In every company I’ve worked for or with, I’ll confess that I’ve often been a little confused about the tracking setup. I’ve seen plenty of reports that simply don’t make sense, but once these got sorted out, I usually found that we were putting developments live with only having a top level understanding of their effect – basically ‘improving’ a product based on conjecture. But speaking to quite a few experienced analysts led me to believe that this problem wasn’t just following me around – it was everywhere! So I learnt what I needed to do about it, and I thought I’d share that knowledge here.
Eliminating Vanity Metrics
The big metrics in publishing are Unique Users and Pageviews, but they are surprisingly meaningless unless the website sells out of advertising inventory, which often doesn’t occur. In e-commerce, the audience and engagement metrics can become even more meaningless – more traffic doesn’t always equal more revenue, and ‘more’ engagement via pageviews could mean a user is struggling to find what they came for. In SEO we are obsessed by the organic visit, yet it means little if the user doesn’t convert or add further value.
Thus it’s important that we move away from the Audience Overview and simply reporting ‘vanity metrics’ in our analysis, and focus on user actions that really mean something.
Website Goals
Most talks or posts about analytics will tell you to set up goals, and they are easy to setup in Google Analytics. It might be a given to put the receipt page as a goal on an e-commerce website, but which other goals are you tracking?
These goals should be considered as fundamentals alongside e-commerce receipts – they’ll inevitably help you as you make subtle website changes and improvements.
Assigning Value
At last year’s A4U Expo I saw a great talk about assigning value to all of your goals. This doesn’t have to be a literal £ value, but there needs to be a relative value assigned.
For instance, you might consider that newsletter capture is perhaps the most valuable of your marketing goals and assign it a value of 100 (although, if you also knew the e-commerce value of email traffic, you could assign it this value).
Then, working back, you could work out the relative value of your other goals. For instance, if your advertising clicks have 1% equity of newsletter clicks, then they could be assigned a value of 1.
The beauty of this approach is that you can measure your traffic sources goal conversion rate by Goal Set. So you could have an engagement Goal Set, or a commercial Goal Set – click these at the top of the traffic graph on All Traffic to see which traffic sources are giving you the most value. You can then spend more time on the traffic sources that matter.
Event Tracking on Steroids
Google’s suggests that you apply event tracking to the following:
- Any Flash-driven element, like a Flash website, or a Flash Movie player
- Embedded AJAX page elements
- Page gadgets
- File downloads
- Load times for data
Some of your goals will inevitably be some of the above, so it’s important to have these event tracked. As an aside tip, to track Youtube plays, use this JavaScript.
But we can take it much further than this; why settle at just key engagements when you can track every link? If you’re likely to be working on a site through a number of development cycles, then I seriously recommend this. Even on very large sites, it’s possible to event track everything relatively simply, by having a system for placing event tracking in links throughout the templates.
Fundamentally, Google Analytics Event tracking works through 3 components – the Category, Action and Label. The label is optional.
onclick=”_gaq.push([‘_trackEvent’, ‘category’, ‘action’, ‘opt_label’]);”
More detailed posts on event tracking are available here:
- Event Tracking 101 for Google Analytics (Search Engine Land)
- Google Analytics Event Tracking to Monitor Calls to Action (SeoMoz)
- Idiot’s Guide to Event Tracking (.eduGuru)
- Event Tracking Naming Strategy (SEER Interactive)
Think of these as a hierarchy. In navigation with dropdowns you could (and maybe we will) have the following for State of Search:
onclick=”_gaq.push([‘_trackEvent’, ‘Navigation’, ‘News’]);”
Then the homepage carousel might be:
onclick=”_gaq.push([‘_trackEvent’, ‘Homepage’, ‘Carousel’, ‘1’]);”
Or on the homepage top post you could have:
onclick=”_gaq.push([‘_trackEvent’, ‘Homepage’, ‘Posts’, ‘1’]);”
If you followed the same logic for all links throughout the site, with dynamic insertions using .PHP, tracking all the links wouldn’t take that long. I’ve had developers do practically an entire site in half a day! It’s not as big a task as it seems, but if you’re short on resource you can iterate it to include just core elements to begin with.
You would soon know which pieces of functionality are working best for you. And as you tinker with the design (as you should be doing in the AGILE age) then you will find it much easier to see what is and isn’t working for you.
Some interesting findings…
In applying as much click tracking as we could to a site, we were able to gather a few things that were critical to future developments:
- Not nearly as many people were clicking the nav as we first thought; very few visits included a sub nav click.
- Most clicks were coming through the homepage listings and right hand columns, but articles published earlier in the day would often get less pageviews.
The big problem here was that our homepage and right hander sizes just couldn’t cope with the volume of articles being published. Subsequently, articles were getting buried quickly. Also, our navigation clearly didn’t get information to the user very well. They weren’t that interested in sub-categorisation, just content.
Our Response: Extend everything! Create larger pages with more content options and create a mega nav.
The result: 25% increase in pageviews in one month! (In this case, it was a publishing site, so it wasn’t a vanity metric! H.T @JamesGurd)
Conclusion
You might think that this could be an overkill of data, and that having so many different events will make it troublesome to analyse your most important data. However, I can assure you that if you categorise your events correctly in a hierarchy, then you will be able to make sense of your data very easily. Most importantly, you’ll have a better understanding of where you should invest time and energy in the next development cycle.
