Key takeaways
A marketing persona is the semi-fictional portrait of a typical customer, built from real data: interviews, customer history, search queries. A target defines a segment through criteria. A persona describes one person in that segment: their goals, their obstacles, the questions they ask before buying.
- Plural in French: “des personas” follows the 1990 French spelling reform; “des personae”, the Latin plural, is still common, but “un personae” is wrong.
- Three types: proto-persona (team assumptions), qualitative persona (interviews), statistical persona (large-sample survey), according to the Nielsen Norman Group (NN/g).
- Method: gather data, spot shared patterns, write a profile linked to the queries people actually type. NN/g advises starting with 5 to 30 interviews.
What is a marketing persona?
A marketing persona is a fictional person who represents a group of real customers: same needs, same obstacles, same way of deciding.
Definition
Persona: a representation of a typical customer, given a first name and a context, that sums up the goals, motivations and obstacles of an audience segment. The character is invented; the information behind it is not.
The word comes from the Latin persona, which meant the actor’s mask, then the role he played (Universalis). Personas are also used in interface design (UX) and in public service design, to keep the user in mind when decisions are made.
NN/g, a user experience research firm, sets three requirements: a persona must be memorable, actionable and distinct from the others (Page Laubheimer, NN/g, June 21, 2020). A profile that people forget, that changes no decision or that looks like its neighbor is useless.
In a content marketing strategy, the persona helps choose topics, level of detail and vocabulary. It replaces the question “what do we have to say?” with “what is this person looking for, and when?”.
Persona, target, buyer persona: what are the differences?
The target defines a segment, the persona gives it a face and a line of reasoning, the buyer persona focuses on the purchase decision.
| Concept | What it describes | Fictional example | Main use |
|---|---|---|---|
| Target | A segment defined by criteria (industry, size, area, age) | Building tradespeople with 1 to 10 employees, in France | Sizing a market, buying audience |
| Marketing persona | A typical person in the segment: goals, obstacles, questions, information sources | Karim, a plumber who runs a team of 6 and loses hours scheduling jobs | Choosing messages, topics and channels |
| Buyer persona | The buying reasoning: trigger, expected results, obstacles, criteria, journey | What pushes Karim to look for software now, and what stops him | Decision-stage content, sales pitch, B2B |
| UX persona | The user of a product: tasks, context of use, difficulties | Karim planning from his phone, on a job site | Product or website design |
Two personas can belong to the same target. A tradesperson who wants to save time and another who fears losing control of his schedule look alike on paper. They do not read the same content.
The buyer persona has a stricter definition among those who have turned it into a method. The Buyer Persona Institute (page updated on August 15, 2024) contrasts two models. The one-page “traditional” avatar says little about the decision; the “modern” buyer persona centers on a specific purchase. It documents five points: what triggers the search, the expected results, the perceived obstacles, the decision criteria and the buying journey. The example published by the institute is based on interviews with recent buyers.
Persona or personae: which plural in French?
In French, the recommended form is “un persona, des personas”. It follows the 1990 French spelling reform; “personae” is the Latin plural, still common, never a singular.
The report of the Conseil supérieur de la langue française (Higher Council for the French Language) was published in the French Official Journal of December 6, 1990, with the unanimous approval of the Académie française. Its rule 7 gives borrowed words a regular plural. The rule applies to nouns of Latin origin: the text cites “des maximums” and “des médias”. Only words that keep the value of a quotation remain invariable, such as “des mea culpa”.
Usage remains divided, including in the French civil service. The Interministerial Directorate for Public Transformation (DITP) titled a section “Un persona, des personae” on December 15, 2022. On the same page, it refers to “la méthode des personas” (the persona method).
Key point
In French, “des personas” is the recommended form, “des personae” is accepted, and “un personae” is a mistake. In French marketing usage, the word is masculine (“un persona”), as the DITP uses it. In the feminine, “la persona” mainly refers to the psychological concept taken up by Carl Gustav Jung around 1920.
On Elev8 Lab, we write “personas”.
What are the 3 types of persona?
The same NN/g article sorts personas into three types according to the data they rest on: the proto-persona, the qualitative persona and the statistical persona.
| Type | Data | Effort | Limitation |
|---|---|---|---|
| Proto-persona | Team assumptions, without new research | A 2 to 4 hour workshop, producing 3 to 6 personas | Often inaccurate: it reflects what the team already believes |
| Qualitative persona | 5 to 30 interviews, user tests, observation | Low, compatible with the team’s other work | Does not say what share of the audience each persona represents |
| Statistical persona | Qualitative research, then a survey of at least 100 respondents (ideally 500 or more) and statistical clustering | High, requires a statistician | Costly, for a result often close to the qualitative persona |
For most teams, NN/g recommends the qualitative persona. The proto-persona still has value: it makes explicit the assumptions everyone has in mind. It becomes a problem when it is presented as knowledge of the customer.
The same article advises against personas built only on demographic or analytics data. That data shows what people do, not what they were looking for or why.
How to create a marketing persona?
A marketing persona is built in three steps: gather data, spot shared patterns, write the profile.
- Gather data: interviews, CRM, search queries and questions asked to AI search engines.
- Spot shared patterns: group people who share goals, obstacles and ways of searching.
- Write the profile: one page per persona, each field linked to its source.
Gather data (interviews, CRM, searches, questions asked to AI)
Start with interviews: they give the “why” that numbers do not show.
5 to 30
To create qualitative personas, NN/g advises interviewing 5 to 30 users, in rounds of 5, until each new interview adds almost nothing.
Page Laubheimer, NN/g, “3 Persona Types: Lightweight, Qualitative, and Statistical”, June 21, 2020
Four sources complement each other.
- Interviews. Talk to recent customers, lost prospects, the sales teams and support. Ask what triggered the search, what nearly blocked the purchase, which words they used. The DITP bases its personas on this immersion work: interviews, questions, observation.
- CRM. Customer data from the CRM shows who buys, what, how often and through which channel. It helps choose who to interview, then check that a persona matches a real volume of customers.
- Search queries. In Search Console, the “Queries” dimension groups data by the query typed by users (Google Search Console Help). Sort these queries by search intent: each frequent question of a persona should match queries people actually type.
- Questions asked to AI. Ask your persona’s questions to ChatGPT or to Google’s AI Mode and note the sources cited. You see which content already answers these questions, and what is missing.
Watch out
A persona profile must not copy a real customer. For the CNIL, the French data protection authority, a person can be identified by combining several pieces of information (definition of personal data). An invented first name is not enough if the age, city, job and company point to a single person. Aggregate information from several customers.
Spot shared patterns
Reread the interview notes and sort each answer by theme: trigger, goal, obstacle, selection criterion, information source, vocabulary. NN/g calls this step data coding.
Then look for broad overlaps. Two people with the same trigger and the same obstacles belong to the same persona, even if their age or industry differ. Conversely, two identical demographic profiles can give two personas.
Keep few personas. NN/g points out that no one makes decisions while keeping dozens of profiles in mind. In the proto-persona workshop it describes, the group ends up with 3 to 6 personas.
Write the persona profile
A useful profile fits on one page and keeps only the fields that change a decision.
| Field | Question it answers | Where to find the information |
|---|---|---|
| First name and situation | Who is this, in one sentence? | CRM, interviews |
| Trigger | What starts the search now? | Interviews |
| Goal | What result does he expect, in his own words? | Interviews |
| Obstacles | What delays or blocks the decision? | Interviews, support, loss reasons in the CRM |
| Selection criteria | What does he compare? | Interviews, sales teams |
| Questions and queries | What does he type, what does he ask an AI? | Search Console, tests in AI search engines |
| Information sources | Where does he get information? | Interviews |
| Verbatim | Which sentence sums up his expectations? | Interviews |
The verbatim also appears in the DITP’s profiles, with a fake first name, the story that leads to the service, the obstacles and the motivations. Leave out details that change nothing, such as hobbies unrelated to the offer or an exact income when price is not an obstacle.
The profile then becomes the reference for choosing the tone and themes of the editorial line: the verbatim gives the vocabulary, the questions and obstacles give the topics.
Marketing persona example, with comments
The example below is fictional: it was built to illustrate the method and describes no real audience. It concerns an imaginary company that sells scheduling software to tradespeople.
| Field | Content (fictional example) | Source to use |
|---|---|---|
| First name and situation | Karim, 44, plumbing and heating engineer, runs a company with 6 employees | CRM (customers’ trade and size) |
| Trigger | A technician went to a customer who was not home: half a day lost | Interviews |
| Goal | “Know in the evening who goes where tomorrow, without spending an hour on the phone.” | Interviews |
| Obstacles | Fear of wasting time learning a tool; reluctant technicians | Interviews, support |
| Selection criteria | Phone app, monthly subscription, free trial | Interviews, loss reasons in the CRM |
| Questions and queries | “scheduling software for tradespeople”, “how to organize your technicians’ jobs” | Search Console, questions tested in AI search engines |
| Information sources | Colleagues, equipment supplier, short videos | Interviews |
| Verbatim | “I want a tool my guys use without training.” | Interviews |
The queries in the table are plausible wordings, not measured queries. In a real persona, they come from Search Console.
Four remarks on this profile:
- the trigger and the obstacles matter most: they say when Karim searches and what stops him. Content that answers “will my technicians use it?” is more likely to help him than a feature list;
- age adds nothing here; it could be removed without losing anything;
- the “Questions and queries” row links the profile to SEO: each query becomes a possible topic, attached to an intent;
- each field states its source. A field without a source is an assumption to check at the next interview.
What mistakes should you avoid with a persona?
Five mistakes make a persona useless or misleading.
- Describing a demographic profile. Age, gender and income do not say why someone buys; NN/g advises against personas based on this data alone.
- Presenting assumptions as facts. A proto-persona is useful if it is presented as such, then checked through interviews.
- Copying a real customer. The persona becomes too singular to represent a group, and the customer may be recognized.
- Multiplying personas. Beyond a few profiles, nobody uses them to make decisions anymore.
- Not linking the profile to decisions. Each topic in the editorial calendar should be traceable to a persona and to one of their questions.
Once your marketing personas are validated, you still have to turn them into a plan: which topics, for which persona, at which stage of their decision. That is the subject of the guide to building your content strategy.
Sources
- Encyclopædia Universalis, Persona (dictionary) (in French), undated page. Accessed on September 26, 2026.
- Nielsen Norman Group, Page Laubheimer, 3 Persona Types: Lightweight, Qualitative, and Statistical, June 21, 2020. Accessed on September 26, 2026.
- Buyer Persona Institute, Buyer Persona Example, updated on August 15, 2024. Accessed on September 26, 2026.
- Conseil supérieur de la langue française, The spelling reform (Les rectifications de l’orthographe) (in French), French Official Journal, Administrative Documents no. 100, December 6, 1990 (rule 7), PDF hosted by the Académie française. Accessed on September 26, 2026.
- Interministerial Directorate for Public Transformation (DITP), Understanding public service design: focus on personae (in French), December 15, 2022. Accessed on September 26, 2026.
- Google, Performance report (Search results): Overview and basic setup, Search Console Help, undated page. Accessed on September 26, 2026.
- CNIL, Personal data (Donnée personnelle) (in French), undated page. Accessed on September 26, 2026.
Digital marketing, SEO and GEO consultant
Article checked and updated by the author. Sources consulted on the date shown.
Cite this article
Clair, B. (2026, September 27). Marketing persona: definition and method. Elev8 Lab. https://elev8-lab.fr/en/content-marketing/persona/