Customer Discovery: How to Talk to Users the Right Way
Photo by Unsplash
Most founders I meet in Nairobi can pitch their product in their sleep.
Ask them when they last sat with a real user and listened for an hour, and the room goes quiet.
That gap is where good ideas go to die.
Customer discovery is the craft of talking to the people you want to serve so that you learn what is true. Done well, it tells you whether the problem you are solving is real, painful, and worth paying for. Done badly, it gives you a folder of nice comments and a runway burning down to zero.
There is a famous example sitting right here in Kenya. M-Pesa began life as a tool for microfinance borrowers to repay loans over their phones, built with the firm Sagentia. During an eight-month pilot in 2005 across Nairobi, Mathare, and Thika, the team kept watching people do something they had not designed for: sending airtime credit to relatives upcountry as a way of moving money home. Safaricom listened to that behaviour and rebuilt the product around "Send Money Home" before the 2007 launch (McKinsey, 2022). Today M-Pesa processes more transactions within Kenya than Western Union does globally (IMF). That entire pivot came from discovery, from watching what users actually did.
This guide is the practical version of that lesson. Here is how to talk to users the right way.
Pack this before you go
You do not need a research budget to do this well. You need preparation.
A real hypothesis written in one sentence. Something like "Mama mbogas in Eastlands lose stock because they cannot track daily sales." If you cannot state who and what, you are not ready to interview.
A list of 8 to 12 people who actually have the problem, not your friends and not other founders. Drivers, traders, nurses, farmers, whoever lives inside the problem.
A notebook or a recorder (ask permission first). Your memory will betray you by the third conversation.
One free hour and zero slides. No demo, no deck, no pitch. You are there to learn.
A teammate if you can bring one, so one person talks and the other writes everything down.
Step 1: Find the people who actually have the problem
The fastest way to ruin discovery is to interview the wrong people.
Your aunt will love your idea. Your co-founder will love your idea. None of that is signal. You want people who feel the pain today, with their own money and their own time on the line.
Go where they already are. If you are building for boda riders, stand at a stage in Kasarani and buy a few of them tea. If you are building for SME shop owners, walk into the shops on a weekday morning when business is slow enough to talk. Africa has over 700,000 professional developers as of the Google and Accenture Africa Developer Ecosystem report (2021), and the number has grown sharply since, yet most of us still build for users we have never sat beside. Closeness is the edge.
Aim for at least 10 conversations before you draw any conclusion. Patterns do not show up in two.
Pro tip: Recruit one person per channel, then ask each one, "Who else do you know who struggles with this?" Referrals reach the people surveys never find.
Step 2: Ask about the past, not the future
This is the rule that separates real discovery from theatre.
Never ask "Would you use this?" or "Would you pay for this?" People are kind. They will say yes to be polite, and that yes will mislead you for months. Future questions invite fantasy. Past questions surface facts.
So ask about what already happened:
"Tell me about the last time this problem cost you money."
"What did you do the last time this happened?"
"How much time or cash did that take?"
"What have you already tried to fix it?"
These come straight from the spirit of Rob Fitzpatrick's "The Mom Test," and they work because behaviour that already occurred cannot be invented to please you. If someone has never bothered to solve the problem before, that tells you the pain is mild, and mild pain rarely opens wallets.
Pro tip: When someone says "I would definitely use that," gently steer back to reality: "Have you ever paid for anything like it?" The answer to that one is worth ten enthusiastic maybes.
Step 3: Stop selling and start watching
The best discovery often happens with your mouth closed.
Talk for less than a third of the conversation. Ask a question, then sit in the silence until they fill it. The pause feels uncomfortable to you and completely normal to them, and the richest answers usually arrive after that pause.
Where you can, watch people do the task in front of you. Ask the shopkeeper to show you how she records today's sales. Ask the farmer to walk you through how he checks the day's market price. The M-Pesa team did not learn about "Send Money Home" from a focus group. They learned it by watching what people actually did with the phones in their hands.
What people do and what people say are two different datasets. The doing is the truth.
Step 4: Listen for the workaround
Every real problem already has a clumsy solution attached to it.
Your job in discovery is to find that workaround, because it tells you three things at once: the problem is real, the person already spends effort on it, and there is a budget, even an informal one. A shopkeeper tracking sales in a worn exercise book is showing you both the pain and the price she already pays in time.
Internet access shapes which workarounds even exist. Around 38% of Africans were online in 2024 (ITU and Statista, 2024), and roughly 74% of African web traffic comes through mobile phones. So the workaround you uncover might be a WhatsApp group, an SMS chain, or a cash ledger. That context tells you what your real competition is, and it is usually an existing habit.
When you hear a workaround, dig in. How often do they do it? How much does it cost them? What do they hate about it? That is your product brief, written in the user's own voice.
Step 5: Make sense of what you heard
Ten good conversations produce a pile of notes that means nothing until you process it.
Do it within 24 hours, while the tone and the body language are still fresh. For each interview, write down three things: the exact problem in their words, what they currently do about it, and any moment that surprised you. Then lay all the interviews side by side and hunt for the pattern that repeats across at least half of them.
A few honest outcomes you should be ready for:
Strong, repeated pain with active workarounds. Green light. Build a small thing and test it.
Mild interest, no workarounds, nobody has paid to solve it. The problem is not painful enough yet. Change the segment or the problem.
One loud user and nine quiet ones. Resist building for the loud one. Follow the pattern that the quiet majority shows you.
Discovery does not end when you start building. It becomes a loop. Keep talking to users through every release, because the market keeps moving and so do their needs.
Common mistakes to avoid
Pitching instead of listening. The moment you start selling, the learning stops.
Asking leading questions like "Don't you think this would be useful?" You are writing the answer for them.
Talking only to friends and fellow founders. Comfortable, flattering, and useless.
Counting compliments as validation. Praise is not a purchase order.
Stopping after three interviews. Patterns need a crowd to appear.
Ignoring the workaround because it makes your idea look less original. The workaround is the gold.
Forgetting the context. A beautiful app means nothing to a user who lives on patchy data and a feature phone.
Talk to users the right way and discovery becomes the steady drumbeat underneath everything you ship.
The founders who win in African markets are the ones closest to the people they serve. Go and sit with them.
Further reading
Over to you: What is the most surprising thing a user ever told you that changed your product? Share it in the comments.
Go deeper with us. Join the Hackhouse community for conversations that go beyond the surface, where builders share the hard-won lessons that never make it into press releases.
