Customer investigation · running0%
Customers you know
Segment found1.9M reachable
Segment found640K reachable

Illustrative

The discipline

Customer Diagnostics

Customer Diagnostics is the process of finding overlooked customers and growth opportunities hiding inside — and sometimes just outside — the data a company already has.

The distinction

Not analytics. The opposite direction.

Most work filed under "customer analysis" looks backwards at the customers a brand already has, and asks why they behaved the way they did. That is a useful discipline and it is not this one. A diagnostic runs the other way: it starts from demand nobody asked for and works out who produced it.

Diagnostic analyticsCustomer Diagnostics
QuestionWhy did this happen?Who else has this problem?
Looks atCustomers you haveCustomers you do not have yet
Starts fromA metric that movedDemand nobody can explain
OutputA root causeA group, and a cheap test against it
Fails whenThe data is incompleteThe anomaly gets explained away

It is also not a quiz. "Customer diagnostic" is sometimes used for an assessment built as a lead magnet — a scored questionnaire a prospect fills in about themselves. That is a marketing asset. This is an investigation run on a brand's own data, and the brand is the one being diagnosed.

The method

Strip the category. Follow the anomaly.

  1. 01

    Say what the product does

    Not what it is sold as. A sports drink is an electrolyte solution with no added sugar. The category is a distribution decision somebody made once; the underlying job is what decides who else could possibly want it.

  2. 02

    Find the demand nobody asked for

    Unexplained traffic, reviews describing a use case nobody planned, orders that do not fit. These arrive as anomalies, which is why they get explained away — every reporting system is built around the customer the company already decided it has.

  3. 03

    Separate a market from noise

    Same underlying problem or merely similar demographics. Arrived uninvited or responded to a campaign. Dense enough to reach each other or diffuse enough to need paid reach for every customer. Servable as it stands or only after a new product.

  4. 04

    Make being wrong cheap

    The point of the test is not to be right. It is to size the downside so being wrong costs less than the information is worth — which is what lets a team be wrong four times and still finish ahead of one that spent a quarter building certainty about a single option.

What it finds

The customer was already there.

Greater Than launched as a sports drink at eight dollars a bottle. The traffic spikes nobody could trace turned out to be breastfeeding mothers, tagging ten friends each inside private groups. Three years later a three-hundred-thousand-dollar-a-year business was doing ten million, profitably. The signal that started it was about twenty odd comments.

Questions people ask

The short answers.

What is Customer Diagnostics?
Customer Diagnostics is the process of finding overlooked customers and growth opportunities hiding inside — and sometimes just outside — the data a company already has. It starts from what a product actually does rather than the category it is sold in, and treats demand nobody can explain as evidence rather than noise.
How is Customer Diagnostics different from diagnostic analytics?
Diagnostic analytics asks why something that already happened, happened — why churn rose, why a campaign underperformed. It looks backwards at customers you already have. Customer Diagnostics looks for customers you do not have yet, starting from demand that arrived without being asked for.
Is this the same as customer segmentation?
No. Segmentation divides the customers you already know about into groups. Customer Diagnostics starts from behaviour you cannot explain and asks who is behind it. Segmentation sharpens an existing picture; a diagnostic is how you find out the picture was incomplete.
What data does it need?
Mostly data a brand already owns. Unstructured text first — support tickets, reviews, DMs, comments — then traffic that resists attribution. Structured reporting is organised around the customer a company already decided it has, so a buyer who does not match gets averaged into a column nobody reads.
How large does an overlooked group need to be to matter?
Smaller than most teams assume. Greater Than’s breakthrough started from roughly fifteen to twenty unusual customer comments. What decides it is not the size of the signal but whether the group shares the underlying problem and can reach each other.

Your next customers are already out there.

Find the opportunity before you spend more trying to reach it.