Best for
- Teams whose personas are demographic fiction nobody consults
- Compressing research into a form product decisions can actually use
- Aligning everyone on who this is for — specifically
What you give it
- Your evidence: interviews, support themes, usage patterns, sales notes — whatever exists (gaps get named, not invented)
What you get back
- Personas built on behavioural segmentation: what they are trying to do, how they decide, what constrains them — each claim traced to evidence
- The decision-relevant layer: what each persona needs to believe to buy, what makes them churn, how they prefer to learn — the parts teams actually consult
- The honesty markers: which attributes are evidenced, which are hypotheses — and the anti-persona: who this is NOT for, stated
How it works
- Segments by behaviour, not demographics: how they decide, what they are trying to achieve, what constrains them — the differences that change what you build.
- Traces every attribute to evidence: each claim footnoted to its interviews or data; the number of personas follows the evidence, not the template.
- Includes only decision-consuming attributes: if no product, marketing or sales decision reads a field, the field goes — the stock-photo layer is where credibility dies.
- Marks the epistemics: evidenced versus hypothesis, with the anti-persona stated — who this is not for is half the alignment value.
Example
You: Turn our 18 interviews and support history into personas for the product team.
Result: Two personas, not five — the evidence supported two genuinely different behavioural patterns (the hands-on owner who configures everything herself and decides alone in an evening, versus the delegating manager who buys what his accountant endorses and never opens settings). Each: goals, decision style, constraints and churn triggers — every claim footnoted to interviews or support data; age, hobbies and stock photos omitted on purpose (no decision consumes them). Plus the anti-persona (enterprise teams — politely out of scope, with the two sales near-misses as evidence) and three marked hypotheses awaiting the next research round. The product team quotes them weekly, which is the test personas usually fail.
Limits — please read
- Thin evidence makes thin personas; it says 'the data supports one pattern, not four' rather than decorating.
- Personas compress; the underlying research stays linked for the questions compression loses.
- They age: usage shifts and markets move — the review trigger is part of the deliverable.