Once upon a time, Adobe Analytics had a metric that tracked the average page load time but, then suddenly, it was gone. Wouldn’t it be great to have this back in the UI so you can slice and dice data based on it?
The thing to remember with page load time is that, as the load time for your site goes up, the bounce will increase while conversions from that page decreases. This is terrible for organizations and needs to be addressed ASAP. So, how do we fix that within Adobe Analytics?
The first thing you will need to do is leverage the getLoadTime plugin. As the name suggests, this plugin captures load times in a variable, such as prop or eVar depending on your set up. The load time is recorded in tenths of a second with no decimal. So, 11 is 1.1 seconds, 15 is 1.5 seconds, 19 is 1.9 seconds, and so on. Keep that in mind as Adobe records the values.
Next, you set up regular expressions to handle, which makes the report a lot easier to analyze. Here are some suggestions for how it can be bucketed:
Less than 1 second
1 to 2 seconds
2 to 3 seconds
3 to 4 seconds
4 to 5 seconds
5 to 10 seconds
More than 10 seconds
Regular Expressions Needed:
\b(0-9)\b
\b(1[0-9] | 20)\b
\b2[0-9]\b
\b3[0-9]\b
\b4[0-9]\b
\b[5-9][0-9]\b
^[1-9]\d\d\d*$
Reporting on load time via intervals (1-2 seconds, 4-5 seconds, etc.) is directional and not ideal for reporting out site speed. It’s hard to optimize this. So, if you look at browser load times, wouldn’t it be nice to have just one number to measure this, as opposed to a bunch of new data points created?
Well, now you can take the page load time received and convert it into a calculated metric to report. To set this up, you will first need to create a segment for each second.
Within the segment, you want to have page load times equal 1 all the way to 150. This will guarantee that you have most of your site’s total load time factored in. This means you have 150 segments — with a segment for each second ranging from one-tenth of a second to 15 seconds.
Next, you multiply each segment you just created by their respective amount, then divide by the total page views within a calculated metric. For example:
Page Load Time = 1 is Page Views x 0.1
plus
Page Load Time = 2 is Page Views x 0.2
plus
Page Load Time = 3 is Page Views x 0.3
You repeat this equation until page load time = 150. This will cover all of your bases from the segments setup. You then need to divide this number by total page views. This can take several hours to bake so give yourself time.
Once completed, you have now converted page load time from a variable value to a calculated metric that will reveal average load times. From here you can set up reports including average load time by browser, by mobile device, etc.
That’s it! Now you have average page load time again in your reports and can leverage it whatever way you wish.