Adoption begins with the first action that gives a customer value, not the first event that makes an investor dashboard move. For Amel’s Tunisian fintech, that action happens when a cash-first customer successfully brings money into the product, not when someone links a bank account.
In London in 1854, physician John Snow faced a measurement problem with fatal consequences. Cholera deaths were accumulating around Soho, while the dominant explanation blamed contaminated air. Snow suspected the water.
The map did not settle the argument
Snow mapped deaths near the public water pump on Broad Street and investigated where affected households obtained their water. The pattern supported his theory, but the evidence was imperfect. Some nearby residents had not become ill. Some victims lived farther away.
The exceptions mattered because Snow was challenging the accepted account of how cholera spread. He could not treat every address as an identical data point. He had to ask what people had actually consumed.
That distinction led him beyond proximity. Workers at a nearby brewery largely drank beer rather than water from the pump. A woman who had moved away still preferred Broad Street water and had it delivered. The useful measure was exposure to the water, not distance from the pump.
Local officials eventually removed the pump handle. The outbreak was already declining, so the intervention did not provide a clean experiment. Snow’s work still helped establish the link between contaminated water and cholera. Steven Johnson documents the investigation and its complications in The Ghost Map.
The shape of Amel’s problem is smaller, but familiar. Her investor sees a linked bank account as evidence that a customer has entered the product. Her customers begin with cash. The dashboard records the behaviour the company expected, while missing the behaviour the market actually performs.
Signup was measuring access, not use
A linked account can be valuable. It may reduce later friction, support recurring transactions, or make reconciliation easier. None of that proves the customer has reached the first useful moment.
Amel’s customers are giving her a harder signal to interpret. They may register, inspect the product, then stop when the next step assumes money already sits in a compatible account. They may intend to fund through a cash-based path. They may need help from an agent, merchant, or another person before value appears inside the product.
Counting linked accounts makes those customers look inactive or incomplete. Counting registrations makes the funnel look larger without explaining who can transact. Both measures describe system events. Neither establishes adoption on its own.
This is where a founder has to resist the clean chart.
The first question is practical: what must a customer accomplish before the product becomes useful enough to return to? For this product and this market, the answer may be a first successful cash-funded balance, payment, transfer, or other core transaction. The exact event depends on what Amel has built. It cannot be borrowed from an investor template designed around banked customers elsewhere.
I have seen this error appear whenever teams measure the easiest event to instrument. The event gets a name in analytics, the name becomes a board metric, and the metric slowly replaces the customer’s actual job.
Redefine adoption around completed value
Amel needs two views rather than one disputed number.
The first view should show access paths. How did the customer attempt to enter the financial system behind the product? A bank link belongs here, alongside the cash path and any other supported funding route.
The second should show the first completed act of value. This is the adoption measure. It should require a successful outcome that the customer can recognise, rather than a button tap or initiated transaction.
That separation changes the product conversation. A customer who funds through cash and completes the core action has adopted, even without linking a bank account. A customer who links an account but never completes that action has finished setup, not adoption.
The team can then examine conversion by starting condition:
- How many cash-first customers begin a funding attempt?
- How many complete it?
- Where do unsuccessful attempts stop?
- How long does it take before the customer reaches the first useful outcome?
- How many return and repeat the action?
These questions expose product work. A single signup number conceals it.
The same distinction appeared in Youssef’s product launch, where making the cash step visible protected the real route to use. Visibility matters because a hidden step can look like weak demand when it is actually an unsupported path.
Put the disagreement into the dashboard
Amel does not need to persuade her investor with a better definition alone. She needs a cohort comparison.
Keep linked accounts visible. Add the cash-first cohort beside it. For each group, show the share that reaches the first completed value event, returns, and performs it again. If bank linking predicts repeated use, the investor’s measure earns its place. If cash funding predicts it better, the adoption definition must change.
John Snow’s map became useful because he investigated what the dots represented. Amel’s dashboard needs the same discipline. A dot for signup, a linked account, or a funding attempt only becomes meaningful when it connects to what the customer actually did.
The next product meeting should end with one sentence everyone can test: “We count a customer as adopted when they complete this action.” Then Amel can instrument every route to that point, including the one that begins with cash.
Comments
No comments yet.