Tipping Test (Winner-Take-All Conditions)
When you need this method
You invoke network effects and conclude that the market will end up belonging to one provider. Expensive decisions follow from that assumption: below-cost pricing, exclusive partnerships, speed ahead of margin. The assumption itself is never tested. If the market does not tip, you are funding a race that does not exist and end up without a position.
Approach
- 1Name both sides of the market and record, for each direction, how strongly one side attracts the other. Cross-side effects are usually positive but can be negative, for instance when readers want fewer ads.
- 2Estimate multihoming costs per side: acquisition, operation, training and the time a second platform consumes. For the first condition it is enough that these costs are high on at least one user side.
- 3Check whether the network effects are positive and strong precisely on the side with high multihoming costs. Only there does the size of the other side actually decide the choice of platform.
- 4Establish whether either side insists on special features. Where special needs exist, smaller differentiated providers survive, and the market does not converge.
- 5Read the result and state the consequence. All three conditions met means the market will be served by one platform, and the next question is whether to fight or to share it. If one condition is missing, differentiation is the more viable position, not a land grab.
- 6Only then set the pricing sides. Subsidize the side that is sensitive to price and quality, and charge the side whose demand rises most strongly with the growth of the other. The subsidy is wasted if the subsidized side can also transact with a rival's money side.
Typical application
A typical B2B SaaS in the building trades wants to build a marketplace between contractors and suppliers and plans to serve both sides free of charge for two years in order to take the market. The test brings the assumption down: suppliers already maintain their catalogue on several portals, so one more connection costs them almost nothing. On the side that creates the value for the other, multihoming costs are therefore low, and the first condition fails. Instead of a land grab, the team aims the offer at one trade segment with requirements of its own, charges contractors from day one, and subsidizes only those suppliers whose catalogue the contractors actually look for.
Limits and counter-indications
The test is a judgement about market structure, not a calculation. The three conditions are estimated, not measured, and multihoming costs in particular are hard to quantify. It only applies where two distinguishable user groups attract each other; ordinary vendor-customer businesses do not qualify, however freely network effects are invoked there. The conditions are also unstable: when the cost of a second connection falls, a market that had tipped opens up again, as happened in the game console market. And the test says nothing about who wins, only whether there will be a winner at all. The evidence comes mostly from large consumer and standards markets such as payment cards, consoles and operating systems; for small B2B niches there is almost no empirical work.
How to measure impact
Track three figures per side of the market: the share of users who are also active on competing platforms, the estimated cost of one additional affiliation, and the share of revenue that comes from special requirements. A rising multihoming share on the decisive side is the earliest sign that the tipping assumption no longer holds.
Related methods
Sources
- 1.Eisenmann, Parker, Van Alstyne: Strategies for Two-Sided Markets, Harvard Business Review 84(10), Oktober 2006, S. 92-101 (opens in a new tab) · Harvard Business Review (Volltext über KTH Royal Institute of Technology) · 2006-10 · academic and scholarly literature · describes the methodCarries the procedure verbatim: a networked market is likely to be served by a single platform when multihoming costs are high for at least one user side, network effects are positive and strong for exactly that side, and neither side has a strong preference for special features. It also carries the pricing rule with subsidy side, money side and marquee users, plus the warning that a subsidy is wasted if the subsidized side can reach a rival's money side. Limit: a practitioner-magazine article by academic authors, not peer reviewed; the evidence consists of selected cases such as DVD, payment cards and game consoles, not a sample. The common reading that multihoming costs must be high on both sides is not what the source says.
- 2.Rochet, Tirole: Two-Sided Markets: An Overview, Arbeitspapier, März 2004 (opens in a new tab) · IDEI Toulouse (Volltext über MIT, Kurs 14.271) · 2004-03 · academic and scholarly literature · supports the underlying mechanismProvides the economic reason why multihoming is the pivot. With platforms that are not interconnected, multihoming on at least one side is necessary for any gains from trade to be realized; and the less one side multihomes, the greater the platform's pricing power over the other side (the single-homing index). It also carries the marquee-user logic: those with strong pull on the other side are won through price. Limit: a formal model without empirical testing; the three conditions of the test do not appear there as a checklist, and the authors note that platform competition changes the price structure in ambiguous ways.
- 3.Rysman: The Economics of Two-Sided Markets, Journal of Economic Perspectives 23(3), 2009, S. 125-143 (opens in a new tab) · American Economic Association (Volltext über George Mason University) · 2009 · academic and scholarly literature · evidence of effectivenessIndependently names three factors that decide whether a market tips: whether standards can differentiate from each other, how easily agents can use several standards, and whether complementors can differentiate after choosing a platform. For the second factor there is empirical evidence: as the relative fixed cost of porting a game fell, titles spread across several systems and the console market became less concentrated. Limit: a survey article, not a test of the checklist; the empirical evidence comes from a single industry. The third factor cuts against the test's third condition and shows that the need for differentiation can work in either direction depending on which side of the market it sits.
Origin: Eisenmann / Parker / Van Alstyne