Measuring A Page’s Effect On Conversion

How likely are people to call you once they’ve visited your website’s About page? How likely are people to purchase your product once they’ve read your most popular blog post?

For digital marketers and analysts, questions like these are important. Answering them can help us determine what value propositions are most important to site visitors and what page elements best communicate them. It can lead to smarter decisions regarding how to shape user paths, and it can inform our decisions about conversion points.

However, despite its importance, the question of page attribution (which is similar to channel attribution) is only partially answered by the standard reports in Google Analytics.

In this blog, I’ll explain 1) what reports come standard with Google Analytics and why they fall short for the analysis that we want to perform, and 2) the method I’ve begun to use to get an idea of a page’s effect on conversions.

Built-in reports for similar questions

The Channels report offers one method for partially gauging a page’s effect on conversion. In this report, after selecting a Default Channel Grouping, you’ll have the option of setting Landing Page as the Primary Dimension (if it isn’t already). With Landing Page as the Primary Dimension, you’ll be able to select your conversion of interest from the dropdown menu in the Conversions header (as long as you have it set up as a goal in your account) and see the rate of conversion for sessions landing on your page of interest.

Google Analytics

This method of measuring a page’s effect on conversion can be thought of as first-interaction attribution for pages, because it only associates a conversion with a page if the session in which the conversion took place began on the page. If the page you’re interested in typically isn’t a landing page, or if you’re interested in seeing the page’s effect on conversions for sessions where it wasn’t the landing page, this report doesn’t offer much help.

How about our other options?

The Goal URLs and Reverse Goal Paths reports in the Conversions tab provide other methods of measuring a page’s effect on conversion. However, like the Channels report, they don’t fully answer our question. These reports operate on the other end of the attribution spectrum, showing us how often a page was the last page- or one of the last pages- that a user visited on their way to converting.

Here’s how they work, and why they’re imperfect for our analysis.

The Goal URLs report tells us which page a user was on when they converted. It calls this page the Goal Completion Location. If your conversion point is a contact form, this is the page that will display in the report. If your conversion is a phone call, the report will show you which pages visitors were viewing when they decided to call.

Google Analytics

The problem with this kind of report is obvious; it’s the same as the Channels report’s issue, but in reverse- it only gives credit to the page in question if it was the site of the conversion. If you’re interested in how a page “upstream” of a goal completion location affects conversions, your only recourse is to use the report’s “Goal Previous Step” dimensions. These dimensions are displayed by default in the Reverse Goal Paths report.

Put simply, the Reverse Goal Paths report shows the last three pages visitors passed through on their way to your Goal Completion Location. It lists the paths in order of frequency, with the most common at the top and the least common at the bottom. Because this report gives data on up to four pages that were visited in every goal conversion path, it gives us an opportunity to capture more information about how a specified page affects conversion rates. To do so, you could export the report’s information and use some quick Excel functions to determine how often your page of interest shows up in the last four steps before conversion.

Google Analytics

This gets us closer to answering our question, but it still provides an incomplete picture if our page of interest sometimes falls outside of the last four pages in the conversion path. This is a concern especially if your site’s conversions typically take place on a contact form or after a series of checkout pages.

So, if we want to know how often a page makes up any part of a session’s path towards conversion, what are we to do?

Are you giving credit to *all* the pages in your site’s conversion funnel?

The workaround

Before diving into the solution, let’s return to our initial question. We were curious to know how likely someone was to convert if they visited a specified page on your site. The rate we were after, then, was the rate of conversion among sessions that visited the specified page- in other words, the number of converting sessions that visited the page in question divided by the total number of sessions that visited the page in question.

To find this rate, I use a simple but perhaps non-obvious solution. Here’s how to do it.

First, go to the “All Pages” report in Behavior > Site Content. Create a segment consisting of site visitors who made the conversion(s) that you’re interested in, and apply the segment (without removing the “All Users” segment). Notice that Google Analytics will limit your date range to 93 days from your start date, due to the Converters segment containing user data. Adjust your date range accordingly. Your report should now look something like this.

Google Analytics

Next, export the data to Excel or your statistical program of choice. Add a column to your dataset and populate it with the quotient of Converters’ pageviews divided by all pageviews for each page (or the same calculation, but with unique pageviews, if that’s your preference). This is the rate we’re after. Find your page of interest and compare its rate with the rate for other pages.

Excel Spreadsheet

Is the conversion rate of sessions visiting your page of interest relatively high? If so, the page may be an important part of a common conversion path within your site. It may show a value proposition that persuades visitors to convert, or it may communicate more effectively a value proposition that is found elsewhere on the site. Whatever the case, the beauty of our analysis is that we’ll be able to see these effects on conversion even if the page isn’t a landing page, and even if the conversion typically occurs after browsing several more pages.

Do you have a different way of finding a page’s effect on conversion rates? Have you spotted a flaw with my method? Have you thought of other benefits for this kind of analysis? Share your thoughts with me on Twitter (@MikeGardo)!