Skip to main content
Back to Glossary (P)
Glossary · P

Persona.

A persona is a condensed model of a user group: a fictional person derived from real data, with a goal, a task, prior knowledge and obstacles. It is used to test design decisions against a concrete use, instead of measuring them against an abstract 'all customers'.

Illustration of "Persona": Persona card with goals, frustrations and a quote Sandra, 46Owner of a fiduciary firm, WinterthurGOALSMore suitable clientsLittle time spentFRUSTRATIONSAgency jargonUnclear costs"I want to know what it costs and what it brings."

Persona — Explained in Detail

A useful persona contains what influences decisions and leaves out what only adds colour. Useful elements are: the trigger that starts the search, the goal, the concrete task, prior knowledge of the subject, the environment including the device, past obstacles, the decision criteria and a verbatim quote from a real conversation. Invented details about hobbies, car brand and favourite drink are not useful. They create an appearance of closeness but drive no decision: no layout, no text and no order of steps gets better because of them.

The term goes back to Alan Cooper, who popularised it in software development in the late 1990s. There are two kinds. A research-based persona is built from interviews, observations, analyses of enquiries and usage data; every statement can be traced back to a source. A proto-persona is created in a few hours from a team's existing knowledge and deliberately records what is assumed and not yet evidenced. Both are legitimate; it only becomes a problem when an assumption is later quoted as if it were a finding.

A use case: a trade business serves two very different groups. The first is the private homeowner who buys such a service once in her life, knows little of the jargon, is looking for orientation and a price range, and builds trust through references and photos. The second is the buyer at a general contractor who awards twenty jobs a year, thinks in line items and deadlines, and wants to see capacity, references in commercial construction and a contact person within two minutes. These two models lead directly to two entry pages with a different order, not one page with duplicated content.

The criticism of the tool is justified and belongs with it. Personas can reinforce stereotypes when age, gender and origin are described without relevance, and they can legitimise decisions instead of testing them: 'Ms M. wouldn't want that' ends discussions without proving anything. A strict rule has proven to be the antidote: every statement in a persona points to its source, and anything without a source is marked as an assumption. Personas also do not replace testing with real people. How the data basis is created is described in the entry User Research; the result is checked with usability testing; and the models are translated into concrete sequences in the user flow.

Related Page

User Research

Frequently Asked Questions About Persona

Usually two or three, in exceptional cases four. What matters is not covering the whole target group but being able to tell them apart: two personas are only justified if they lead to different decisions, such as a different order, different content or a different entry page. If they lead to the same result, they are really one. More than four models are no longer used in everyday work and end up in a drawer.

A target group describes a segment by characteristics such as industry, size or region and is used for looking at the market. A persona describes a single use with a goal, task and obstacles and is used for design. The two are not mutually exclusive: several personas can occur within one target group, such as the person who decides and the person who uses, with completely different questions.

Yes, with a caveat. Enquiries in the inbox, notes from sales conversations, queries from Search Console and the internal site search provide a lot of material at no extra cost. Five phone calls of twenty minutes each with existing customers noticeably round out the picture. Labelling remains important: whatever comes from data is referenced; whatever comes from experience is kept as an assumption and tested later.

Ready for Your Project?

Apply this knowledge to your website — DLM Digital will help you.