Business model innovation: how to test a new way of earning revenue

Business model innovation changes how a company creates value, delivers it or earns revenue. An established SME should test the uncertain parts of that change before committing to a full transition. Define what must be true, choose an experiment that can reveal it, set an evidence threshold and limit the resources at risk. Customer enthusiasm is useful, but a workable model also needs acceptable prices, repeatable delivery and a manageable transition from the existing business.

What makes a change a business-model experiment?

A new machine can improve an existing operation without changing the commercial model. Offering customers guaranteed output from that machine, with the supplier retaining ownership and maintenance responsibility, changes payment, investment and risk together. Similarly, a consultancy moving from projects to a subscription must change more than its invoice schedule. It needs a continuing benefit, an account-management process and capacity to meet the recurring commitment. Describe those changes explicitly before deciding what to test.

Strategyzer distinguishes improvement of an established model from the search for an unproven one. The distinction matters because historical performance may provide a sound base for one and little evidence for the other. An SME should still budget an experiment, but an elaborate forecast cannot establish that customers want the new offer. Use the current business as a source of knowledge while recognising where the proposed model breaks with its experience.

Which assumption should be tested first?

Select a claim whose failure would materially change the decision. A manufacturer considering an equipment subscription might depend on customers accepting a minimum term, sufficient utilisation and the ability to refurbish returned units. Testing a redesigned brochure would do little to resolve those questions. Rank uncertainty by its effect on the model and by how cheaply it can be investigated. Sometimes a review of maintenance records resolves the largest risk before any customer pilot is necessary.

The Strategyzer Test Card asks teams to make their hypothesis, test, measurement and success threshold explicit. An SME can apply that structure in a short working document. Replace “customers want flexibility” with a claim about a defined customer group accepting particular commercial terms. Decide what evidence would make the team proceed, revise the offer or stop before seeing the result. Otherwise a disappointing experiment can be reinterpreted as success after the money has been spent.

What evidence is stronger than an encouraging conversation?

Interviews help explain a problem and the language customers use to describe it. They are weaker evidence of purchasing behaviour when the discussion imposes no cost or commitment. A request for a quotation, provision of operational data or a paid trial can reveal more, depending on the buying process. The experiment should reach someone who can authorise the purchase. Approval from an enthusiastic user may not resolve procurement, security, finance or budget constraints.

Prototype only what is needed to answer the question. The UK Government Service Manual’s alpha guidance uses prototypes to test ideas before a live service is built. Applying that principle commercially, a supplier might demonstrate a sample report before developing the entire reporting platform. Make the prototype’s status clear to participants. A manual pilot can establish whether an output is useful, but its labour cost must remain visible when assessing a future automated operation.

How should a paid pilot be designed?

Choose a scope that exposes ordinary difficulties without overwhelming the existing operation. A hypothetical industrial supplier could test an output-based service with a small number of suitable sites, a defined operating envelope and a fixed review date. Record installation effort, failures, maintenance, customer acceptance and payment behaviour. A trial performed only at a particularly cooperative customer can still teach something, but the team should not assume it represents less prepared buyers or more demanding sites.

Set the commercial terms as carefully as the technical boundaries. State the fee, included work, access required, ownership of equipment or data and the end-of-pilot arrangements. A discounted pilot measures acceptance of that discount, not necessarily the intended long-term price. Where a customer receives exceptional support, record the time and test whether it would be available after launch. The commercial lesson may be that the service works for a narrower customer group than originally imagined.

What budget and decision thresholds make the test useful?

Limit cash expenditure and management attention separately. An experiment that costs little externally can still consume the founder and delay work for paying customers. In a hypothetical pilot, £12,000 of setup expenditure and £3,000 of monthly delivery cost over three months require £21,000 before receipts. If three customers each pay £1,500 per month for all three months, receipts are £13,500 and the pilot’s net cash requirement is £7,500, before other costs and timing differences. That loss may buy useful knowledge, but it should be deliberate.

Evidence thresholds should reflect the uncertainty and the cost of the next commitment. Three pilot renewals would be informative for a specialised service, yet they would not establish demand across an entire international market. An acceptable technical result might justify further customer research rather than a factory investment. Ask what the next spending decision requires, and avoid combining every favourable result into a broad claim that the model has been validated.

How do you decide whether to stop, adapt or expand?

Separate observations from interpretation. The Strategyzer Learning Card makes that distinction explicit before identifying the next action. “Two customers renewed” is an observation. “Customers prefer subscriptions” is a wider inference that may require further evidence. Preserve cancelled trials, objections and unusually expensive cases alongside successes. These can reveal the eligibility conditions under which the offer is commercially sensible.

Expansion also needs a transition plan. Decide who supports existing customers, whether current incentives favour the old offer and how inventory, contracts or staff capacity will change. A recurring-revenue model can delay cash that previously arrived on delivery, even if lifetime revenue appears attractive. Expand in stages that allow those effects to be observed. The business model canvas can record the revised system, while the guide to manufacturing recurring revenue examines the service obligations behind selling output. A test has done its job when it supports a better commitment, including a decision to stop.