Personas aren't bad, and you're not a bad designer if you use them | February 20, 2019

With disturbing regularity, my Twitter stream seems to explode with posts demonising commonly used, yet seemingly harmless design tools. These posts take great pains to document all the ways these tools have failed people in the past (while downplaying or ignoring situation where they may have helped). Reading one of these threads you’d be forgiven for thinking these tools were actively harmful; some sort of public health nuisance this brave whistleblower was uncovering for the first time. Although they generally read more like the hysteria surrounding MMR vaccines to me, rather than any reasoned or measured debate. Instead they’re full of personal anecdotes and half truths, wrapped up in a thin veneer of scientific speak. One such debate I witnessed recently revolved around Personas.

Now the love/hate debate around personas is nothing new; in fact it’s been rumbling on for some time. I personally have no great love for personas—that would be a little like loving a hammer or a chisel—but I have no great hate for them either. Personas are tools like any others, with their own relative strengths and weaknesses. So while every few months a new tool comes along claiming the previous one is now dead–in this case Jobs to be Done—I’m much more interested in developing a rich toolbox of tools to draw upon, than engaging in a Battle Royal style deathmatch. That is to say, I’m not on team persona, team JTBD, or any other team that may be having its moment in the sun. Instead I’m on team human.

Whenever the vilification of persona’s pops up its ugly head, it always goes something like this. “Personas are bad because they’re usually made up and therefore have no empirical backing”, “Personas are bad because they try and segment people and people are too unique to be segmented”, “Segmenting people is bad on principal and tantamount to discrimination”, “Personas are just a bunch of demographics and are therefor completely useless”, “Even if personas aren’t just a bunch of demographics, they are so badly used that we should just stop using them”.

While all these arguments annoy the hell out of me, the last one is the worst. The idea that because something is badly used, we should ditch it, is a blatant call for ignorance. Personally I think if something is poorly or wrongly used, it’s our responsibility as experts, craftspeople and educators, to help people understand how a particular tool or technique works, rather than just throwing it–and the people using them—away like yesterdays garbage.

So let me address some of these other common misconceptions about personas. The first and most pernicious is the idea that personas are just a bunch of—often made up—demographics featuring a persons age, gender, job, hair colour and little else. Now if this was the main content of personas I could understand the frustration. However that’s just not the case.

I think this misconception comes from the world of marketing, where marketing personas generally are more focussed on demographics. That’s because marketing personas are generally used to try and capture, understand and categorise fairly broad ranges of preferences, and then understand where people holding those preferences spend their media browsing time, to better focus marketing spend. That’s all well and good for our friends over in marketing, but this sort of approach provides very little value when it comes to Interaction Design.

Instead, the use of demographics in personas is really there to add a bit of background detail, to make these pen portraits a little more realistic and memorable. This is because the true value of personas is a communication tool. As a way of synthesising a large amount of rich and complex data, generated through observations, interviews and surveys, into something that can travel around an organisation, and be consumed by people who weren’t necessarily involved in the conversation, or even have regulate access to customers. In these situations it’s helpful if the persona is recognisable as person, with a name, a background story and a set of attitudes and beliefs.

The question then is, if personas are a communication tool, what are they communicating? Well in general, the most useful part of any persona is the User Scenario; the area on the page that describe the sort of problems these users are facing, the behaviours they exhibit, and if you want to use that particular language, the jobs they want to get done.

I think this is one of the reasons why a lot of people claim that Job Stories and JTBD do away with the need for Personas, and I think in some situations that may be true. If you’re working in a fast moving engineering focussed start-up, servicing an incredibly wide demographic with a similar set of needs, and everybody in the company is very user centric and in touch with those needs, you can probably do away with the touchy-feely fluff and focus on the core jobs; it’s just a lot of unnecessary distraction.

However if you’re working in a traditional organisation serving a smaller range of fairly distinct behaviours, and you need to get your stakeholders out of the mindset that all customers think and behave the same, or worse, all customers think and behave the way the executive team do, personas may be of some use.

In some way you could argue that personas and a useful step on the journey towards JTBD, but if you’re already there, they probably hold little additional value. Personally I like the ability to reference John, Mary, Prisha, or Joaquin as a placeholder for a set of behaviours, rather that sifting through half a dozen job stories to explain what set of problems we’re solving for today. It’s efficient and and travels well, even if there is a certain loss of fidelity.

The best personas generally are based on research data. You’ve surveyed hundreds or thousands of customers, you’ve interviewed dozens more, you’ve crunched the data and found certain patterns. Maybe you’re a travel company and have noticed that parents and children tend to travel shorthaul for a week or two, over the school holidays? Maybe the data has surfaced another cluster of younger travellers who prefer cheap city breaks, anytime other than the school holidays. Maybe you discovered a third cluster of people who travel for business and generally (but not always) fly business class, use the lounge and prefer to be back home for the weekend. This is a fairly trivial example, but I hope you get my drift.

Knowing this information may allow different product teams to tailor different experiences for different clusters of behaviour. This obviously beats treating every customer as the same, but isn’t as good as building up a rich and individual picture of behaviour by acquiring tonnes of data about the user, and feeding it into a sophisticated CRM system. Many tech firms have these capabilities, which is another reason why they probably look down on personas and those who use them. However you’d be amazed how few traditional businesses are able to significantly alter the merchandising offering, let along the whole user journey, based on detailed data capture.

Now it’s worth pointing out that nobody is saying that you can’t be both a business traveller, a city breaker, and somebody who is looking for a fly-and-flop holiday with their family. Similarly while the Personas my suggest that city breakers tend to be younger than people traveling with families, the actual data shows considerable overlap. In fact you may find a cluster of retirees who take their grandchildren on holidays on their own, and also like city breaks.

Fortunately personas aren’t rigid definitions, but rather porous archetypes, so people will move between the two. Something this will happen over years. Other times it will happen in a single session. The key to any successful model is realising that it’s only a model, and therefore has limited scope. Sometimes it’s useful to think about light in terms of a wave, other times in terms or a partial. A good craftsperson understands the strengths and weaknesses of each model and uses it appropriately, while a poor craftsperson generally blames their tools.

Which comes on to my last point, the inevitable cargo-culting of any and all design tools. As a quick recap, cargo cults emerged amongst Pacific Islanders who tried adopting the object and rituals of western colonial power, without realising how things really worked. So they would build wood and reed airplanes, control towers and airport terminals, in the hope that American GIs would return with their cargo, without realising the intricacies of world politics, aviation, and international trade. They basically just copied what they saw other people doing in the hope it’d work.

I see this all the time in our industry. You read an article about personas and throw up some badly researched, demographic heavy documents. You then use them to prover the validity of your pet feature to a bunch of skeptical engineers and business people, and wonder why it failed. It couldn’t have been you, it must have been the tool, so you go to Twitter or Medium and vent your anger at the uselessness of Personas.

It’s frustrating to see people behaving this way, and if it was just themselves they were hurting, I wouldn’t mind as much. However when these things below up, the outrage and misinformation tends to travel much further than a reasoned argument about the pros and cons of a certain tool or technique ever could. The resulting blast radius is huge, and a lot of innocent people often get caught up in the ensuing chaos. After all, if you were a relatively newly minted designer and somebody you trust with 20k followers on Medium, tells you that the thing you were taught as school is no longer useful, and all the smart people are now using something else, what are you going to do? Well, you’re probably going to ditch one tool in favour of the other, and may go around telling others to do the same. This is how MMR scares and market panics start.

Fortunately we’re not dealing with anything nearly as serious here, and at least you’ll end up with a tool in your tool box. However as the saying goes, if all you have is a hammer, everything looks like a nail. Even if you spend most of your time hammering nails into blocks of wood for a living, I think it’s worth understanding what a screw looks like, and when to use it, rather than trying to hammer a screw into a block of wood once or twice and then claiming that screws are rubbish.

As a professional we all benefit from having a variety of tools at our disposal, instead of falling for the hype around silver bullets and one size fits all solutions. We’d also benefit from a bit more critical thinking in our industry. Arguments that go along the lines of “This thing hasn’t worked for me, so it cannot work for anybody” are super reactionary. They stifle intellectual curiosity, rather than encouraging people to expand their horizons and learn new things.

If you’ve discovered a new tool that works for you, by all means share what you think is great about it. Similarly if you’ve found legitimate weaknesses in existing tools, please share that information so we can all learn and grow. However we live in incredibly polarising times, and debate through opposition is super easy. Just remember that it isn’t always necessary to raise one thing up by knocking something else down. That it’s a large world out there and we benefit more for variety and choice than a limited set of (largely manufactured) binary opinions.

Comments (0)