Building a Buyer Persona: A Practical Guide for SMEs

A buyer persona is the fictional but data-based description of a concrete customer. It has a name, a job, goals, pain points, media use and ideally a typical quote. The point is not the document itself but a shift in perspective: marketing, design and product decisions are no longer tested against an anonymous gut feeling but against a recognisable human being.

In most marketing blogs, the persona shows up as an Excel profile sheet with ten fields. But that sheet is only the artefact — the value emerges in the hour when a managing director has to call their own customer for the first time and realises they don't know them as well as they thought. That insight changes decisions — the profile sheet alone does not.

A persona that has impact in daily work is three things at once: a research tool, a decision tool and a shared mental model in the team. If one of those is missing, it becomes decoration.

Personas are not useful for every company. With five employees, a clear niche and daily customer contact, owners know their customers personally — a written persona adds little additional insight there. But as soon as several people make marketing, sales or product decisions without all of them having the same customer conversations, the method pays off.

Four recurring triggers have emerged from consulting Swiss SMEs.

In most cases on the left side, customer knowledge is already there — distributed across the team, in the owner's head, in individual emails. The persona is then the tool that makes this knowledge visible, testable and shareable. During a generational handover or a larger relaunch, the exercise is both stocktaking and realignment — the knowledge is gathered anew, not just consolidated. If you're unsure whether your SME is at the right maturity level, an honest assessment of your digital presence often delivers a quicker answer than a persona workshop itself.

Four terms come up in practice that are often used interchangeably but serve different purposes. Separating them cleanly avoids the most common misunderstanding: that a single method answers all marketing, sales and product questions.

The methods complement each other, they do not replace one another. A typical Swiss SME first needs a clean target group (for media planning), then a persona (for tone of voice and website), and with a clear B2B focus also an ICP (for sales). Jobs-to-be-Done is the most useful addition when features have to be prioritised — for example in a product or service roadmap.

A persona workshop without preparation produces what most SMEs end up with in a drawer: a plausible but generic profile. The line between useful and arbitrary lies largely in the preparation — and a second time in the disciplined translation afterwards. Investing one evening in the following questions saves two hours of warm-up in the workshop and avoids the usual loop towards interchangeable mission statements.

The corporate version of a persona process takes three days, involves twenty stakeholders and ends in a twelve-page document. For a Swiss SME with twelve employees, that is simply disproportionate. Noevu therefore uses a compact variant that deliberately limits stakeholders to three to five, finishes within half a day, and makes the result directly usable on the website — instead of for a report.

The typical flow has four steps. First, a preparation call with management, in which the six guiding questions from the previous section are clarified. Second — where possible — two or three short customer conversations conducted by the SME itself, with notes for the workshop. Third, a two- to three-hour workshop with three to five people from the team, in which a concrete persona is condensed from data plus empathy-map quadrants (says, thinks, does, feels). Fourth, the translation into hero claim, navigation, copy and CTA labels — the actual design step that takes two to four weeks separately.

The recurring aha moment often happens in the first hour: when the team reads a concrete customer quote for the first time and realises that their own website does not speak to that quote. From this moment, the best decisions emerge — not from the demographic description.

The method is always the same — the results depend on maturity and data. Three recent examples from Noevu's practice make this concrete. At a Zurich-based security service provider focused on events, the workshop produced three clearly different personas: an event manager in her mid-30s focused on reliability, a production lead who needs crowd management and quick response above all, and a young security professional looking for flexibility and training. From these three persona profiles the website navigation was structured directly — Security Services, Event Services, Academy.

At a Swiss compensation office — a context with a complex public-sector customer journey — three personas and the matching customer journey map were condensed in the workshop together with the internal team. The insight was uncomfortable: the existing website did not answer the very question every persona asks first — "Am I in the right place, and what concretely do I have to do?" At an SME in the lighting expertise space, the data was too thin for solid personas. The workshop produced clearly marked hypothesis profiles with explicit provisional status; they were tested against real enquiries during the first six months after launch and only then condensed into personas.

In all three cases, the workshop alone delivered only half the impact. Preparation and the disciplined translation of the persona insights into hero claim, navigation and copy made the difference — complemented by coming back six months later to test the hypotheses.

Most persona articles end with the description of the persona. In practice, that is the beginning, not the result. A persona that has value directly influences four areas of the website — and is measurably effective there, not in a PDF.

A good persona shows up concretely in the hero section, navigation, about-us page and CTAs.

Whoever makes the jump from persona to web decisions cleanly ends up with a website that does not sound like every other one. Skip that jump and you have a beautiful persona slide and an interchangeable website. A complementary exercise is Simon Sinek's Golden Circle — a compact method to articulate the why behind persona needs. More on that in Simon Sinek's Golden Circle: what's actually useful for your brand.

At first glance, personas and search engine optimisation have nothing to do with each other: empathy on one side, keywords and load time on the other. The connection runs directly nonetheless — through content that emerges from real experience with the persona.

Google has been weighting the acronym E-E-A-T for years: Experience, Expertise, Authoritativeness, Trustworthiness. Translated, this means: content has to come from lived experience, not from generic copy blocks. A website whose team actively uses the persona in briefing and writing almost inevitably formulates around the real questions, in the real language, with the real examples of its customers. That is rewarded — more directly than any keyword research.

With AI-generated search results — AI Overviews, ChatGPT and Perplexity answers — this dynamic intensifies in 2026. AI answers cite sources more often that stand out from interchangeable copy: concrete examples, local references, a recognisable voice. Interchangeable content disappears into the summary "various providers offer similar services". Anyone writing from a clearly defined persona perspective is, by contrast, cited, linked and recognised as a source.

The following mistakes show up in persona projects so regularly that they can be addressed up front. They share one thing: they do not lie in the method itself, but in how it is handled — and that means they can be avoided reliably.

The most common mistake is the persona without a data base. It is "invented" in a workshop on the basis of gut feeling and assumptions, instead of being derived from customer conversations. The result feels plausible but does not describe the real customer base. The second mistake is persona inflation: five, seven or ten personas, because the team does not want to say no to anyone. In the end, none of them is referenced in daily work. Third — and perhaps the most expensive — the persona as a marketing island: it is created by marketing and never reaches sales, support or product development. Personas only unfold their full impact across departments.

AI tools promise personas in minutes — ChatGPT, HubSpot Make-My-Persona, Delve.ai, 121WATT. The 2026 market is full of vendors positioning "synthetic personas" as a substitute for research. Current research — among others from ACM and Springer — sees this far more critically and talks about the "synthetic persona fallacy".

The finding is clear: AI-generated personas reflect the training corpus of the models, not the real audience of a Swiss trades SME. The training corpus is predominantly English-language, urban, tech-affine. A Bernese trust office or a Valais family business is missing from it. Anyone adopting an AI persona without scrutiny optimises their website for an audience they don't even have — copy, hero message and service order then aim past the real audience.

AI remains useful as a tool in three clearly bounded roles: first, as a structural aid after real interviews — summarising notes, clustering themes, ordering quotes. Second, as a hypothesis generator before research — formulating assumptions that are then tested in interviews. Third, as a stress test for existing personas — "which assumption in this persona could be wrong?". In all three roles, AI complements human research. It does not replace it.

In consulting practice, we often see the same pattern: a large share of all personas is no longer referenced in daily work after six months. They don't die because they were bad — they die because no one is responsible for keeping them alive. An outdated persona is more dangerous than none at all: it suggests customer knowledge that no longer holds.

The pragmatic way out is a light quarterly refresh. One person from the team — ideally not from marketing alone — runs two short customer conversations once a quarter, checks CRM and analytics data for anomalies, and adjusts one or two fields of the persona. Once a year a larger review follows: is the persona as a whole still right? Has a new segment shifted enough to justify a second persona? This routine costs about two days of work per year — and reliably prevents the drawer fate.