Market Sizing and Category Economics: Questions and Common Misconceptions

No. A market size estimates demand within a defined boundary at a stated time or under a stated scenario. A forecast adds assumptions about how that demand may change. A present addressable revenue pool, a five-year adoption scenario, cumulative spending, and total economic impact are different measures. A number without its measurement date and unit cannot answer whether the category exists now, may develop later, or creates broader social value.

The distinction affects executive action. A product team deciding whether to launch a narrow offer needs current evidence about eligible buyers, willingness to pay, implementation friction, and alternatives. A long-range strategy discussion may consider conditions that could expand the boundary. The future view can inform exploration, but it should not be presented as current demand or inserted into a sales plan without the intermediate adoption and capacity assumptions.

No single total should settle a material decision unless its definition closely matches the question and its method can be evaluated. Published estimates may count software, services, infrastructure, productivity value, or some combination. They may use different geographies, buyer segments, currencies, base years, and category rules. Two totals can both be internally coherent while neither describes the market a specific company can reach.

Treat a published total as one item in a source register. Recover the original methodology, covered population, inclusions, exclusions, and forecast logic. Classify it as a direct estimate, an adjacent signal, an upper bound, or a value-pool reference. If the method is unavailable, use the number for orientation only. Authority of publication does not remove the need to test fit between the source and the decision.

Total addressable market should represent revenue available if the defined offer served every eligible buyer inside its category boundary. It should not automatically include all labor associated with the problem, all infrastructure used to deliver the product, and all adjacent software. Those pools may help explain buyer value or ecosystem structure, but adding them together can double count the same economic activity and obscure what the company actually sells.

TAM can be useful as a ceiling for a specific revenue model. It is less useful when the category definition is so broad that nearly every organization or technology purchase qualifies. A good test asks whether two independent reviewers would classify the same buyer and expenditure consistently. If not, narrow the category rule, specify the funded job, and identify the revenue unit before refining the calculation.

Serviceable available market applies limits such as geography, regulation, product capability, integration coverage, customer size, workflow readiness, and commercial packaging. Serviceable obtainable market adds company-specific reach, sales capacity, implementation throughput, retention, capital, and time. SOM is therefore not an arbitrary small percentage of TAM. It is a model of what the organization could plausibly acquire and serve under declared conditions.

These layers should change as evidence changes. A new connector may expand technical serviceability but not procurement readiness. A partner channel may expand reach but strain implementation capacity. A more repeatable deployment may improve obtainable revenue without changing TAM. Keeping the filters explicit helps leaders identify whether the next investment should improve category evidence, product coverage, distribution, delivery, or economics.

A workflow may reduce preparation time or accelerate a handoff, but released hours do not automatically become payroll reduction or incremental revenue. The buyer may use the capacity for quality, resilience, backlog reduction, or work that was previously deferred. Those outcomes can matter, yet they require observation. A value model should state the mechanism by which an operational change could influence a financial or strategic result and avoid claiming the result before it is measured.

Value capture also depends on alternatives and bargaining power. Buyers retain much of the benefit, implementation partners may capture some, and suppliers receive input revenue. A vendor price must remain credible against switching cost, risk, budget ownership, and substitute offers. Applying an unexplained capture rate to a large labor pool produces a precise-looking revenue estimate without showing why a buyer would approve that payment.

Vendor revenue must be read beside model usage, data access, infrastructure, storage, monitoring, human review, support, implementation, sales, and compliance costs. Some costs vary with each run, some with customer complexity, and others with the maturity of the product. A category can show growing revenue while providers absorb expensive exceptions or services that make repeatable delivery difficult.

Profit is also a company outcome, not a market-size layer. Different providers may serve the same revenue pool with different architectures, channels, prices, and cost structures. The category model can test plausible unit economics, but it cannot infer company profitability from demand alone. Keep gross revenue opportunity, contribution after direct delivery, and broader operating result as separate lines with separate assumptions.

Top-down analysis begins with a documented population or spending pool and applies filters that approximate the target category. It can reveal whether a bottom-up estimate exceeds a credible ceiling and can provide context about adjacent budgets. Its weakness is distance from the buying unit. A broad technology total may conceal workflow eligibility, organization readiness, purchase timing, and price, which are often the variables a launch decision needs most.

The responsible method records every filter and explains why it belongs. A percentage borrowed from a different geography or buyer segment should be labeled as a proxy, not an observation. The analyst should avoid multiplying several uncertain filters without showing the compounded effect. A small change in each rate can produce a large change in the result, so sensitivity matters more than decorative precision.

Bottom-up analysis counts eligible buyers or workloads and connects them to a price or usage model. It is often more useful for product, distribution, and capacity decisions because the assumptions resemble commercial operations. Its weakness is sample bias. Interviews, pipeline records, or pilot activity may overrepresent interested, reachable, or unusually mature organizations and may not generalize to the full population.

Reconciliation does not mean averaging the two answers. Compare their boundaries, units, and time periods, then investigate the largest gap. The top-down model may include buyers the bottom-up filters correctly exclude. The bottom-up model may reveal an overlooked segment or an implausible price. The disagreement becomes a research question. Averaging would hide the question without improving either source.

Hypothetical top-down model: begin with 20,000 organizations, assume 30 percent fit the target size and industry boundary, 40 percent have the qualifying workflow, and 25 percent are ready during the period. The resulting modeled population is 600 organizations. At a hypothetical annual price of 60 units, the addressable annual revenue is 36,000 units. Every percentage and price is illustrative, not observed.

Hypothetical bottom-up model: count 450 qualifying workflows across a defined account list, assume 70 percent map to an accountable budget owner, and apply a hypothetical price of 80 units per workflow. That yields 25,200 units. The difference is not resolved by taking 30,600. The top-down model counts organizations, while the bottom-up model prices workflows and may include more than one workflow per buyer or exclude unreachable buyers.

First normalize the models. Determine whether the offer is priced per organization, per workflow, or through a mixed contract. Check whether the account list represents the same geography and segment as the population source. Then test workflow count per eligible buyer, readiness, and price using source-backed interviews, procurement evidence, disclosed contracts where available, or a bounded offer experiment with appropriate consent and review.

The result may remain a range. That is acceptable if the range preserves the material uncertainty and supports the decision. Record observed values separately from inferences, retain rejected sources, and note what evidence changed each version. If research cannot support the critical price or adoption input, the company should limit the capital, forecast, or public claim attached to the model rather than fill the gap with confidence.

Founders and executives can use the answers to challenge category claims before allocating capital. Investors can test whether a narrative moves improperly between economic layers. Product, finance, and go-to-market leaders can align around the same buyer, workload, price, period, and serviceability filters. Researchers should preserve source definitions, confidence, disagreement, and refresh dates so later users do not inherit a number without its conditions.

No method guarantees adoption, price, growth, share, customer outcome, or profitability. Surveys can measure stated interest rather than purchase. Pipeline can reflect a company's channel rather than total demand. Public disclosures can combine categories. Forecasts depend on future behavior. Material legal, accounting, investment, competition, and public-claim decisions require qualified review. A good answer narrows uncertainty; it does not pretend uncertainty has disappeared.

Within a verified configuration, OmegaOS can help connect the question, Hermes source packet, model assumptions, owner review, scenario decision, related work, and later observations. That continuity can make it easier to see whether a change came from new evidence, a revised definition, or a strategic preference. It cannot turn a proxy into primary evidence or make a broad category claim safe for publication without appropriate review.

Use the smallest operating loop that serves the decision. Record the disputed assumption, commission bounded research, attach source references, route consequential interpretations to accountable owners, and compare the resulting decision with later signals. Verify current access, connectors, and data custody before use. The proportionate OmegaOS bridge is traceability and learning around the model, while executives and specialists retain authority for the commercial conclusion.