Test the business you actually plan to run

A test is useful when its result can change a decision. If every encouraging result leads to the original launch and every discouraging result leads to more testing, the plan has become difficult to question.

In Choosing the Next Bet, Vijay Sheth considers an expansion trial using spare capacity in an existing operation. Paying customers and satisfactory delivery costs are real findings. They do not establish that the same customers can support a permanent location with its own fixed costs. That distinction lets the smaller business be worthwhile without making it a rehearsal the company is obliged to outgrow.

Choose the uncertainty before the test

A trial can answer whether customers will pay a particular price, whether a route can be served efficiently or whether a supplier can deliver when needed. Those are different questions. Choose the one whose answer could alter the next commitment.

The SBA’s market research guidance separates demand, location, competition and pricing, and distinguishes broad existing information from research specific to a business’s customers. An industry report and a paid order can both be useful without answering the same question.

For an invented example, a repair company is considering a second workshop. It first collects inquiries from 40 nearby businesses. That shows interest from people willing to inquire. It has not established which work they need, the price they will accept or when their existing service agreements permit them to change suppliers.

For the next step, suppose six customers buy a defined service at the intended price, with the work completed by the current workshop during unused hours. The company now has evidence of paid demand and some experience delivering the service. It still lacks evidence that six customers, or the next six, justify the second workshop.

Record what the trial borrowed

Spare workshop hours have a particular advantage: they already exist. A permanent location may need another lease, equipment and supervision before it earns its first sale. A test can be profitable because it uses those existing resources, while the proposed standalone version would lose money at the same volume.

The opposite mistake is to reject the profitable smaller version because it has not proved the original idea. The current workshop might continue serving these customers well. The next decision could be to improve that arrangement, extend it carefully or leave it at a size that fits.

The last question belongs in the trial’s result. An attractive new line can borrow the experienced employee who made the old service reliable. Its own figures may improve while the combined company becomes harder to operate.

A second site cannot spend the first site's money twice

Consider a separate, entirely fictional framing studio reviewing a proposed satellite room on November 5, 2026. An earlier popup showed that people would pay for the service. The owner now needs to decide before the new room's November 6 payment deadline. Nothing below changes the repair-company trial above.

The company has $5,000 in cash. Of that, $4,000 is already assigned to the existing site's customer work and the owner's pay through November 22. Only $1,000 is unassigned. The existing site's work is funded separately and cannot be moved to the satellite or counted as satellite revenue.

Twelve different prospective satellite customers have made written offers for $150 jobs, six for pickup on November 13 and six on November 20. Each job needs two attendant hours and $30 of materials. They would pay on pickup. No offer is accepted, no deposit has arrived and no additional demand is assumed. The branch below assumes the owner accepts all twelve before opening, after establishing funding and capacity; declining now breaks no accepted customer promise.

The two-week room arrangement costs $600, including every premises and utility charge for November 9–22. Setup costs $500. Both are payable November 6, with no refund assumed. All $360 of materials must be bought November 9. The paid attendant costs $600 each week for twenty-four available hours, including every supplied employer cost. Twelve production hours a week cover the six jobs, including checks and customer handoffs; the remaining paid hours cover opening availability. These are complete invented costs, not wage or tax benchmarks.

Supervision needs a separate transfer. The owner already has thirty-six hours of old-site work inside a forty-hour paid week. The satellite requires six more hours a week. A confirmed, suitably qualified cover person can take two hours of old-site work for $80 a week, including every associated employer cost. That leaves the owner with thirty-four old-site hours plus six satellite hours: forty in total. The cover is available for the exact duties and dates. Without it, the plan needs forty-two owner hours, not a motivational speech.

The owner's existing pay covers that forty-hour week and is inside the protected $4,000. The satellite attendant and cover person are separate paid people; neither supplies both jobs at once. There are no other receipts, costs, taxes, equipment payments or financing within this limited two-week cash model.

Money needed before the first satellite collection

Scroll across to compare the available money with each dated requirement. Use the arrow keys when the table area is focused.

Even if all pickups and collections happen as supplied, the twelve jobs bring $1,800 against $2,820 of total payments: $600 room, $500 setup, $360 materials, $1,200 attendant and $160 cover. They use $1,020 of cash over the period. Starting with the unassigned $1,000 would leave a $20 shortfall at the end, while failing much earlier at the payment deadline. A forecast with only the final balance would understate the first funding gap.

On November 5, no extra funding has been committed. The supported decision is to leave the room and customer offers unaccepted while looking for a genuinely funded alternative. Moving $460 from the protected old-site money would leave that site's commitments short; labeling it an internal transfer creates no new company cash.

The useful next question is precise: can another arrangement fund the early payments and release the required supervision without weakening existing service? More inquiries alone would not answer it. Keep the current operation and the proposed one on separate demand, resource and funding records before combining them. The transition-funding example follows an established handover; the founder-absence exercise examines authority and follow-through. Here the unresolved choice is whether this next location can meet its own dated requirements.

A test record that can change the decision

Write the following before starting. The fields concern the evidence this test should produce; the decision notebook can hold the wider choice.

Decision the result will inform

Name the next commitment, such as taking a lease, hiring a second team or continuing with current capacity.

Question and observable result

State what must happen to count as evidence. Distinguish an inquiry, an accepted offer, a completed paid job and a repeat purchase.

Conditions of the test

Record the price, service, customer group, location and period. Identify favorable conditions unlikely to persist.

Resources borrowed

List existing staff time, equipment, relationships and facilities used. Say what would have to be added or replaced in the planned operation.

What the test cannot establish

Identify the difficult dependency removed by the test, and the evidence or funding still needed for it.

Decision date and alternatives

Choose when to examine the results and what would support proceeding, revising, remaining smaller or stopping.

When a small test removes too much

Macau of the Americas complicates the instinct to test everything cheaply. A destination needs enough of a complete visit for someone to want to make the trip. Testing one attractive component can omit the transport, accommodation and opening dates on which its appeal depends.

Some questions cannot be settled by a miniature version of the final undertaking. That does not make evidence irrelevant. It means identifying which uncertainty can be reduced before the large commitment, and which still has to be carried. A narrower first stage is useful when it finishes a real job or tests a consequential question, rather than merely looking affordable.

When the results arrive, keep the finding specific: six customers bought this service, at this price, under these conditions. Decide whether to extend that operation, change it or take the next commitment. A test that cannot alter any of those choices needs a better question.