How Do Authorities Define Markets and Measure Dominance?

Market definition and dominance assessment decide competition cases before any conduct is discussed: draw the market narrowly and a 70% share appears; widely, and it dissolves to 20%. Understanding how authorities actually draw the line — and where the battles are won — is foundational for merger filings, abuse defence and complaint strategy alike. This is the analytical toolkit article of our dominance pillar in the Competition & Antitrust hub.

Why does market definition matter so much?

Because every structural presumption keys off it: dominance thresholds, merger safe harbours, notification analyses and fine calculations all consume a market share, and the share is an artifact of the boundary. Epic lost against Apple on market definition (mobile-gaming transactions, not iOS apps); the EU’s Google cases succeeded on narrow, layer-by-layer markets; Kroger/Albertsons turned on whether supercenters discipline supermarkets.

Its power is also its limit, and modern practice holds the two in tension: the EU courts insist definition is not an end in itself and effects can be shown across plausible boundaries, US agencies plead multiple alternative markets, and ex ante regimes (DMA, DMCC) deliberately sidestep the exercise with designation criteria. But for the working lawyer and economist, the discipline remains where cases are framed — and the side that controls the market question usually controls the narrative that follows.

How does the hypothetical monopolist (SSNIP) test work?

Start from the focal product and ask: could a hypothetical monopolist of it profitably sustain a 5–10% price rise? If enough customers would switch to defeat the rise, add the closest substitute and repeat, stopping at the smallest set over which the rise is profitable — that set is the relevant market. The test disciplines intuition by forcing the substitution question to be quantitative: how many customers, moving to what, how fast.

Implementation leans on critical-loss analysis (how much volume loss makes the rise unprofitable) against estimated actual loss (from diversion ratios, survey and switching data). Two classic traps: the cellophane fallacy — at an existing monopolist’s current (already supra-competitive) price, everything looks substitutable, so the test must run from competitive-level prices in abuse cases; and asymmetric markets — A may constrain B while B barely constrains A, producing different markets from different focal points. Geographic scope follows the same logic through transport costs, regulatory borders, language and shipment-pattern evidence (the Elzinga-Hogarty tradition, used cautiously).

What evidence do authorities weigh in practice?

A hierarchy that surprises newcomers. Internal documents rank at the top: board decks naming the competitors management actually watches, pricing committees reacting to specific rivals, win/loss analyses — authorities treat the company’s own map of competition as the best evidence of it. Customer evidence ranks beside it: procurement testimony about real alternatives moves decisions more than models.

Quantitative strands then corroborate: diversion ratios from switching data or surveys (who do your lost customers go to?), price-concentration and price-correlation studies, natural experiments (entry, exit, outages — what happened to volumes when a rival’s plant burned or a platform went down), bidding data in tender markets, and margin evidence. The craft is convergence — no single method is dispositive, and the losing side in most fights is the one whose story requires dismissing three independent evidence types at once. For companies, the standing implication: your documents are drafting your future market definition today; write market descriptions with the same care as financial statements.

How is dominance assessed once the market is drawn?

Share first — level, stability and trajectory: a stable 55% over a fragmented field reads differently from a volatile 55% trading share with a 30% challenger. Then the constraint analysis: entry and expansion barriers (sunk costs, scale and scope economies, network effects, switching costs, brand, regulation, key-input control, data advantages), rivals’ capacity to expand if the leader raises prices, and countervailing buyer power — concentrated, sophisticated buyers who can sponsor entry or self-supply can neutralise high shares.

Durability completes the test: dominance requires the position to be defensible over a horizon that matters (the EU thinks in years, not quarters). Modern additions include profitability evidence (persistent returns above cost of capital suggest weak constraint) and, in platform markets, tipping analysis — whether network effects have converted leadership into lock-in. The assessment is holistic by design; parties that treat it as a share-threshold checkbox concede the richest terrain, since most winnable dominance defences are constraint stories, not share disputes.

How do zero-price and multi-sided markets change the analysis?

Zero-price services (search, social, many apps) break the SSNIP’s mechanics — you cannot raise a zero price by 5% — so authorities substitute quality-adjusted variants: would users tolerate a small but significant degradation in privacy, ads load or features (SSNDQ)? The EU’s revised 2024 Market Definition Notice codifies this, alongside attention and data as the relevant currencies.

Multi-sidedness requires holding both sides in view: a platform’s power over advertisers is disciplined (or not) by its position with users, and single-side analysis misprices the constraint. Aftermarkets and ecosystems raise the boundary question vertically: is the market ‘smartphones’ or ‘iOS app distribution to iPhone owners’? Case law accepts single-brand aftermarkets where switching the primary product is costly and information is poor — the analytical bridge to the layer-level definitions that carried the Google cases. The honest state of the art: the toolkit adapts, contested case by contested case, and the regulatory bypass (designation without definition) exists precisely because litigation over these boundaries consumes years.

How should companies manage the market-definition battlefield?

Prospectively. Maintain a considered internal position on your plausible markets and shares — for threshold analysis, dominance triggers and litigation readiness — built on the same evidence authorities use, refreshed annually. Discipline documents: marketing’s “we have no real competitors” and strategy’s “category king” slides become exhibits; train teams to describe competition accurately (broadly when true, and with named constraints).

In live matters, invest early in the empirical record: switching studies, diversion surveys and natural-experiment evidence commissioned at complaint stage shape the authority’s first framing, which anchors everything after. And argue in the alternative without embarrassment — sophisticated parties run their best boundary and show their story survives the opponent’s boundary, because decisions increasingly do both. The same evidentiary work doubles for merger filings, where the Phase I framing contest is market definition under another name.

How does market definition differ between merger and abuse cases?

The forward/backward lens. Merger review asks how markets will behave after a structural change — definitions accommodate entry trajectories, pipeline products and closeness-of-competition analysis that can matter more than the boundary itself (unilateral-effects cases increasingly argue diversion, not definition). Abuse cases look backward at conduct in markets as they were, and the cellophane problem looms: current prices already reflect the dominance under investigation, so competitive-level benchmarks must be reconstructed.

Consequences follow for evidence and strategy. Merger practice tolerates ‘in any plausible market’ pleading; abuse decisions need a definite market to ground dominance findings that courts will test. And timing diverges: merger definitions are negotiated in weeks under deadline, abuse definitions litigated over years. Companies operating near dominance should keep both frames in their share book — the merger-review market you argued for in last year’s acquisition will be quoted back to you in next year’s abuse complaint.

What role do sector inquiries and market studies play?

They are market definition at wholesale: authorities map entire sectors — e-commerce, mobile ecosystems, cloud, groceries, fintech — outside any case, building datasets, questionnaire records and published framings that later cases inherit. The UK’s market-investigation regime can impose remedies directly from a study without any infringement; the EU and Türkiye use inquiry findings to open targeted cases with the analytical work pre-done.

Businesses treat inquiry questionnaires as compliance chores; they are actually framing contests. The definitions, share estimates and constraint narratives submitted in a sector inquiry become the authority’s baseline — and inconsistencies with later case submissions are read as opportunism. Participate strategically: senior review of responses, consistency with the share book, and voluntary context where the questionnaire’s framing misleads. The companies that shape the study shape the docket that follows it.

Frequently Asked Questions