A key component of web governance is an ongoing programme of monitoring site analytics. The question is; what do you monitor if you are not an ecommerce site?
You’ve got an information rich site and you make some changes to it. How do you know whether you are heading in the right direction with your changes if the site isn’t an ecommerce site?
The great Avinash Kaushik has some good suggestions in his article “I Got No Ecommerce. How Do I Measure Success?”
Let’s look at some of Avinash’s ideas in practice for an information rich site.
Establishing measures of success
First we need to think about some measures of success. From a business perspective we want to understand whether the changes we’ve made are improving the level of engagement of visitors with content.
If we look at the behaviour of engaged visitors on the new site compared with the old site we might hope to see the following.
- A higher percentage of all visits being by engaged visitors.
- A higher percentage of all page views being made by engaged visitors.
- Engaged visitors spending longer on each page because our content is so mesmerisingly wonderful.
There are all sorts of other behaviours you might want to see but that’s plenty for now.
What is an engaged visitor?
It therefore makes sense to come up with some definitions of engaged visitors. Once we’ve defined engaged visitors we can set up segments for them in Google Analytics and analyse what they do.
Here are a few possibilities.
- Duration of visit. The site is content rich, so it makes sense that genuinely engaged visitors are going to spend a reasonable amount of time on it. Let’s say that one definition of an engaged should be visits that last longer than 180 seconds.
- Number of pages viewed in a session – ‘depth’ as Google Analytics calls it. Of course, some users may find exactly what they want directly from a search engine. However, most truly engaged users will move around the site, view some high level pages and some deep pages. Let’s use another definition of an engaged visitor as one who views more than 3 pages in a session.
- Number of visits. An engaged visitor should be one who visits the site more than once, a lot more than once. Let’s use another definition of an engaged visitor as one who has been to the site more than 3 times.
The thresholds (in this case 180 seconds, 3 pages and 3 visits) might be different for you. It doesn’t really matter what they are as long as they are high enough to be a pretty good indicator of genuine engagement. If we need to we can do some combining. The key thing is that we’ve now got some definitions of engagement.
Setting up Google Analytics
How do we set them up in Google Analytics? We use Advanced Segments. Here are the segments we’ll need.
So now we’ve got our set of engaged visitor segments, what do we do with them?
Compare and contrast
The first thing to do is to set up comparable time windows. For example, for reasons best known to myself lets say we are interested in comparing what our so-called engaged users are doing in 1 – 30 November 2012 with the same period in August 2012. Google Analytics allows you to set a range and a “compare to” range.
Start off without selecting any segments. Now we start to see some interesting information.
Firstly, we can see that visits have gone up by over 23% and visitors by over 19%.
That’s good, but what about our engaged visitors? We want to see more engaged visitors. Let’s switch on the three segments we just set up.
Here we see increases in the proportions of engaged visitors. This is also good news. However, the increase is on a pretty low baseline of just over 8% of all visits being more than 180 seconds for example.
So we still have work to do.
Let’s take a look at how much content our engaged visitors are consuming. Here are two ways we can do this.
- Page views per visit. We’d like to see this going up.
- Time spent on each page. Assuming our articles are about the same length on average in August and November we would expect the average time per page to increase as engagement with our content increases.
Here’s what page views show us.
So good, but not quite as good as what we saw for visits. For some reason the proportion of all page views made in visits over 180 seconds has dropped a little. No reason to panic, but worth watching in future months. The other measures are still pointing in a good direction.
Finally, let’s look at what is happening to the average time spent on a page by engaged users.
Ouch! So our lovely engaged users are spending significantly less time on each of our lovingly written pages. Perhaps our articles are getting shorter and more punchy – in which case fine. Or perhaps our users are starting to lose interest in our content.
We need to find out more. Time to put together a more complete programme of monitoring. But, that can wait until next week.