How to test whether clients will pay for a new service before you build it

Ask about the last time the client needed the job

Ask the client to describe a recent request, what triggered it, how they completed it and where they needed help. Look for the actual work and consequences. A general question about whether they like self-service is harder to interpret.

• When did you last need this result?
• How did you get it done, and who was involved?
• What information or judgement was missing?
• What did you use instead, and what did that cost in time or money?
• How often does this need return, and who approves spending on it?

Keep your explanation of the proposed service until you understand the existing job. The GOV.UK discovery guidance offers a broader method for understanding users and constraints before delivery.

Make the proposed content service concrete

Said sees an opportunity for content businesses to bring former clients back through a smaller ongoing arrangement. The firm may already hold useful context about each client’s business and competitors, which could support a self-service system.

That gives you a specific group and a starting hypothesis. Do former clients still have recurring work that the existing service no longer fits? Would using the context themselves help them get an acceptable result? How much expert review would they still need?

The current account establishes the idea, not customer demand or revenue. A useful test should be capable of showing that the offer needs changing or is unsuitable for the chosen group.

Offer a small version with a clear boundary

Choose one job and describe its inputs, output, review and support. A manually supported first version can help test the service before building every automation. Be clear with participants about what the version does and how their work is handled.

Make a concrete commercial offer with an agreed scope and terms. Record who accepted, who declined and why. A paid pilot provides different evidence from a free trial; both still need observation of useful work and the reason someone returns.

Avoid promising broad results from a small pilot. Agree the output the client can inspect and the support you can actually provide. The productisation worksheet helps describe that offer.

Keep four kinds of evidence separate

Add repeat use when the need recurs. A trial login or complimentary reaction does not establish retention. Record the interval so an absence of use can be interpreted against the actual frequency of the job.

Compare the arrangement with the cost of delivering accepted work. Include the effort to refresh client context and handle exceptions. Check whether the offer adds revenue or replaces existing work at a different fee.

Decide what to do from the observed evidence: continue, change the client group, narrow the job or stop the proposed route. Keep the next test small enough to answer the remaining question.

Plan the evidence before the test