At some point most founders decide to "ask customers what they think of the brand." Then they ask "do you like the logo?" and get a polite "yeah, it's cool." That answer is meaningless. It tells you nothing about whether the brand is working.
Useful brand research isn't about asking customers to evaluate your taste. It's about understanding what your brand is actually communicating to them. Which is often very different from what you think you're communicating. Here's how to do it without becoming a market research firm.
The wrong questions (and why)
These are the questions founders default to. None of them produce useful answers:
- "Do you like our logo?" Customers will be polite. Even if they hate it, they'll soften the answer. The data is unreliable.
- "What do you think of our brand?" Way too abstract. The customer has to figure out what you're asking before they can answer. Most will produce some version of "looks professional" because it's the safest response.
- "Does our brand resonate with you?" "Resonate" is jargon. Customers don't think in those terms about anyone's brand including yours.
- "How would you describe our brand in three words?" Better than the others but still produces vague answers ("modern," "clean," "trustworthy") that don't help you decide anything.
These questions fail for a common reason: they ask customers to be brand experts evaluating your work. Customers aren't brand experts. They're people who interact with your brand and have impressions formed by that interaction. The research should surface those impressions, not ask for evaluations.
The right questions
Five questions that produce useful answers:
Question 1: "If you had to describe what we do to a friend, what would you say?"
This is the single most useful brand research question. The answer reveals whether your positioning has landed. If customers describe you the way you describe yourself, the brand is doing its job. If they describe you differently, "you're like Notion but cheaper" when you're trying to be "a complete brand identity in 10 minutes". Your positioning isn't reaching them.
Question 2: "When you first landed on our website, what did you think we were before you read anything?"
The first-impression question. Asks about brand signals (visual identity, layout, tone) without asking customers to evaluate them. The answer tells you what your brand is signaling at a glance. Most founders find this answer surprising. Customers often categorize them as something the founder didn't intend.
Question 3: "What made you think we might be the right tool for [their use case]?"
Tells you which brand elements actually drove conversion. The answers reveal whether your headline, your social proof, your pricing, or something else did the work. You can then double down on what's working and rewrite what isn't.
Question 4: "What almost made you not sign up?"
Brand friction surfaces here. Maybe your pricing structure was confusing. Maybe the homepage didn't make clear who you're for. Maybe you looked too expensive, or too cheap, or too unfamiliar. The "almost" framing gives customers permission to share doubts they'd otherwise keep to themselves.
Question 5: "Have you recommended us to anyone? What did you say about us?"
The "told a friend" version of Question 1, but with social-proof framing that captures even more honesty. People share things in their own words; they don't borrow yours. The exact phrasing they used is gold for your messaging.
How to ask
The mechanics matter as much as the questions. Three principles:
Ask in writing, not in person. In-person conversations create politeness bias. Email or async survey produces more honest answers. Tradeoff: you can't follow up. Tradeoff is worth it.
Ask recent customers, not your most enthusiastic fans. Customers who signed up in the last 30 days remember the brand impressions; longer-term customers have integrated you into their workflow and can't reconstruct what they thought initially. Power users have selection bias. They're the people the brand worked best on.
Ask 8-12 people, not 100. Brand research isn't statistical. You're looking for patterns. Eight thoughtful responses produce more useful pattern recognition than 100 rushed ones. Most founders waste time chasing larger samples; smaller samples with deeper answers beat the math.
What to do with the answers
Don't average them into recommendations. Pattern-match instead.
Read all the responses to each question. Note what showed up more than twice. That's a real signal. Single outliers are interesting but not actionable.
Then ask yourself, for each pattern: is this what we want customers to think? If yes, the brand is working and you should amplify the signals that produced it. If no, the brand is communicating something off-target and the next quarter's brand work should address it.
Three patterns is a lot of data. Pick the most consequential one. Address it. Run the research again in six months. Compare patterns. That's how brand research compounds.
The mistake most founders make with this data
They run the research, get clear answers, and then... do nothing. The answers sit in a doc, the brand stays the same, and six months later they run the research again and get the same answers.
The point of asking is to change something. If you're not going to act on the answers, don't ask the questions. Customer time is a real cost; respect it by using what they tell you.
And if the answers come back and the brand is working. You find out the customers describe you exactly as you'd want. That's the most useful possible outcome. You can now move brand work to maintenance mode and spend your energy on the product. That's a real win worth the 30 minutes the research took.
Your brand kit, ready in 10 minutes.
Five quick taps. Free preview before you pay.
Start building free →