those 10 visits resulted in a goal conversion (in this case it was an RSS subscription). It does not matter if the visitor viewed the About Me page immediately before converting.
We can validate this by looking at the All Navigation report for the About Me Page. I don’t have a screenshot right now, but the data shows that there was only one click from the About Me page to the RSS subscription.
Amit Thanks for the knd words. It’s good to be back. To answer your questions:
1. The All navigation report and the Initial Navigation reports are listed under the “Analysis Options” when viewing a page detail in the Top Content Report.
2. I’m not sure I understand this question. Could you elaborate a little bit more?
Thanks for the questions!
Justin
Hi Justin…
So when you say “Analysis options”
you’re referring to the old GA version aren’t you?
:)
Thanks for a great tip! I’ll definitely be using this in the future.
Amit,
Sorry, it’s not Analysis Options, it’s “Analyze”.
Go to the Top Content report and choose a page from your sight. You should notice that there is an “Analyze” drop down box just above the data table. It’s near the Segment drop down box.
Using the options in the Analyze drop down box you can analyze how people navigate to and from the content that you chose from the top content report.
Hope that clears up any confusion.
Justin
Yep it does.
It’s the same as “Navigation Analysis”
which is a fix area right under the segment drop down..
Well you sure got me frightened for a moment
:)
Very cool tip Justin! Thanks for that :)
Good to have you back in the blogosphere!
Best,
Mike
Site overlay will not show the same information?
Clicks, revenue, Conversion.
Hi Eugene,
Great observation. Site overlay does something very similar. However, site overlay is buggy at best. Plus, if there are multiple links on a page that point to the same destination, then the site overlay will show the same data for both links.
Finally, this method can be used to literally associate a click with a conversion, something that can not be done with site overlay.
Thanks for the comment,
Justin
Hi Justin,
May I know that did there happened technical issue in GA please?
It’s because I found out that my site’s total visits & pageviews has dropped more than 50% as compared with previous days.
Thus I try to observe in another report:
Visitors>>Visitor Trending>>Pageviews by hourly
The pageviews has a sudden drop from almost 3k to 100+ since 11am which seems very abnormal. And this scenario has continued till 29thJuly..
Do you have any idea on this please?
Thanks,
Crystal
Crystal,
I don’t know of any issues. Have you segmented the data to determine if the drop off is due to a particular segment?
Also, have you made any changes to your account? A new filter might be having adverse effect.
Hope that helps.
Justin
Justin,
I have set up funnels in GA that look to be working properly. The 1st step of the funnel is the home page and is not marked as required. When i look at the number of visits in on the homepage within the funnel visualization report, there is a discrepancy with that of the content report for the same page. Are the Funnel reports tracking absolute unique visits, total visits….? can you help explain why the numbers on the content report would differ to those on the funnel report for the same page?
thanks for the help.
-Michael
Michael,
The funnel visualization report is based on unique pageviews. So, the number of times that a visitor hits a certain page in the funnel should match the number of unique pageviews in the content reports.
Try reconciling those numbers, they should match.
I also wanted to point out that if you plan to use the technique above you should make the first step in the process a required step. That’s the key to showing the relationship between the conversion and a piece of content.
Hope that helps and thanks for the question.
Justin
Justin-
Your blog posts–and followups–are amazing… well done! I’m starting to feel pretty good about GA, but find some basic things are surprisingly non-obvious to achieve… maybe you can point me in the right direction–perhaps a custom report or even your own AnalyticsView tool…
I’d like to compare add-to-cart conversions for a bunch of different products… say p0.php to p9.php, with the cart page being cart.php. I guess I’m looking for a table with the following columns:
– product page URI
– # distinct product page views (a)
– # *subsequent* cart views (b)
– % add-to-cart conversions (b/a)
like this:
p0.php 1234 123 10%
p1.php 1024 51 5%
etc.
The challenge I realize is that GA seems visit/session-based, so that it normally doesn’t report the immediate sequencing of pages (with the exception of the Navigation Summary)…
If I didn’t care about sequencing, the closest I can imagine is making Goal1=cart.php a custom report with dimension=Page and metrics={UniquePageViews,Goal1_Completions} … but I’m not sure if that data is right, and besides, the lack of sequence-awareness dilutes its meaning significantly.
So since goals don’t seem to consider sequence, it seems like that’s not the right tool. Is there another way to get the above table out of GA?
Daniel,
Thanks, glad you find the blog useful.
I think the biggest problem is that some people will add items to the cart during their visit. Because a visitor can only convert once per visit we won’t capture every time a product is added to the cart.
With that said, I wonder if it’s possible to use event tracking to track something like this. Perhaps you can use capture the product view as an action and then the add to cart for that product as an action. You won’t get the report that you’re talking about, but you’ll have all the data in the event reports, which should make it fairly easy to calculate.
Thanks again for reading,
Justin
Hi,
I just stumbled across your blog whilst searching for an answer to a particular question – seems very useful and thorough, I’ll be keeping an eye on it from now on! :)
However, my question is, is there any way to set an event tracking value as a Goal? e.g. if I have an action as ‘Click’, can that be one of my goals? This would be very useful as the Event tracking does not seem to be as ubiquitous as Goal reporting.
Thanks,
Tim.
Hi Tim,
Glad you like the blog and find it useful.
To answer your question, no, there is no way to make an event a goal. Goals are pageviews. And pageviews are fundamentally different data types in GA. They’re different than events.
My suggestion would be, if you need to simply count something, like a click, use an event. But if you need that click to be a conversion, use a pageview.
Thanks for the question,
Justin