Choose a first segment without sales history by comparing three buyer groups through problem evidence, buyer access and equal paid offers before committing.
Short answer: Choose no more than three candidate segments that share the same core problem, then give each an equal test: five recent-problem interviews, five conversations with budget owners and five paid offers at a viable price. Start with the segment showing the strongest combination of repeated costly behaviour, direct buyer access and paid acceptance. Do not choose from total market size or survey enthusiasm when no transaction evidence exists.
No sales data means you cannot compare retention, realised margin or acquisition cost. It does not mean you have no evidence. Workarounds, budgets, buying roles and paid test response can reveal which segment deserves the first operating cycle.
The danger is compensating for missing transactions with an elaborate score built from assumptions. A precise spreadsheet does not make the inputs true.
Use the Three-Round Segment Tournament
The Three-Round Segment Tournament gives each candidate the same evidence allowance. A segment advances only when it passes the current round.
- 1. Problem: Question: Does the loss recur and cause sacrifice?; Evidence allowance: Five event-based interviews; Advance condition: At least three show the same costly mechanism
- 2. Buyer: Question: Can you reach someone able to act?; Evidence allowance: Five buyer conversations; Advance condition: A clear owner and approval route repeat
- 3. Payment: Question: Will buyers accept this scope and price?; Evidence allowance: Five equal paid offers; Advance condition: Payment or one repeated, fixable barrier
These counts are practical comparison rules, not statistical proof. High-value, regulated or varied markets need deeper work. Equality matters: one segment should not receive 30 approaches while another is judged after two.
Limit the candidates to one core problem
Comparing restaurants that need stock control with consumers who want language tuition is a comparison of business ideas, not customer segments. Keep the core problem and result substantially stable so differences in buyer behaviour become visible.
Candidate segments can differ by industry, size, workflow volume, trigger or buyer role. For example, a warranty-claim administration service can be tested with appliance, bicycle and furniture retailers because the operational mechanism is similar.
My view is that three candidates are enough for the first decision. More segments dilute interviews and allow you to keep researching instead of offering. Park the rest until one tournament finishes.
Use behavioural evidence before market estimates
In round one, reconstruct the last occurrence. Ask what happened, who intervened, what time or cash was used and what remained unresolved. Look for existing spending, overtime, lost contribution or accepted risk.
Do not ask which features a hypothetical service should contain. You are comparing the problem mechanism. Record:
- Frequency over a defined period
- Cost or consequence per occurrence
- Current workaround
- Person carrying responsibility
- Trigger that makes action timely
If the same segment contains several unrelated mechanisms, narrow it before advancing.
Published market estimates can provide context but often use categories broader than your buyer. Treat them as indicative and verify definitions, geography and date. A small identifiable group with repeated behaviour can be a stronger start than a huge category with no direct access.
Test access to the actual buyer
A user interview does not prove you can reach the payer. In round two, ask who approves a purchase at the proposed level, what evidence they need and when decisions occur.
Track founder hours needed to produce five genuine conversations. The segment reached through eight hours of direct work differs from one requiring 40 hours, even before sales. Note gatekeepers, trusted intermediaries and whether your current credibility meets the standard.
Do not use an existing friendship as representative access. Label relationship-assisted conversations and include prospects with no personal reason to help.
Privacy and direct-marketing rules vary by location, buyer and channel. Use current official guidance and qualified local advice before collecting personal data or contacting individuals for commercial purposes.
Make equal paid offers
Use the same core outcome and comparable price logic in round three. Segment-specific wording is sensible, but do not give one group a free trial and ask another for £500.
The price must be viable. Calculate direct cost and founder delivery time first. A low test price can make a weak segment appear responsive while proving demand only for underpaid work.
Record the first failed point: no problem, no timing, no authority, no trust, price below value or delivery mismatch. One paid customer is valuable but not an automatic winner. Compare how many qualified offers were required and whether the work produced positive contribution.
Where the offer handles customer data, regulated decisions or safety-critical work, define responsibility and obtain appropriate professional advice before a paid test.
Worked example: Raj's warranty-claim administration service
Raj has no customers for a warranty-claim administration service. He compares independent appliance retailers, bicycle shops and furniture retailers. Each receives the same evidence allowance.
In round one, four of five appliance retailers show repeated monthly claim delays, three of five bicycle shops show the mechanism, and two of five furniture retailers do. Appliances and bicycles advance.
In round two, Raj reaches the owner or operations buyer in all five appliance businesses after six founder hours. Reaching five bicycle-shop buyers takes nine hours because several owners work mainly on repairs. Both still advance, but appliance access is stronger.
Raj offers a bounded one-month claim-administration test for £300. Customer-specific cash costs are £24, delivery takes six hours and he values his time at £28 an hour.
Economic surplus per paid test is: £300 minus £24 minus (6 × £28) = £108.
Of five offers, two appliance retailers buy and one bicycle shop buys.
- Appliance retailers: Paid tests: 2; Delivery surplus: 2 × £108 = £216; Buyer-access hours: 6
- Bicycle shops: Paid tests: 1; Delivery surplus: 1 × £108 = £108; Buyer-access hours: 9
- Furniture retailers: Paid tests: 0, did not advance; Delivery surplus: £0; Buyer-access hours: Not tested further
Raj should begin with appliance retailers. They show more repeated problem evidence, cheaper access and twice the paid response under equal terms. The result does not prove a large market or recurring retention. It earns that segment the next ten paid offers and a repeat-purchase test.
Raj must define data access, customer communication and the boundary between administration and product or legal judgement. Requirements vary by sector and jurisdiction.
Do not collapse evidence into one arbitrary score
Weighted scores can hide fatal flaws. A segment should not compensate for zero buyer access with a large market-size score. Use gates:
- The problem and consequence repeat.
- A reachable buyer can approve spending.
- The intended price supports both customer value and your contribution.
- Paid evidence justifies a longer test.
If two segments pass equally, choose the one that produces a paid answer faster and uses less irreversible cash. You can test the other after the first operating cycle.
Keep notes on uncertainty rather than assigning fake precision. “Buyer role unclear in three of five firms” is more useful than giving authority 6.4 out of 10.
Choose from a fixed review date
Set the decision date before starting. Without one, a founder can keep adding interviews until the preferred segment wins.
At review, compare only evidence gathered under equivalent conditions. If one segment encountered a temporary event, record it and decide whether to rerun that round rather than blending incomparable results.
The selected segment remains a hypothesis. Its first jobs will reveal realised time, corrections, payment and retention. Update the decision after ten paid customers or one complete buying cycle.
Related guides
Run the tournament in 14 days
On day one, define the core problem, result, viable price and three candidate segments. Prepare five suitable prospects per segment for the first round.
Then act in this order:
- Complete event-based problem interviews and eliminate weak mechanisms.
- Speak with five actual buyers in each surviving segment.
- Make five comparable paid offers to each survivor.
- Calculate delivery surplus and access hours for every payment.
- Select one segment on day 14 and give it the next ten-offer cycle.
No sales history should make your first decision smaller and faster. It should not make it theoretical.
Frequently asked questions
Can I choose a segment from competitor customers?
Competitors can reveal recognised demand, buyer language and common alternatives, but their customers may buy because of capabilities or relationships you do not have. Study public positioning, reviews and buying contexts without taking private information. Then run your own problem, buyer and paid-offer rounds.
Do not assume a crowded segment is bad or that an underserved one is good. The exception is a formal market with published procurement and supplier requirements, where competitor presence may provide stronger structural evidence. You must still meet the standard and test your own access.
What if nobody accepts the paid offer in any segment?
Diagnose the first repeated failure before adding segments. The problem may be weak, the buyer wrong, price unsupported, trust insufficient or the offer unclear. If all groups reject for different reasons, your core problem or result may be too broad. Change one element and rerun with a fixed evidence allowance.
The exception is a long buying cycle where payment could not reasonably occur within the test. Use an earlier commercial milestone involving the budget owner, but do not treat a meeting as equivalent to a sale. Set a later decision date appropriate to the real process.
Should I pick the segment with the biggest total market?
No. Choose the segment with repeated valuable behaviour, reachable buyers and paid acceptance at viable economics. Total market size can support future capacity decisions after you verify definitions and access. A smaller segment may produce faster learning and enough customers for your first target. The exception is a model with large unavoidable infrastructure or development costs. It genuinely needs a larger market, but that raises the demand evidence and capital required before proceeding. A broad estimate alone does not fund or validate that commitment.
Can surveys replace the first interview round?
They can collect comparable facts when respondents are well qualified and questions concern recent behaviour. They are weaker for reconstructing workarounds, authority and consequence because follow-up is limited. Use a short survey to screen, then conduct event-based conversations with suitable buyers.
Avoid asking whether people would buy a hypothetical idea. The exception is a simple high-volume consumer behaviour where a larger structured sample adds value, but observed purchase or a paid offer remains stronger demand evidence. Check privacy and research obligations for the data you collect.
What if I have personal experience in only one segment?
Treat that as an access and competence advantage, not automatic demand. You may ask better questions and reach buyers faster, but you can also normalise problems outsiders would challenge. Give the segment the same rounds and terms. For unfamiliar candidates, obtain enough operational understanding to avoid unsafe or irrelevant offers. The exception is regulated or specialist work that requires formal competence. Personal familiarity cannot substitute for qualification, permission or professional responsibility. Confirm requirements before including such a segment in a paid test.
What if two customer segments finish level?
Do not break the tie with market size or personal preference. Identify the uncertainty most likely to change the operating decision, such as repeat purchase, payment delay, delivery exceptions or buyer access, then give both segments one equal extra round focused only on that variable. Keep the offer and evidence allowance comparable.
If the tie remains and both produce viable contribution, choose the one with lower irreversible cash and easier direct access for the next cycle. The exception is a strategic or legal constraint that makes one segment unavailable, which should have been treated as a gate rather than scored as a small disadvantage.
Comments (0)