Opportunity discovery and market sizing · Lesson 4

Size an opportunity without false precision

6 min read

Lesson objectives

  • Size an opportunity without false precision: execute the named specialist task with a checked output.
Reading tools

After discovery interviews, a manager may ask, “How much can we sell?” Do not multiply an entire population by the product price. A specialist distinguishes the accounts matching a segment definition, the subset actually reachable and serviceable, and the number obtainable within a specified period using a particular sales channel and delivery capacity. Each number can change for a different reason. Bottom-up sizing starts with countable accounts and visible exclusion rules, then passes through assumptions that can be tested. It does not attach one arbitrary percentage to a huge market total.

Build an account table before announcing a market value. For each row, record the counting unit, source and date, geography, whether the account is active, whether it is already a customer, and whether the current product can serve it. Remove duplicates when the same branch appears in two directories, but keep branches separate if they buy independently. The resulting list might contain 800 accounts apparently matching the segment definition. Eight hundred is a research ceiling given the quality of that list, not evidence that 800 buyers exist. If forty percent of contact records are old, access to all accounts cannot be assumed.

Inspect denominators before comparing percentages. Fifteen percent here applies to the 120 potentially fitting accounts, not the original 800 organizations. A colleague who says “fifteen percent of the market will convert” would expect 120 customers instead of eighteen. That is about six and two-thirds times as many customers because earlier filtering stages disappeared. Likewise, 480 reachable accounts is a planning assumption, not evidence that 480 people replied; contact data might be invalid. Display the denominator beneath each rate and preserve the version used for the decision, so a target does not turn into its own justification.

Open an uncertainty range instead of reporting one exact figure. Suppose reachability ranges from forty to seventy percent, relevant-problem prevalence from fifteen to thirty-five percent, and conversion from eight to twenty percent. A coherent low case is 800×0.40×0.15×0.08=3.84, reported as roughly four customers rather than 3.84 actual contracts. A high case is 800×0.70×0.35×0.20=39.2, or roughly thirty-nine customers. The range is wide because the assumptions have not been verified. Do not use its high end as a staffing plan. Use it to choose a test that reduces uncertainty. If the current team can activate only twelve new customers a year, the eighteen-customer base scenario exceeds capacity. Delivery is capped at twelve unless the responsible manager approves a capacity increase, even if demand reaches eighteen.

Test sensitivity by changing one assumption at a time. Hold base reachability and conversion constant. Raising problem prevalence from twenty-five to thirty percent increases potentially fitting accounts from 120 to 144 and expected customers from eighteen to 21.6, about twenty-two. Instead, cleaning contact data so reachability rises from sixty to seventy percent yields 800×0.70×0.25×0.15=21 expected customers. Which uncertainty is cheaper to reduce? List cleaning may be quick; learning problem prevalence needs varied interviews. The point is not to choose the assumption that creates a larger number. Choose the test that improves the next decision at a reasonable cost.

Check the estimate independently against team activity. Could eighteen deals be supported by the number of conversations the team can hold? If one specialist can conduct five useful interviews per week across forty working weeks, the ceiling is 200 interviews before allowing for nonresponse. A plan requiring deep interviews with all 480 reachable accounts would be infeasible with that capacity. Not every account needs an interview: public information can filter a list before interviewing a varied sample. This is not another market-sizing equation. It is an operational consistency check against a number the organization cannot possibly reach with its present process.

A specialist's report states a range and conditions: “The fictional list contains 800 organizations. Our base case suggests eighteen customers in one year before a capacity ceiling of twelve. Problem prevalence is the largest unknown. We will contact thirty varied accounts, hold ten interviews, and revise that assumption.” Submit any request for extra delivery capacity separately from the demand estimate. The model does not guarantee growth or return. If the list's source or deduplication method cannot be checked, say that the estimate is unsuitable for a major spending decision. Clean the base rather than polishing decimal places.

Keep expansion revenue from an existing account separate from acquisition revenue in a new-customer opportunity estimate. A current customer may enlarge a subscription, but its sales effort and probability differ from a new prospect's. If eighteen contracts are expected to close throughout the year, they may not all produce 2,400 during that same financial year. A deal starting in month ten yields only three months of service within that year if revenue is recognized monthly. Label 43,200 as the nominal annual subscription value of the hypothetical group, and ask finance to confirm recognition treatment before using it in financial forecasts. The specialist explains the distinction; the specialist does not set accounting policy.

Schedule recalculation when evidence arrives, not simply at month end. Thirty contacts can update reachability, ten interviews can update problem prevalence, and a set of proposals can begin to reveal conversion. Do not silently combine stage rates drawn from different samples or periods. Multiplying them can look precise even when each rate belongs to a different market. A new model version shows changes in numerator and denominator and their reasons, while preserving the earlier version for forecast review.

Lesson summary

Size an opportunity without false precision with stated assumptions and a reviewable output.