Anna Holopainen asked this question on LinkedIn:
What’s wrong with most buyer personas? Aka my favorite pet peeve 👇
She also provided an answer straight away, listing the following points. In the following, Anna’s points are shown in bold text, followed by my response in unbolded text. I found these points very fruitful for a serious discussion about the value of personas, and they are fairly common so it is worthwhile to address them. In my rebuttal to Anna’s points, I’m not trying to claim personas as a perfect vehicle for design or decision making, but rather to expand on possible remedies and fixes to the points Anna mentions.
► They focus too much on demographics or psychographics and ignore contextual motivations (how is the fact assumption that the buyer is “business-savvy” going to help you sell CRM?)
>> It is true that demographics play a large role in personas. In our research, we have observed that users place great importance on demographic factors and also the looks of the personas, which are factors that could often be considered irrelevant for understanding user needs. So, the rebuttal to this point contains two suggestions: (1) create personas without demographics (totally possible, albeit potentially limiting the perceived realism by the persona user); or (2) make sure that you do include heavy emphasis on information beyond demographics and standard psychographics. An interesting alternative is to portray your personas in a comic strip, in which the persona goes through various “real life” situations involving your product. Doing this would introduce flexibility and fluidity to communication of the persona. Another alternative is the use of interactive persona systems, such as the Automatic Persona Generation (APG), that enable interaction with personas’ information (e.g., filtering quotes by sentiment) and “simulating” the persona’s reactions to your content. Ultimately, the information design of personas should be thought case-by-case: what is the information your sellers need to sell CRM? That is the information that goes into the persona profiles.
► They typically assume that people with the same title (e.g. CTOs) are a homogenous mass, with the same motives and pain points — regardless of the context (e.g. a mature Fortune 500 company or a fresh hi-tech startup with novel technology).
>> If this is the case, then the created personas would not be what we call high-quality personas. Meaning, personas need to precisely separate one user type from another. If all CTOs would be identical on their information content, then the persona creator (or the algorithm creating the personas) has failed in their task. Nonetheless, this is an easy fix: create better personas, precisely ensuring that the motives and pain points have separation. If you are using external consultants or agencies for persona creation, make sure to communicate in your design brief that you want “Startup CTO buyers” and “Fortune 500 CTO buyers” that need to be different.
► They only focus on buyers, not users (in SaaS, we can’t afford to overlook expansion revenue, can we?)
>> I would argue that this point is somewhat a fallacy. In fact, the origin of personas, at least in Human-Computer Interaction and user-centered design, is precisely the portrayal of users (not buyers). Buyer personas, to which Anna refers, came up somewhat later, when marketing and business domains embraced personas. Nonetheless, I get the point – that buyers are not all that matters; the end-users of the product matter as well. But for this there is an easy fix, in fact the same as in the previous case. Just create personas of users, in addition, to buyers. In fact, this point is crucial. The person who is ordering the personas to be created needs to map out his/her information needs: hence, if it is important to understand not only buyers but other stakeholders in the value chain (such as end-users of the product), this requirement must be made explicit before starting the persona creation. In SaaS, for example, there buyers and users, certainly, but even the users can be further segmented into active and inactive users. Hence, understanding the business requirements, and communicating those requirements to persona creators, are crucial activities to undertake. But, there is absolutely no reason why personas should focus only on buyers.
► They only focus on, well, personas and not segments, which is an obvious problem in B2B
>> This point is slightly unclear to me. Personas are very similar to user/customer segments; in fact, I often say that personas are segments with a name and face. Conceptually, personas give a face to user segments; they personify nameless and faceless user segments. This means they do contain the same information as segments presented in graphs, tables, or other ways. So, this point escapes me.
► They only focus on the current target audience, not adjacent customers.
Personas created algorithmically can benefit from lookalike-creation algorithms, much like lookalike audiences in platforms like Facebook Ads. Moreover, data-driven personas can be traced back to individual or aggregated user data to e.g. form marketing email lists. Finally, by relaxing/tightening hyperparameters, it is possible to create personas that cover more/less of the baseline user data. So, there are ways of making personas contain different subsets of customers. Nonetheless, the crux of the point Anna is making might be that “personas are married to their data.” Meaning, whether using data-driven algorithmic methods or qualitative methods like interviews, we cannot extrapolate beyond the data at hand. Meaning, we cannot create personas for a startup company that does not have any data on their customer yet! This is a valid concern. The same concern applies to creating personas of competitors’ customers – we don’t have access to their proprietary data, so personas cannot be created. Yet, I’d still argue this matter can be resolved with the use of creativity. Meaning, in the case of the startup, we can still interview prospective customers or even invite to user testing – the resulting material can be used for persona creation. Even in the case of rival companies, we can find online reviews, feedback and complaints that can help us understand the type of customers the competitor has. So, I would not say the lack of adjacent customers is a property of personas per se; rather, it’s a shortcoming or lack of creativity in their application.
► They’re typically pulled out of thin air as a result of an internal brainstorming session (or superficial interviews)
>> This point is obviously from the handbook of “How to Create Bad Personas” 🙂 Seriously though, this point is a major concern. “Garbage in, Garbage Out” means that imaginary personas would play zero importance in actual decision making – nobody should take such personas seriously or base any choices on their information. However, “typically” does not mean “always” — what I mean is that this is a malpractice of personas. It is similar to somebody inventing numbers to a profit/loss statement. We wouldn’t say profit/loss statements are useless because they can be abused, right? The same is true for personas – just because they are done wrong, we cannot say they are wrong. To support quality persona creation, we have published a lot of resources, including a book on Data-Driven Personas, over 35 research articles on persona design or generation, and more than 150 blog posts addressing various myths and misconceptions about personas, while offering solutions to those issues. These resources have been created in the course of our research and persona system development over the past five years.
In conclusion, personas are not perfect because the people creating and using them are not perfect. Personas come from a qualitative research tradition that has been less preoccupied by scientific standards and quality than the digital analytics field, for example. This is why the reputation of personas is worse than analytics. Yet, there is absolutely no fundamental reason as to why personas would lose to numbers. The many benefits of personas, relating to immersion, empathy, and communication, can triump over cold numbers, graphs, and tables. There is also scientific evidence showing personas outperforming Web analytics systems for specific tasks such as user segment identification.
With this, I thank you for reading this post and wish you a productive day working with customers, personas, and data!