Category: UX advice / Research articles

Quick and cheap personas from Google Analytics data

Getting knowledge of who your users are when you design or market a website is an important step. You may not always have time or budget for user research but I've found data from things like Google Analytics can come to the rescue.

When it comes to assessing the likes of Google Analytics, I see there being two main types of quantitative data: performance data and user data. Each of them can be used in different ways.

You're going to want to watch performance data like conversion rates regularly and see how they change. For this I suggest creating a dashboard (for more on that see this article). However user data that tells you who your users are is something you only need to visit occasionally. Generally this is less likely to change suddenly. When working with clients I gather this type of data at the start of a project so I have a clear idea of who I'm designing for.

I then use this data to create outline personas. These help me keep the site users in mind at all times and enable me to judge if I’m designing the right thing.

An outline persona

For the uninitiated, personas are ways of summing up classes or archetypes of user and are a good method of keeping focused on your users as you design. They give you someone to design for rather than making assumptions and just making your product for yourself.

It’s good for showing the different groups of user you have too. Often it isn’t just one type, even though it’s easy to fall into the trap of talking about your users as if they are all the same.

What you can create with user data is what I call an outline persona. It gives you the what, where and who of the person but not a fully fleshed out picture. It’s not going to be as good as going out and doing user research and talking to actual people but it is certainly better than nothing. The main thing it will be missing is the insights into ‘why’ they do what they do but it will give you a solid starting point. By making a mark in the ground, it’s something that can be challenged or validated later.

How I do it

I’m a fan of keeping personas minimal and not filling them with lots of irrelevant stuff: keep it focused on things that will help you design. The kind of data I’m looking for from Google Analytics is age, gender, device, location, source of traffic, new vs returning, and whether they convert or not.

Let's say you look at your data and see the following:

I would take a look at this first row and think that summarises quite nicely into 4 personas: two would be American, one British, one Spanish (chosen from the biggest represented nation). Then moving onto the second row I would make three of those personas women and one a man (lets say one of the Americans). Next up using the split suggested by the data, I would make two of my personas desktop users, one a tablet user and one a mobile user (I could then look further in each category to fill in if they are a Windows/Mac or iOS/Android user).

Following the process for the rest of the data I could pretty quickly have four personas with six or seven useful pieces of information about them, including picking appropriate representative ages for them. I find Google Analytics data helps summarise the technological behaviour of the users. Learning which devices are most likely to convert on your goals helps define what is needed in the user flow for each. Also the source data gives you clues about what type of person they are and likely how knowledgeable they are likely to be on the company (article referrals = more informed, Google search = less so).

If you can then combine this with any external data you may have, like support emails and calls you can flesh the personas out a bit more and validate them. When doing this with clients, it’s handy if they chip in at this stage and inform you whether or not they think they sound believable.

If you don’t have the time, budget or the instruction of a client to produce more in-depth user research, outline personas are a great starting point for being user-led in your thinking. Whilst not as good as getting to speak to lots of customers, they’re much better than assumption personas or guesswork, as it’s real data from your actual users.

For a step-by-step guide to how to do this, sign up for my upcoming book on Data-Driven UX Design, which will explain this technique and more in depth.

Last updated on 10 June 2016

users / user need / ux / personas / google analytics / metrics / data /

14 tools to gather evidence – guide

Articles on similar topics

The 6 step process for running free remote desktop user tests

Build your own website dashboard (for free)