Low Funnel Conversion Rates
If you’re using Google Analytics you’re probably measuring your funnel conversion rate wrongly. Really. But I’m about to tell you why and what you can do to get better quality information from it.
Does your goal funnel have just one route users can take? Are you sure? They’ll never go back, refresh pages, duck out to add another item to their cart (starting them again in the funnel from the beginning), change delivery details, decide to give you a different email address, etc? The simple fact is, everybody browses the web differently. We all look at different things on a website and behave in different ways.
On your website, you might have envisaged the only logical route through to completing your goal. But your logic is somebody else’s confusing layout. People on a site click on lots of things, whatever they might think is a link. They won’t look at the entire page before they decide where to go next, they’ll scan it and click on the first thing they can find that looks plausible.
For instance, the screenshot below is a heatmap showing where users click on our homepage. The links are in the tabs at the top and the two titles on the right hand side. But a lot of people click on the man’s face at the bottom, expecting a link, or for him to do something.
The result of this, is that users are probably not going to do what you think they’re going to do, or what you think is logical.
How does this apply to your funnel conversions?
Well when you look at your funnel in Google Analytics it will give you a list of what pages people left the funnel to at each stage. You’ll almost certainly find these are a combination of exits (people who have changed their mind (for now at least) and left the site), page refreshes (for instance, updating delivery details might refresh the page), users splitting between a new user account creation and an existing user sign-in, people going back to the previous page to change something, people continuing to shop, etc etc. Of these, a smaller amount of people are actually getting out of the buying process than you think.
What this means is that when you view your funnel, you are probably seeing many more pageviews of the early pages than are actually occurring. People revisit these pages multiple times for a whole range of reasons. So many reasons in fact, that even if each one happens with just one person, your numbers for the initial stages of your funnel may be inflated many times over.
But there is another way to view this. When you look at the report for your funnel it gives you an idea of navigation paths about where people are going next. Starting from the final stage of your funnel (not the goal itself) work backwards looking at the navigation summary (available in the content report) for for each stage looking at where people go next. Portion these pages out into three categories:
- Exiting the funnel - made up of exits, going back to the homepage, and various other parts of your site that are no longer part of the goal process, be that shopping, registering, etc.
- Staying within the funnel - this is people who go somewhere else, but are still clearly interested in continuing to shop (or engage in whatever manner is relevant). They could be looking at delivery options, they could have followed a “continue shopping” link, they could have gone back to the previous stage to change a detail. Any people who haven’t progressed but are still part of your buying process fall into this category.
- Progressing along the funnel - people who move on to the next stage as expected.
Once you’ve assigned visitors to these categories for your final step, add up portions 3 and 1. These are your real numbers of people who entered that stage of the funnel. The assumption I’m using here (which I stress isn’t perfect, but is an improvement on the normal funnel functionality) is that people who are just moving laterally or backwards are still in your funnel. They haven’t left yet, they’ve just done something you didn’t expect or plan for. So these people should not be counted twice. Eventually they will go on to perform one of actions 1 or 3.
So the number you have now represents more accurately the number of users who have really entered that stage of the funnel. Using that figure as the number of people leaving the previous stage of the funnel, perform the exercise again for the prior stage.
And so on.
Eventually you’ll reach the beginning with a number of people you think have actually entered the first stage of your funnel, having removed the major causes of double counting.
In the image above, you can see the main exit sources from this stage of the funnel. The biggest one is /baskethold.aspx, which is actually this same page. So those are just refreshes, likely from people who have updated quantities or similar. The others include some specific product pages (likely from recommended products links) and the page users are taken to from the “continue shopping” button. Clearly these people should not be counted as exits, but as people still in the funnel. So on top of the 255 who progressed, only 23 went somewhere we can consider an exit. So the true number of people (not pageviews) of this funnel stage is more likely to be in the region of 278. So instead of the 58% progression rate, we’ve actually got more like 92%, which is more what we would expect from this stage of a sales funnel.
Now take your goal completions, divide it by your new initial figure, and voila! Your new funnel conversion rate.
You can take this a step further, and start looking at reasons people might be exiting once they have moved laterally. For instance if a lot of users leave your shipping page to look at the delivery options explanations page, then leave the site without going back to the shipping page (a double count that we want to avoid!) then you need to look at why they don’t go back. It could be that there is a condition in there that is putting them off.
This is a fairly advanced stage of analysis, but you can use the navigation summary in the content report to determine things like this. The users who leave your funnel but don’t come back… what happened to them? What changed their minds?
As far as I’m concerned, that’s the interesting piece of data your funnel can give you.
But don’t expect it to be easy. Google Analytics works hard to make things as simple to digest as possible, but if you start to rely on that instead of doing your own analysis you may not see the information that’s of most use to you.
If you’re interested in learning more about Google Analytics or heatmapping, give us a call on 0207 234 0500.