What Is Game Theory and How Can Businesses Use It?
You set a price, launch a product or enter a market, and the outcome turns out to depend less on your decision than on what everyone else decided at the same time. That interdependence is the subject of game theory, and it is the reason ordinary planning tools underperform in competitive situations: a plan assumes the world holds still while you execute it.
The formal version involves mathematics. The commercially useful version mostly does not. What it requires is the discipline of thinking one step further than feels natural, which is a habit rather than a technique.
The basic apparatus
The field was formalised by John von Neumann and Oskar Morgenstern in the 1940s and extended by John Nash. Four elements describe any strategic situation.
Players. Whoever's decisions affect the outcome. In business this is wider than the competitor list: suppliers, distributors, regulators and large customers all make moves that change your payoff.
Strategies. The options available to each player. Not intentions, options: what they could actually do given their capacity and constraints.
Payoffs. What each player gets under each combination of moves. This is where most analysis goes wrong, because payoffs are assumed rather than investigated.
Interdependence. Your result depends on their choice as well as yours, which is what separates a strategic situation from a planning one.
Dominant strategies, and why they are rare
A dominant strategy is one that is better for you regardless of what anyone else does. Where one exists the analysis is finished: take it.
They are rarer than they look. Most apparent dominant strategies turn out to depend on assumptions about the other side that were smuggled in unexamined. Testing for one is still worthwhile, because it is quick and because finding that your best move changes with theirs is itself the useful information.
Equilibrium, and why it is not the same as a good outcome
A Nash equilibrium is a combination of choices where no player can improve their position by moving alone.
The commercially important point is that an equilibrium can be bad for everyone in it. Consider a market where four competitors have all discounted to the point where margins barely cover cost. Every one of them would be better off if all four raised prices. None can raise alone without losing volume to the other three. That is a stable and unpleasant resting point, and it will not resolve through effort or determination.
Recognising that you are in one changes what you do. The move is not to try harder within the equilibrium but to change the game: compete on a dimension the others cannot match quickly, serve a segment with different requirements, or alter your cost base so the same price yields a different margin.
The prisoner's dilemma, which is usually a price war
The best-known structure in the field is also the most common in commerce.
Two competitors each choose to hold or cut price. If both hold, both keep healthy margins. If one cuts and the other holds, the cutter takes share. If both cut, both keep their relative positions at materially lower margins.
Each firm's individually rational move is to cut, whatever the other does. So both cut, and both end up worse off than if neither had. The logic is airtight and the outcome is destructive, which is precisely what makes it worth recognising in advance.
The practical response is not to move first into the trap, and to compete where matching is slow. A discount can be replicated within a day. Contractual guarantees, deep integrations, accumulated switching costs and service commitments cannot.
Sequential situations, and working backwards
Most competitive decisions are not simultaneous. Someone moves and someone else responds.
Sequential situations are analysed from the end. Start with the last move and ask what the other party will rationally do once you have acted, then work back to the present.
A firm considering entry into a market occupied by a large incumbent, for example, should not ask whether the incumbent would prefer to keep the market to itself, which is obvious, but whether cutting prices to deter entry would be worth it to them once entry has occurred. If the incumbent's margins are high and the entrant's target segment is small, retaliation may cost the incumbent more than tolerating a niche competitor. Their threat, however loudly made, is then not credible.
Credible commitment
Which brings us to the concept that does the most practical work.
A commitment matters only if it would be irrational to abandon it, and that requires it to be costly to reverse. Saying you will never discount is not a commitment. Publishing a price list with a public no-discount policy is closer, because reversing it is visible and embarrassing. Signing a long-term supply contract at a fixed price is a commitment in full, because breaking it costs money.
The counterintuitive implication is that removing your own options can strengthen your position. A supplier who genuinely cannot go below a floor price negotiates from a stronger place than one who can and says they will not. Threats and promises that cost nothing to withdraw change nobody's behaviour, including the behaviour of people who believe them.
Repetition changes everything
A one-off interaction rewards taking the immediate advantage. A repeated one makes reputation an asset.
In markets where the same firms meet repeatedly, on tender after tender, in the same accounts, at the same conferences, behaviour that would be sensible once becomes expensive. Aggressive undercutting on one deal invites retaliation on the next three. A reputation for consistent pricing is met more predictably and, over a series of interactions, more profitably.
This is why competitive advice that ignores the time dimension is frequently wrong. The question is not whether a move wins this deal but what it teaches the other side about the next ten.
Co-opetition and the value net
Adam Brandenburger and Barry Nalebuff made the most useful business adaptation of the field in Co-opetition. Their central observation is that competition and cooperation are not opposites and usually occur simultaneously.
Their Value Net adds a fourth party to the familiar three. Alongside customers, suppliers and competitors sit complementors: parties whose presence makes your offering more valuable. A payments provider makes an e-commerce platform more valuable; a consultancy makes a software vendor more valuable.
The reason this matters is that the same organisation often occupies several roles at once. A large platform can be your distribution channel, your complementor and your direct competitor in the same quarter. Treating it as purely one of those produces a strategy that fails on contact with the other two.
A worked situation: the tender
Public procurement is unusually close to a formal sealed-bid auction, which makes it one of the clearest places to apply the analysis.
Players: the buyer, the credible bidders, and any incumbent supplier.
The buyer's payoff is stated in the evaluation criteria and is not always what they are actually optimising for. A weighting that awards heavy credit for a narrowly defined reference, or for membership of a local register, tells you something the stated criteria do not.
Competitors' strategies depend on their capacity. A firm with idle capacity will bid lower than one operating at the limit, because their marginal cost of the work is different.
Working backwards from the award produces the most valuable conclusion, which is frequently that you should not bid. Preparing a losing submission has a real and recurring cost, and the discipline of deciding not to compete is worth more over a year than incremental improvements to the documents.
How to run the analysis in an hour
- List the players. Include the ones who are not competitors: the regulator, the distributor, the customer with enough volume to move your pricing.
- State what each is optimising for, and be explicit that this is a guess. Write down the guess.
- List each player's real options, constrained by what they are actually capable of.
- Sketch the payoffs for the two or three combinations that matter. Directional is enough; precision here is false comfort.
- Look for the equilibrium. Where does this settle if everyone acts in their own interest?
- Ask whether the equilibrium is acceptable. If not, the question becomes how to change the game rather than how to win it.
- Test your assumptions about their payoffs. Ask what would have to be true for their observed behaviour to be sensible, then check whether any of it is.
What the method cannot do
It does not predict. It identifies which outcomes are stable and which moves are self-defeating, on assumptions that are often wrong.
The assumptions fail in three regular ways. Competitors hold information you do not. They optimise for things you have not guessed, including internal targets unrelated to profit. And they sometimes act irrationally, or rationally against an objective nobody outside the company can see.
Used as prediction, the method produces confident and incorrect forecasts. Used as a discipline, it makes your own assumptions explicit and rules out moves that fail under every reasonable reading of the other side. The second use is the valuable one.
The point
Game theory earns its place in commercial work as a habit of thought: name the players, guess what they want, look one move further than instinct suggests, and ask where the situation settles.
Most of its value comes from three recognitions. That a stable outcome can be bad for everyone in it. That commitments only matter when they are expensive to reverse. And that a move which wins today teaches the other side something about tomorrow.
At go:lofty we use this kind of analysis in competitive positioning and bid decisions, where the most useful answer is often that the game is not worth entering.
