Opportunity discovery and market sizing · Lesson 2

Segment and select a serviceable market

6 min read

Lesson objectives

  • Segment and select a serviceable market: execute the named specialist task with a checked output.
Reading tools

“We will sell to small businesses” does not yet tell a specialist whom to contact or what to test. A segment is a group of accounts whose problem, buying setting, and service needs are similar enough to test one offer. Size alone does not make a segment useful. A large market may be hard to reach or may require changes the delivery team cannot make. Start with the customer's job, behavior, and serviceability. Use industry labels and population estimates to locate research candidates, rather than treating those labels as a substitute for a buying decision.

First define the unit you count: company, branch, user, or purchase decision. If a head office buys software for all its branches, each branch is a place of use but there may be only one commercial account. If branches sign separate contracts, the denominator changes. Never add companies and branches into one total. Choose variables that actually distinguish commercial decisions: severity of the problem, access to the buyer, fit with the current product, delivery burden, and likely contribution margin. A long descriptive profile adds little if it changes neither offer nor channel nor delivery cost.

Take a fictional appointment-booking software provider. The current product manages one site's calendar and customer reminders, but does not coordinate complex schedules across dozens of locations. The team considers independent clinics, electronics repair stores, and multi-site training offices. Desk research and internal lists yield preliminary, not fully verified counts: 400 reachable clinics, 900 repair stores, and 150 training offices. In a few conversations, clinics describe appointment cancellations. Repair stores describe unpredictable walk-ins, which booking alone may not solve. Training offices have a strong need to unify several sites. It would be unjustified to select repair stores solely because 900 is the largest figure.

Do not let a precise score disguise weak evidence. If the clinics' problem-severity score comes from only three conversations, mark confidence as low or use a plausible range. Were that score to fall from 4 to 2 after further interviews, the total would fall by 0.50 to 3.60. Clinics would still lead this particular comparison, but their advantage would narrow. If product fit instead fell from 5 to 2 because a required integration proved costly, the score would fall by 0.75 to 3.35. The decision would now be sensitive to a product and delivery check. Sensitivity analysis asks which changed assumption could change the decision; it is more useful than a generic “risk” box with no follow-up.

Compare alternatives customers already use. A clinic may rely on a paper diary, phone messages, or an existing system, each of which changes the value of reminders and the willingness to pay. A repair store might address waiting time by adding a staff member rather than buying booking software. A multi-site office might reject a system that cannot allocate capacity across locations. “Do nothing” is sometimes the main competitor. The US Small Business Administration's market-research guide suggests examining demand, size, saturation, prices, and direct and indirect competitors. Those are prompts for investigation; they do not validate the fictional counts or scores in this case.

After choosing a segment, define it so the team can use it: “Independent, single-site clinics with three to ten employees, recurrent scheduling needs, an accessible operational buyer, and no need for medical-system integration during the trial.” This is a test definition, not a permanent exclusion of larger clinics or a regulatory conclusion. The team can now select a list of thirty matching accounts, hold ten varied interviews, and record how often the problem occurs and what serving it costs. If most clinics actually need an integration beyond current capacity, narrow or stop the trial before proposing a product change without approval.

The specialist may build this scoring table, recommend a test segment, and measure response. Allocating a large budget, changing the product roadmap, or making claims about sensitive health data requires the relevant decision owners. If an interview reveals sensitive information, follow the organization's handling rules and keep it out of a widely circulated segment sheet. Segmentation matters because it reduces the number of assumptions tested at once. That makes a failed or successful offer easier to interpret and turns a vague target market into a bounded learning decision.

Check list coverage before interpreting response rates. Suppose the team targets thirty clinics, but only eighteen email addresses are valid. Six replies and three interviews should be recorded as separate events. It is fair to say that six of the eighteen valid addresses replied. It is not fair to claim that one third of all clinics in the market are interested. Clinics with better digital records may be easier to find and more inclined to buy software than others. This is a sampling limit that belongs next to the conclusion before spending is expanded.

A lower score does not mean a segment has no future value. Multi-site offices can remain on a watch list if the product later gains the necessary coordination capability, but their trial results should not be mixed with clinic results. When new evidence arrives, change only the criterion justified by that evidence, recalculate the score, and explain whether the ordering changed. The table then becomes a reviewable decision record rather than a number that overrides customer reality.

Lesson summary

Segment and select a serviceable market with stated assumptions and a reviewable output.