spacepolicies.org
Market Sizing & Segmentation
The Number That Was Never Really a Number
A pitch deck announces a trillion-dollar market. A research report cites it. A government strategy echoes the citation. A competitor’s deck picks up the figure. Within two years the number has been rounded up twice, cited dozens of times, and embedded in the assumptions of investment decisions that together commit substantial real capital to the segment. At no point in this chain has anyone examined how the trillion dollars was computed, what it actually included, or what fraction of it any specific company could plausibly capture. The number was never really a number; it was a rhetorical object that acquired the appearance of empirical authority through repetition.
Space has a particular vulnerability to this pattern. The segments being sized are nascent; historical data is thin; hype cycles inflate near-term projections; government spending blurs the line between market demand and policy-driven allocation. Into this environment, Market Sizing and Segmentation introduces a discipline: a small set of rigorously distinguished concepts — TAM, SAM, SOM — combined with transparent methodology and explicit assumption disclosure. The point is not to produce a smaller number than the decks. The point is to produce a number the analyst can defend under scrutiny.
A Practitioner Tradition, Not a Theoretical One
The methodological lineage is less academic than most strategic tools. TAM-SAM-SOM emerged from venture-capital and strategic-planning practice over the 1990s and 2000s, codified in countless business-school cases and investor decks rather than in foundational theoretical texts. Its direct intellectual antecedents are the market-research discipline that developed through the mid-twentieth century (Nielsen, Gallup, the early industry-analyst houses) and the strategic-planning frameworks associated with McKinsey, BCG, and similar consultancies, which taught a generation of planners to decompose markets by segment, geography, and customer type.
The discipline’s maturation paralleled venture capital’s growth as an industry. Early-stage investors needed a consistent language for comparing opportunities, and the TAM-SAM-SOM hierarchy supplied it — Total Addressable Market (the full revenue opportunity if every potential customer bought), Serviceable Addressable Market (the share that current technology, regulation, and infrastructure permit), and Serviceable Obtainable Market (the fraction an actor could realistically capture given competitive dynamics). The distinctions are not theoretically profound; they are operationally crucial, and collapsing them is the first discipline the method enforces.
Space-specific market sizing accumulated as a subfield only over the past two decades, through the work of specialized analyst houses — Euroconsult, Bryce Tech, the Satellite Industry Association, NSR, and a handful of government bodies producing their own series. Their disagreements on definitions, boundaries, and methodology are themselves analytically revealing: a segment where reputable analysts produce estimates that differ by a factor of two or three is a segment whose numbers cannot be averaged without loss of information. The method treats those disagreements as data.
What the Method Actually Does
The characteristic move is definitional discipline followed by triangulated estimation. Before any number is produced, the boundary of the market must be specified precisely: what product or service category is included, what geography, what customer types, what time horizon, and — a question that matters more in space than almost anywhere else — whether government procurement is counted as market demand or treated separately. These definitional choices determine which headline number will result, and different analysts reach different totals mostly because they made different boundary choices silently. The method forces the choices to be explicit.
Once the boundary is set, TAM estimation uses two approaches in parallel. The top-down approach takes a larger aggregate — for instance, total global satellite-services revenue — and allocates a fraction to the segment under study based on structural ratios. The bottom-up approach builds the estimate from unit economics: number of potential customers times average revenue per customer, or number of addressable use-cases times price point. The approaches should converge. Where they do not, the divergence is the analysis — it reveals that one of the approaches is making an assumption the other exposes. Cross-reference with published estimates from specialized houses adds a third leg to the triangulation, not to average the estimates but to identify and explain the disagreements.
Segmentation runs orthogonally through the sizing exercise. The market must be decomposed — by application, by orbit regime, by customer type, by geography, by value-chain position — and each segment sized individually. The aggregate TAM is less informative than the distribution across segments, because the distribution reveals where demand actually concentrates and where the growth is coming from.
Finally, growth projections are constructed using scenario bands — base, optimistic, pessimistic — with every major assumption documented. Point forecasts in nascent space markets have a poor historical track record; scenario bands acknowledge the uncertainty honestly and let the reader see what would have to be true for each band to materialize.
The method sits next to other market-analysis tools. Porter’s Five Forces reads competitive structure without quantifying it; market sizing quantifies scale without interpreting competition. The two together form a fuller reading than either alone. Value-chain analysis identifies where in the segment margin accumulates; sizing tells the analyst how large the overall pie is across all nodes. Technology-readiness assessment determines whether the SAM narrowing from TAM is actually achievable in the time horizon examined.
The Method at Work
Take a generic commercial debris-removal market sizing, constructed with illustrative figures. The TAM, computed top-down by multiplying the number of trackable objects by an average removal cost at plausible future scale, lands around four billion dollars per year. The TAM is a number in the sense that it can be cited, but by itself it is close to useless for decision-making; it describes a revenue opportunity that depends on technology not yet operational at scale, regulatory frameworks not yet in place, and customer willingness to pay not yet demonstrated.
SAM narrows the field. Excluding orbits where no removal operation is currently licensed, objects for which no removal technology is proven, and operators who face no pressure — regulatory, insurance-driven, or reputational — to remove their own debris, the reachable market shrinks to perhaps nine hundred million dollars per year. The reduction from TAM to SAM is not a mistake or a pessimism; it is the correction of an aspirational figure into a figure that corresponds to the conditions currently in place.
SOM narrows further. The currently funded contracts plus announced procurements over a realistic capture window reduce to roughly one hundred twenty million dollars per year. That figure, when decomposed, reveals its composition: approximately seventy percent government-procured through agencies such as ESA and JAXA, thirty percent commercial and driven largely by insurance and reputational pressure on large satellite operators. The driver behind the SOM growth is emerging debris-mitigation regulation; the binding constraint is the absence of proven multi-target removal capability at scale; the segment’s maturity reads as nascent.
The non-obvious insight the method produces is this: the business case for commercial debris removal, at today’s SOM, is almost entirely a government-contracting business, not a commercial-demand business. A company pitching itself as a commercial debris-removal provider is pitching a future that depends on insurance pricing and regulatory tightening materializing on schedule. The pitch may be correct, but the story it tells is misleading — a government-contractor today hoping to become a commercial player tomorrow is a different investment proposition from a commercial player with occasional government revenue, and the SAM-SOM analysis forces the distinction. The TAM figure alone would have concealed all of this.
Where It Shines, Where It Zoppica
The method is strongest when the segment has enough disclosed commercial activity to support bottom-up estimation, when multiple published sources exist for cross-reference, and when the analyst is disciplined enough to disclose disagreements rather than average them away. Done well, it produces numbers that investment committees can defend under scrutiny and that policymakers can use as a basis for calibrated decisions.
Its weaknesses are substantial and should be disclosed to readers of any sizing exercise. Space-market data is fragmented; different analyst houses use different definitions, and an uncritical analyst can produce any desired total by choosing sources with matching definitional preferences. TAM estimates for nascent markets — in-orbit servicing, orbital manufacturing, lunar resources — are highly speculative and assumption-dependent; honest reporting marks them as indicative bands rather than point estimates. Top-down estimates risk circularity when they rest on industry reports that were themselves built on optimistic projections; cross-checking with bottom-up is not optional. Government spending blurs the market-demand boundary in ways no defensible methodology can wish away; the honest response is to segment government and private flows rather than fold them together. And hype cycles have produced a long track record of near-term projections that exceeded outcomes by factors of two or more; historical calibration of the analyst’s own prior estimates is the only defence.
The method does not capture competitive dynamics or strategic positioning; it is purely quantitative, and pairing with Porter’s Five Forces, value-chain analysis, or platform-ecosystem analysis is necessary for any complete market read. It does not distinguish demand that responds to price from demand that responds to regulation; the segmentation must be carried out carefully to separate these drivers. And it is a poor fit for segments where the product category is genuinely novel and the customer base does not yet exist — foresight methods such as Horizon Scanning or Three Horizons Analysis are better starting points than sizing in those cases, with sizing applied only once the addressable customer category has been identified.