Category: UX advice / Research articles

Your website’s persona – how to quickly find out with Google Analytics data

It can be hard to pull together a decent set of personas on your users, but often all you need is a single statement about your them to guide your design. This is something you can find for any website in just 10 minutes.

Full personas

Personas are a design tool for summing up classes or archetypes of user and are a good method of keeping focused on your users as you design. They give you people to design for, which helps you avoid making assumptions about who your users are or falling into the trap of making the product for yourself.

However personas take a while to do properly. An example of how personas would traditionally be created:

A company carries out interviews with 20 users of their service and from the answers people give, discovers that they fall into four different types: super users, regular users, occasional users, absolute beginners. They then create four fictional individual personas to illustrate this, giving them names and characteristics that are an amalgamation of the real people they interviewed. They also give them motivations and pains that match real stories they’ve heard from talking to these people.

They’re a good output to summarise a full set of research and user interviews. However you may not have the time or budget to do this work. You might just need a simple steer on who is using your website.

Outline personas

I used to try and create multiple ‘outline’ personas from website analytics data but grouping users can muddy the waters. Turning a bunch of data into several individuals could take us further away from who are users are rather than closer.

For example we could state that with a gender split of 66% male, 34% female and a device split of 60% mobile, 30% desktop, and 10% tablet, we have three personas: 2 mobile using males, and 1 desktop using female. However this suggests that all of our desktop users are female and mobile users are male, which won’t be the case. It actually gets in the way of seeing reality.

Also whilst it gives you the where and who of the person it’s obviously not a meaningful picture. It’s not going to be as good as going out and doing user research and talking to actual people. It will be missing are the key insights into ‘why’ they do what they do.

It ends up being caught in the uncanny valley: they look like personas but aren’t. This can create more arguments than agreements.

A user statement

So lets not try and get too clever here and lets avoid the traditional idea of personas altogether. Instead we can be much simpler and high level. In my experience, nine times out of ten your analytics data will point towards a single user ahead of all others, so we can use that as our guide.

Lets answer the question: if our users were a person who would they be? State who that ‘person’ is and check if it matches who we want our users to be. I call this a user statement.

You may hear designers talk about another type of persona: assumption personas. These are really just a guess at who your users are, based on previous knowledge of a certain sector or business. It’s a way of trying to be user-focussed in your thinking but not based in data. We can quickly be a bit more informed than this.

Why would we do this?

There are two main use cases for defining a user statement:

How to find out who your users are

Google Analytics contains a few key bits of audience data that I use for constructing a picture of users. Here are how you find them in the software:

  1. Country. To find this go to Audience > Geo > Location
  2. Gender. Go to Audience > Demographics > Gender (check you have at least 20% of your audience included for a good sample size)
  3. Age. Go to Audience > Demographics > Age (check you have at least 20% of your audience included for a good sample size)
  4. Device. Go to Audience > Mobile > Overview
  5. Operating system. Go to Audience > Technology > Browser & OS > Operating System. To find the most popular desktop one add the segment ‘Desktop traffic’ at the top. To find the most popular mobile one add the segment ‘Mobile traffic’ at the top.
  6. Coming from. Go to Acquisition > All Traffic > Channels (or Treemaps for a nice visualisation)

Lets test this with some real data. Here are three sets of real but anonymised Google Analytics data, which I have access to. From each one I can fill in the following statement:

“Our typical user is a [age] year old [man/woman] from [place] on a [device & type], who has found the site by [where they came from].”

Website 1

Our typical user is a 26 year old [the most popular bracket is 25-34, but the 18-24 group is close behind, so the age should skew towards the younger end of the top bracket] woman from the USA on a Windows desktop, who has found the site by organically searching Google.

Website 2

Our typical user is a 32 year old person [as the split is so close, our user could be male or female] from the UK on an Android phone, who has found the site by organically searching Google.

Website 3

Our typical user is a 31 year old woman from the UK on an iPhone, who has found the site by clicking on a PPC link on Google.

Check the statement

Now we can ask ourselves a few questions:

Sometimes the results are bang on, sometimes it’s a bit of a surprise, sometimes it doesn’t really matter but it’s useful to know who you’re actually speaking to. Either way you now have a starting point, a target to move one way or another. This can be very useful for helping you position a redesign to appeal to a different group.

If you don’t have the time, budget or the instruction of a client to produce more in-depth user research, simply making a user statement is a great starting point for being user-led in your thinking. Now you have someone to keep in mind when you talk about who your users are and what they might want.

If possible I thoroughly recommend at some point going out and speaking to users, to interview them and find out what motivates them. The insights are often incredible and lead to lots of ideas for features and how to position a brand.

Last updated on 24 October 2019
14 tools to gather evidence – guide

Articles on similar topics

Mistakes to avoid when doing phone interviews with users and customers

How to do quick guerrilla user tests on your own

How to write remote user testing tasks (that don’t bias your results)