Heatmaps and Web Analytics: A suitable match?

– posted September 21st, 2006 by Laurence Veale Comments (5)

Peter Knight recently posted about heatmaps. Here’s one we prepared earlier.

A heatmap showing clicks on our homepage

What can you infer from the above heatmap?

Generally, heatmaps are the end result of eyetracking, not clicks.

You can easily tell where people are clicking through your web analytics strategy. However, as Peter Knight has mentioned, “heatmapping” can be great for “re-representing existing data in another format”.

Web analytics tell you what, heatmaps tell you why

For me, heatmaps could be a perfect complement to web analytics. Your web analytics data can tell you the “what”, i.e. what your visitors click or don’t click on but not necessarily the “why”. Heatmaps may help you on the road to understanding “why”, why certain links are clicked on and others are not. I’m hardly sticking my neck out here but I wouldn’t be surprised if Google purchased Crazy Egg and incorporate heatmaps into their Google Analytics product. You heard it here first!

Like web analytics, heatmaps should lead to action

It’s worth mentioning that heat maps, like web analytics are of no use unless you can use the data from them to change your website for the better.

Digg this post Submit to del.icio.us Submit to technorati

Categories Web analytics

5 comments so far

1. Paul Adams on Sep 21st, 2006 - 12:30

Hi Laurence,

Why do you think heatmaps tell you ‘why’ i.e. the user’s motivation?

I would argue that neither analytics or heatmaps tell you the ‘why’. Analytics is a valuable quantitative tool to compliment qualititative design research. Especially when communicating to board level executives.

But I think that the ‘why’ can only be understood through conversation with; and direct observation of; the user.

2. Laurence Veale on Sep 21st, 2006 - 12:51

Hi Paul,

I totally agree with you, there is no substitute for direct observation of the user.

When talking about “why”, I was coming from the perspective that in visualising the data, you could postualte theories like “this link probably wasn’t clicked on because it has low visual priority or it looks like an advertisement etc.”.

You can then put the theory to the test through observation.

You’ve also made a great point about the value of heatmaps as a communication tool.

3. Hiten Shah on Sep 22nd, 2006 - 07:46

Laurence,

We think heatmaps are a great complement to web analytics, just as you mentioned. We are working hard to try to pull out more of the “why” from the data we collect and the visualizations we provide. There is still much more that needs to be done in that area. Thanks for the writeup and great discussion.

4. Brian Donohue on Sep 25th, 2006 - 09:18

I agree with Paul — there’s no “why” that comes at all from heatmaps, or web analytics. The “why” comes from the analyst figuring out how to interpret the data — this is the critical piece that makes or breaks web analytics.

Actually, I think it’s misleading to even separate heat maps from web analytics. They’re exactly the same thing. In Google Analytics, you use the Site Overlay feature to get the same information as a heat map tells you. The difference is that heat map presents the information in a more immediately intuitive and attractive manner (shading of colour versus a little bar graph). This indeed can be valuable, but shouldn’t be construed as providing a different layer or type of information.

I think it’s also worth pointing out that though direct user observation can lead to more clear-cut insights as to why things are happening on your site, web analytics have the unique ability to average thousands of users over days, weeks, or months. Unlike user-testing, there’s no debate about whether the data is representative or not. Web analytics provide real information about _all_ of your actual users.

5. Paul Adams on Sep 26th, 2006 - 19:40

Laurence: “You’ve also made a great point about the value of heatmaps as a communication tool.”

Actually, I have never used heatmaps as a communication tool! I think their value is way over-rated and I usually don’t find them particularly insightful. There are way too many unknowns.

I do however value quantitative analytics as a useful tool. As Brian mentions, it tells you something about *actual* user behaviour.

For example: I can use analytics to inform a CEO that over the past 6 months, 60% of the people who attempted to register on his site, didn’t complete registration. These people were motivated to use the CEO’s service, but somewhere along the line, the effort outweighed the benefit and the potential customer was lost. So analytics can give us the worrying figures, but it can’t tell us *why* people dropped out. It can’t tell us what we need to fix. Only direct user observation and/or discussion can pull that motivational data out.

We can use analytics to highlight surface level problems. We can use it to get buy-in from senior management to undertake a user centred design process. But I usually don’t use it for much more than that.