Most personas fail the moment someone has to use one. They describe a person — age, job title, a name, a hobby nobody asked about — and then sit in a deck while the media team makes every real decision without them.
A useful persona is not a character sketch. It is a compressed description of a buying decision: what someone is trying to solve, what would stop them, and where they go looking before they commit. If it does not change a targeting choice, a message, or a landing page, it is decoration.
What a persona is actually for
Personas exist to settle arguments cheaply. When three people disagree about a headline, a good persona ends the debate by pointing at what the buyer is worried about. When a campaign underperforms, it narrows the search: wrong audience, wrong objection, or wrong moment.
That only works if the persona carries decisions rather than description. Age and income rarely change a creative brief. The reason someone delayed a purchase for six months almost always does.
- The job to be done — what outcome they are actually buying, in their words rather than yours.
- The trigger — what changed recently enough to make this urgent now.
- The objection — the specific reason they would close the tab and do nothing.
- The alternatives — what they compare you against, including doing nothing at all.
- The proof they need — the one thing that would make the decision feel safe.
Where the inputs come from
Personas built from internal opinion tend to describe the customer the team wishes it had. The inputs that survive contact with a campaign come from people who have already bought, already churned, or already said no.
Sales call recordings are the densest source most companies already own — objections arrive unprompted and in the buyer language you should be writing in. Support tickets show where the promise broke. Search queries and on-site search show what people call the problem before they know your vocabulary for it.
- 01Gather
- 02Synthesise
- 03Validate
- 04Apply
- 05Revisit
How many you actually need
Teams routinely build six personas and use one. The count should follow how differently you would behave, not how many segments the data can produce. Two groups that see the same creative, land on the same page, and convert on the same offer are one persona wearing two names.
Split when something downstream genuinely changes: a different objection to answer, a different channel to reach them in, a different proof to supply. Merge whenever the difference is only descriptive.
If two personas would never receive different work, you have one persona and a formatting problem.
Putting them to work
In media buying, personas map to intent signals rather than demographics — the queries someone runs, the comparisons they read, the moment in the season they start looking. In creative, each persona should generate a concrete list of angles, because an objection is a brief and a trigger is a hook.
On landing pages, the persona decides what appears above the fold. A buyer whose blocker is trust needs proof early. A buyer whose blocker is effort needs the process made visible. Same product, different first screen.
How they go stale
Personas decay quietly. Objections shift when a competitor changes its pricing. Triggers shift when a regulation or a season moves. The document stays confident while reality moves underneath it, which is worse than having no persona at all — a wrong map is followed further than no map.
Treat them as living documents with an owner and a review date. Re-read a fresh batch of sales calls each quarter and change what no longer matches. The persona that gets edited is the one still being used.
The common ways they go wrong
Persona failures are repetitive enough to be worth naming. Almost all of them come from describing a person instead of describing a decision.
- Demographics standing in for motivation — age and job title rarely explain why someone bought.
- Invented detail, added for colour, that nobody can source back to a real conversation.
- One persona per segment the analytics tool happens to produce, rather than per decision you would actually make differently.
- Written once, approved, and never opened again by the people making daily media choices.
- Aspirational targets — the customer the company wants rather than the one currently paying.
The aspirational failure is the most expensive, because it is the hardest to notice. A persona describing a larger, more sophisticated buyer than you actually serve will quietly pull creative, targeting and pricing toward a market that is not converting, and the campaign data will look like an execution problem for months before anyone questions the brief.
B2B changes the shape
In consumer marketing a persona usually maps to one person making one decision. In B2B it rarely does. The person who feels the problem, the person who signs, and the person who can veto are often three different people with three different objections, and a persona that merges them helps nobody.
Model the roles separately. The champion needs ammunition to sell internally. The economic buyer needs a defensible case. The blocker — usually security, legal or finance — needs a specific risk removed. Same deal, three briefs, and the campaign that only speaks to the champion stalls at exactly the point it looked like winning.
The test is never whether the persona is well written. It is whether someone made a different decision because of it this month.
