How to choose a branding agency: the fit test that survives the pitch
The direct answer
Choose a branding agency the way you would choose a surgeon: by diagnosis, evidence, and the named people doing the work. Shortlist up to three candidates who publish reasoning beside their portfolios, then ask each one what problem they think you have before any deliverable is discussed. Confirm who exactly will do the thinking, and buy a small paid piece of work before the large one. Price predicts quality far less than sequence does: decisions first, design second, years of consistency third.
Five conclusions
The argument, compressed.
- A pitch measures presentation skill. The reliable signals are diagnosis, named people, published reasoning, and willingness to start small.
- Shortlist by evidence: case studies that show the thinking and the numbers, from clients who can be found and asked.
- The first question to any candidate is what problem they think you have. Deliverables before diagnosis is a template, at any price.
- Meet the hands. The people in the pitch and the people in the project are often different, and the difference is where quality leaks.
- Buy a small diagnostic engagement before the large one. Two weeks of paid thinking reveals more than two months of references.
Working framework · 5 decisions
The selection sequence
Five stages, run in order, keep the decision with the buyer instead of the best presenter in the room. Each stage exists to disqualify, which is why the sequence moves fast: a shortlist of three usually becomes one candidate by the third stage.
Decision 01 / 05
Name the problem
Decide what is actually broken before anyone pitches a cure. Buyers who meet the business and struggle to say what makes it different point to a strategy gap. Buyers who can say it while the materials look and sound inconsistent point to an identity gap. Sales that stall with nobody arriving at all point to distribution, and branding money is the wrong money there.
Why comparing agencies is so hard
Every instrument a buyer is given for the comparison measures the wrong thing. Portfolios measure taste, and taste has converged: the same references, the same tools, the same tasteful sameness across the category, which is why ten shortlisted sites can feel like one agency with ten names. Awards measure what other agencies admire, judged by juries who never carried the commercial result. The pitch measures rehearsal.
The question that actually matters stays unmeasured by all three: what decisions will these people make on your behalf, and what evidence will they demand of themselves before making them? Branding is a decision service wearing a design coat. Buying it on visual polish alone is how a business ends up with an attractive identity carrying an unexamined strategy.
So the selection needs its own instruments. The rest of this guide is those instruments: a triage you run before the shortlist, an evidence standard for building it, the questions that expose a proposal, the red flags that end a conversation, and the small first engagement that replaces faith with observation.
The plain version
Portfolios show taste, awards show peer applause, pitches show rehearsal. Diagnosis, named people, and a small first project show the truth.
Name the problem before the shortlist
Agencies sell cures, so an undiagnosed buyer gets prescribed whatever the seller stocks. A design studio finds an identity problem, a strategy shop finds a positioning problem, a performance shop finds a media problem, and each is sincere. The only defence is arriving with your own triage already written.
The triage takes an afternoon. Ask five recent buyers, or read five recent lost deals, and sort the failure: they never understood what made us different points to strategy. They understood it but our materials undermined it points to identity. They never encountered us at all points to distribution. Each answer implies a different vendor, a different budget, and a different definition of success.
Write the finding in one sentence and hand it to every candidate as the brief. Their reaction is the first real datum of the selection: the ones who test the triage before accepting it are the ones who will test their own work later.
The evidence standard for a shortlist
A public record tells more than a reference call, because references are curated and the record is habitual. Before any meeting, hold each candidate to a standard the good ones already meet in public.
Reasoning beside the work. A case study worth the name states the problem as the client experienced it, the decisions taken, and what changed, with numbers where numbers exist and honesty where they do so far. A gallery of finished logos with adjectives underneath is decoration about decoration.
Honest boundaries. Practices that label concept work as concept work, name the limits of their numbers, and publish at least a starting price are showing you how they will treat your evidence later. Hidden pricing moves the money conversation to the end of a sales process, where sunk time does the persuading.
- Case studies with problem, decisions, and outcome, at least one with a number and its baseline
- Clients who are named, findable, and willing to take one short call
- A stated starting price, or a written range on request within a day
- A visible line between commissioned work and self initiated concept work
Red flags a good price fails to excuse
Most bad engagements announce themselves in the first two meetings. The signals below are disqualifying regardless of fee, chemistry, or how much the work in the portfolio resembles what you want, because each one predicts a structural failure rather than a bad day.
- Design begins in week one. Strategy skipped is strategy guessed, and the guess hardens into expensive artwork.
- A proposal arrives before a diagnosis. It was written for the pipeline, and your name was inserted afterwards.
- Guarantees of rankings, followers, or virality. Those are promises about systems nobody controls, made to people assumed to be unable to check.
- The pitch team and the project team are different people, and the agency resists naming who does what.
- The service list covers everything for everyone. A shop with no declared focus has decided its positioning matters less than its pipeline, which tells you how it will treat yours.
- Resistance to a small first engagement. A partner confident in the working relationship has no reason to refuse a paid trial of it.
Start with a small paid engagement
The single most reliable instrument in the whole selection is a short, paid, clearly scoped first project: a brand audit, a positioning sprint, a messaging review. Two or three weeks, a defined deliverable, a price both sides can walk away from.
It converts every open question into observation. How they handle a deadline, how they take a challenge to their thinking, whether the senior person stayed in the room, whether the deliverable states decisions or restates your brief in nicer typography. References describe the past through a curator; the diagnostic shows you the present directly.
Structure it honestly on both sides: pay properly for it, own the output regardless of what happens next, and agree upfront that either side may stop at the boundary with no explanation owed. The agencies that welcome those terms are the ones worth the main engagement.
The working test
Two weeks of paid thinking reveals more than two months of pitches, references, and dinners combined.
Fit beats fame
The final filter is size and shape, and the honest rule is that the right partner is the smallest one that can carry the problem. Many hands moving at once, a campaign, packaging, motion, and identity landing together, is genuine agency territory. A problem of thinking, positioning, messaging, a focused identity, is usually carried better by an independent strategist or a small studio, where the person in the pitch is the person in the work.
Geography stopped mattering when the work moved to documents, calls, and shared screens. A remote specialist who fits the problem beats a local generalist who fits the postcode, and the selection sequence above works identically across borders, which is exactly how buyers in one country now hire practitioners in another.
Whoever survives the sequence, the relationship that follows runs on the same fuel the selection did: named people, stated reasoning, evidence over confidence. Choose the partner who behaved that way before the contract, because contracts change behaviour far less than buyers hope.
Before you use it
Questions that can change the recommendation.
How many branding agencies should I shortlist?
Three at most. The evidence standard above disqualifies most candidates before a meeting, and three parallel conversations is the limit at which a founder can stay a sharp evaluator. Longer shortlists select for stamina, and the pitch machine of a large agency outlasts a tired buyer every time.
How much does a branding agency cost?
The market runs in three tiers: independent strategists and small studios from the low thousands, mid size agencies in the tens of thousands, and large agencies far beyond that. Published starting figures at this practice sit at £1,950 in the UK, $2,800 in the US, and CA$3,200 in Canada; the pricing guide in this library breaks down what moves a project between tiers. Treat any quote as untestable until a diagnosis explains what the money buys.
Should the agency be in my city?
Only when the work is physical: signage, interiors, events. Strategy, identity, and messaging travel as documents and conversations, and remote engagements have been ordinary for years. Choosing from a whole market instead of a postcode raises the quality ceiling far more than a shared office lowers the friction.
What belongs in the contract?
The named people doing the work and a substitution clause requiring your approval. Staged payments tied to defined deliverables. Full transfer of the finished intellectual property on final payment, including source files. An exit point after the diagnostic stage. Everything else is standard commercial hygiene; those four clauses are the ones agencies quietly omit.
What if every shortlisted agency looks the same?
Treat the sameness as the finding. When three candidates show interchangeable portfolios and identical language, the category has converged on taste, and taste is what you were about to buy. Reopen the search one level down in size, toward specialists whose public writing shows a point of view, because a partner with no visible opinions will build you a brand with none either.
