Persona

What it is

A persona is a semi-fictional, research-based archetype that represents a distinct segment of your customers, users, or stakeholders. It condenses patterns found across many real people into a single, memorable profile: a named character with goals, frustrations, behaviours, and decision criteria. A persona is not a demographic bucket and not a guess. It is built from evidence (interviews, surveys, analytics, support logs, sales calls) and kept deliberately concrete so teams can reason about a real human instead of an abstract "average user."

A persona typically includes:

• Context: role, environment, level of expertise
• Goals: what success looks like for this person
• Pain points: obstacles and frustrations they want removed
• Behaviours: how they research, buy, and use products
• Decision criteria: what tips them toward yes or no
• Watering holes: channels and sources they trust

Why it matters

Personas create a shared reference point across teams. When product, marketing, and sales all picture the same customer, decisions get faster and more consistent. Personas force explicit assumptions into the open, where they can be tested and corrected. They also help prioritise: features, messages, and offers can be scored against whether they serve a priority persona.

The main risk is fiction dressed as fact. A persona invented in a workshop without data will mislead. Good personas cite their evidence and are revisited as the market shifts.

How it is used in practice

• Marketing: shapes messaging, channel choice, and content themes.
• Product: prioritises features against real goals and pains.
• Sales: anticipates objections and tailors discovery questions.
• AI systems: guides prompt design, tone, and evaluation criteria for assistants and agents.

Teams usually maintain three to five active personas. More than that dilutes focus.

Worked example

"Compliance-first Clara," Head of Risk at a mid-size bank.

• Goal: adopt an AI tool without creating audit exposure.
• Frustration: vendors overstate accuracy and hide data handling.
• Behaviour: reads documentation before any demo, involves legal early.
• Decision criteria: data residency, explainability, human-in-the-loop controls.

Given Clara, a vendor would lead with a security whitepaper rather than a flashy feature tour, and route her to a solutions engineer, not a generic SDR. The same persona would tell an AI product team to expose confidence scores and citation trails by default.