There is an assumption underneath most business writing about customer listening that I want to challenge directly. The assumption is that customers, given the right channel and the right prompt, will tell you what they think. They will fill in surveys honestly, complain when they are unhappy, praise when they are pleased, and the founder’s job is to install the listening infrastructure that captures these honest signals.
This assumption is broadly false, and it is especially false in African markets. The customer does not always tell you the truth. Sometimes they cannot, because the social cost of explicit complaint is high. Sometimes they will not, because they have learned that complaining produces nothing. Sometimes they would tell you, but you have not designed a channel that fits the way they actually communicate. The founder who waits for the customer to articulate dissatisfaction in clean prose will wait forever, and during the waiting the customer will quietly leave for a competitor who never asked them to articulate anything.
Listening, in the way that matters for a Stay-Up phase venture, is not the act of receiving feedback. It is the discipline of reading the signal that is sitting in plain view, underneath whatever the customer has actually said.
The signal under the silence
In a stable market with educated, articulate customers and a functioning consumer protection regime, explicit complaint is cheap and signals real dissatisfaction. In African markets, explicit complaint is expensive. Customers who complain risk being seen as difficult, demanding, or ungrateful. Customers who have been let down by previous providers learn that complaining is wasted breath. Customers who depend on a relationship for ongoing service know that complaining can damage the relationship and degrade the service.
The result is that the most important customer signals in an African market are not the explicit complaints. They are the silences. They are the customers who used to come twice a month and now come once. The customers who used to call to confirm bookings and now do not bother. The customers who used to refer family members and have stopped. The customers who pay on time but stop talking to the waiter who used to know their order.
A founder who is only reading the explicit complaint channel will see almost none of this. The complaints inbox is empty. The reviews are five stars. The exit surveys, when filled in, are bland and generic. The founder concludes that customers are happy. By the time the financial picture reveals that customers were not happy, they have already gone, and the venture is reading the wrong signal too late to act on it.
The Sprouting Curve framework I have written about elsewhere makes this concrete. Early-stage ventures should be accumulating learning faster than revenue. The learning that matters most is the learning about why customers do what they do. If the only channel for that learning is what customers explicitly say, the channel is broken in markets where customers are reluctant to speak. The discipline of listening, then, is the discipline of building richer channels, ones that capture what customers do as well as what they say, and especially what they stop doing.
What real listening looks like in practice
I want to translate this into something operational, because the abstraction is useless without it. Here is what listening actually looks like in the ventures I run.
At Cafe Oldrock, I look at three numbers every week that no customer ever says out loud. The first is the rate at which regulars are returning. Specifically, the customers who came at least three times in the previous quarter, what fraction came at least once this month. A drop in this number is more diagnostic than any complaint, because regulars who stop coming almost never tell you why. They simply stop. The number tells me, before any individual relationship has been articulated, that something has shifted.
The second is the rate of word-of-mouth referrals. We track, imperfectly, when a new customer mentions another customer who recommended us. A drop in this rate is the second signal. Customers who have stopped recommending have stopped feeling proud of the relationship, even if they themselves still come. The recommendation is the strongest possible vote of confidence; its absence is a silent vote against.
The third is the staff observation log. Once a week, my managers write down three things they noticed about customer behaviour that they cannot explain. Customers who used to tip generously and now do not. Customers who used to bring their children and now do not. Customers who used to order the most expensive bottle and now order the second-most. The pattern of unexplained changes is itself the signal. Individually, each observation is noise. Collected and reviewed weekly, they reveal trends that no survey would capture.
None of these channels involves asking customers what they think. They all involve watching what customers do. The watching is the listening.
Asking, when you do ask, like someone who already knows
There is still a place for asking customers directly, but the asking has to be designed for a market in which customers will not volunteer the truth easily. The pattern that works is to ask questions that already presume the customer’s perspective rather than questions that invite the customer to summarise it.
A founder who asks “are you happy with our service?” gets the answer “yes, thank you” almost regardless of the truth, because answering otherwise feels socially expensive. A founder who asks “what is one thing you would change if you were running this business?” gets a real answer more often, because the framing makes the answer feel constructive rather than complainant. The question is the same in spirit; the framing changes everything about what the customer is willing to say.
The same logic applies to staff. A junior employee asked “is everything fine?” will say yes. The same employee asked “what is one thing about how we work that you would change if you had the authority?” will sometimes say something genuinely useful. The framing matters because it shifts the social cost of telling the truth.
This is also why the most useful customer conversations almost never happen in formal feedback channels. They happen at the table when the customer is paying the bill, at the door as they leave, in the WhatsApp message they send the manager directly because they trust the manager. The founder who designs a venture’s listening infrastructure as if customers will write thoughtful long-form responses to email surveys has designed listening for a market that does not exist on this continent. The founder who designs it for the actual texture of how relationships work in this market builds something different and far more useful.
The slow listening that compounds
The listening I am describing has one feature that distinguishes it from most business advice on customer feedback: it compounds slowly, over years, and it cannot be shortcut by software. A new customer relationship management tool cannot replace the staff observation log, because the observations require human attention to behaviours that no system has been instrumented to capture. A new survey provider cannot replace the question reframing, because the framing has to come from someone who knows the cultural texture of the market.
This is, in some ways, frustrating. Founders prefer scalable solutions. The venture I am describing is not scalable in the conventional sense; it depends on the founder and the senior managers maintaining a posture of attention that does not delegate cleanly to junior staff or to software. But this is also why it is a moat. Competitors who have not built the listening discipline cannot acquire it quickly, and customers who are receiving the benefit of being deeply listened to do not leave for competitors who treat them as data points.
In Stay-Up phase ventures, the founders almost always credit something they call “knowing the customer,” and what they mean by it, when you press them, is roughly what I have described here. Not customer surveys. Not net promoter scores. A texture of attention, accumulated over years, that lets the founder read the silence as well as the speech, and act on what the silence is saying before the customer has decided to leave.
The week’s diagnostic
If this piece has landed, the smallest concrete thing you can do this week is to identify one customer relationship in your venture that has changed in the last six months in a way no one has spoken about. Maybe a regular who comes less often. Maybe a corporate account that used to pay early and now pays late. Maybe a referrer who has stopped referring.
Write down what changed, when you noticed, and what the change might be telling you. Then ask one person on your team if they have noticed anything similar elsewhere. The answer to that question is almost always yes, and the answers, taken together, will reveal a pattern your venture has been ignoring because nobody has been instructed to look for it. That ignoring is the listening failure. The looking is the discipline.
The Stay-Up phase is built on listening of this kind. It is unglamorous, slow, and unteachable in any classroom. It is also one of the few things that produces a venture that can survive its founder, because the listening teaches the senior team to read the same signals the founder reads, and over time the signal-reading becomes the institutional culture, not just the founder’s habit.
That is the listening that matters. Not the surveys. The watching, the framing, and the slow accumulation of an attention that the customer feels even when they have not articulated it.
For the framework that underlies how I think about early-stage learning, see The Sprouting Curve. For the related discipline of building a vision that lets the listening land, see The Vision That Does Work.
