The MAAD approach to personas: what UX can learn from marketing
I’m about to commit the biggest sin in UX — take the marketer side (scandalous!) and claim that we can actually borrow a thing or two from the ancient art of making people buy more.
As any UX’er, I’ve read through countless articles and opinion pieces on personas. My personal relationship with personas has been a rollercoaster — I’ve moved from loving them to avoiding them to seeing them as a superior form of empathy and back to hating. An infinite loop, really. And just by doing a quick search on Medium, I can see that I’m not alone.
Deciding where is it exactly that I stand on this topic wasn’t easy. I’ve worked with teams that had the personas printed out on the wall, knowing them by heart and starting every product-related conversation by referring to their names. Great, I thought — their value is undeniable. But I’ve also seen plenty of teams that created them once, only to have them gather dust somewhere in their Dropbox. I’ve seen personas used as a sales tool, as meaningless anecdotes for the stakeholders, and even as a way to make a keynote prettier.
Because they’re abstract, personas have been misunderstood and misused over the years. They get created and they often fail for a number of reasons. — Kim Flaherty, Nielsen Norman Group
Obviously, a persona is only as useful as you make it be. But how do you measure the usefulness of a persona upfront, how do you define the boundaries between a good and a bad persona? Moving towards a proven, repeatable and scalable set of assessment criteria can help eliminate the guessing game, and this is where marketing comes in.
When it comes to defining audience segments, there is a simple yet powerful framework of MAAD (I’m like 90% sure that’s a thing) that is used by the marketing folks. And the good news — it applies nicely to working with personas as well. Let’s have a closer look at what it entails for the world of UX.
MAAD = Measurable, Actionable, Addressable, Differentiated
Measurable
Personas must be measurable (at least to a certain extent) in that we should be able to understand how representative they are of the overall audience. Different personas will result in different design decisions, often even contradicting each other. If we’re not able to prioritize them based on their frequency among the real audience, we run the risk of designing for the wrong people.
Example:
For a new food delivery app, Mark is a persona that always wants to try new dishes and restaurants, and thus appreciates features that enable discoverability. Muriel is an older persona, she does not care much for experimentation, and prefers ordering the same dishes from the same restaurants — her preferred flow is thus one where she can easily repeat her orders from the past. If we’re able to estimate that Mark represents 70% of audience, and Muriel accounts for just 30% — great, decisions on the app flow have suddenly become much easier. Without the measurable aspect though, personas might be not just of limited use, but even misleading.
Actionable
Personas must be usable from a creative, design or UX perspective. They must provide enough information to guide the team in envisioning different flows or creative messages. Even if we throw in smaller details that are meant to build empathy, these should only be there if they can guide towards a clear decision.
Example:
Let’s assume we’re designing professional accounting software. Dogs and kids might seem like great details to build empathy, but they don’t really help us make decisions for the accounting software interface. They’re not actionable — and thus only distract us from our goals. Now, if we’d be designing a train schedule app, these details might actually be relevant — Mike is likely to travel with kids during the weekend, and Jane might want to be able to take her dog for her next train ride.
Accessible
The key variables linked to a persona must be not just specific, but also accessible in that we can easily identify people representative of those variables. This is especially important when it comes to testing. A certain variable will lead to a specific design decision (which often requires making an assumption on what will work best for that variable) — and if we can’t find people with that variable, we won’t be able to validate that the related design choice actually works.
Example:
Let’s say we’re designing a news portal. One of our personas, Tim, is satisfied with his quality of life (this is an actual variable I’ve seen once in a persona, I’m not even making that up). This takes us to the assumption that Tim is an optimistic person and would prefer seeing the positive news upfront, with all the bad stuff away from his sight. Kind of a big leap with this assumption, agreed, so how do we validate this? If we’d start looking for people who claim they’re “satisfied with their quality of life”, we’d probably discover quite quickly that it’s not that straightforward. The meaning of this variable itself is subjective, not to mention that we’d have to run psychological assessment tests on our participants, trying to determine their overall level of life satisfaction.
Differentiated
Personas must be easily distinguishable from each other, and react differently to different product solutions or brand messages.
Example:
If you’re going for the age factor alone, 18-year olds and 60-year olds are likely to be different enough in terms of lifestyles and preferences, while 18-year old and 20-year olds are likely to not be differentiated enough to warrant belonging to separate personas.
Obviously personas are rarely differentiated on age alone, but the same principle applies when we’re going more granular. Let’s say we’re redesigning an Instagram login flow: If Jane has 2 years of experience with Instagram, and Jack has 5 years experience, does it make them likely to have different needs when it comes to the login flow? If not, then they probably don’t need to be two separate personas.
I’ve found this MAAD framework to be very useful in working with personas, whether as an individual sanity check or as a way to guide the team effort.
Next time you’re debating whether a persona’s pet ownership history is a useful detail or a meaningless distraction — try running it through the MAAD criteria and see where they take you.