A persona is a composite portrait of a real customer type — made concrete enough to design and decide for. Alan Cooper introduced them so software teams would build for a specific someone instead of a vague “user.” A good persona captures who the customer is, the job they’re trying to do, their goals and frustrations, how they decide, and what they’re choosing between — a stand-in specific enough that “would this person buy it?” has a real answer.
The dividing line between a useful persona and a harmful one is evidence. Grounded in real interviews and behavior, a persona is a compression of what the business has learned about a segment. Invented in a meeting to match the product the team already wants to build, it’s a fiction that launders the team’s assumptions — worse than nothing, because it wears the costume of customer understanding.
An invented persona doesn’t represent the customer. It represents the team’s hopes, given a name and a stock photo.
Read personas as decision tools, not deliverables. The test isn’t whether the profile is richly written; it’s whether it changes a choice — would this person use this feature, respond to this message, pay this price? Keep the set small (two or three that matter, not a gallery), and keep the focus on goals, frustrations, and the job, not on demographic flavor that decides nothing.
Structurally, a persona is a segment made human. Segmentation does the analytical work of dividing the market by behavior and economics; the persona makes one chosen segment vivid enough to build and sell for. So it sits downstream of segmentation and Jobs-to-be-Done: those decide which customers and what job; the persona makes that decision concrete. Build personas before that analysis and you’re just sketching characters.
For a business serving genuinely different buyers, personas keep the differences from collapsing. The locksmith platform’s distress residential caller and its facilities-manager commercial buyer are not variations on one persona — different jobs, triggers, decision processes, and price sensitivity. Naming them as distinct personas is what stops a single message and a single offer from being aimed at an average customer that neither of them is.
Reach for it when a team is designing or marketing for a vague “user,” or when distinct customer types are being averaged into one. Be wary of personas built without evidence — they amplify assumptions. Pair it with Market Segmentation (the analysis a persona makes concrete) and Jobs-to-be-Done (the job at the persona’s core).