📖 Glossary

The Double Jeopardy Law in Marketing — Definition & Evidence

The double jeopardy law says smaller brands suffer twice: fewer buyers who also buy less often. A foundational law of growth from Ehrenberg-Bass. Here's what it means.

·6 min read

Definition (≤25 words). The double jeopardy law states that lower-share brands suffer twice: they have fewer buyers, and those buyers buy the brand less often.

TL;DR. The double jeopardy law is one of the most robust empirical regularities in all of marketing. It says smaller brands are doubly disadvantaged: not only do they have fewer customers, those customers are also slightly less loyal — they buy less often and are more likely to also buy competitors. The pattern holds across categories, countries, and decades. Its revolutionary implication is that the difference between a big brand and a small brand is not loyalty — it is penetration: how many buyers you have. Growth, therefore, comes from reaching more buyers, not from making existing ones more devoted. It is the foundation of everything the Ehrenberg-Bass Institute argues about how brands grow.

What exactly does "double jeopardy" mean?

The name comes from the idea that smaller brands are penalised in two ways at once. Imagine two brands in the same category: a large one with 40% market share and a small one with 5%. The large brand has more customers — that is the obvious part. But the double jeopardy law adds a second, less obvious penalty: the small brand's customers also buy it less often. They are lighter buyers. They are more likely to buy other brands in the category too.

So the small brand is not just small because it has fewer people; it is small because its fewer people are also less frequent. Both metrics — penetration (how many buy you) and purchase frequency (how often they buy you) — are lower. That is the double jeopardy.

This is not a marginal effect that occasionally shows up. It is a strikingly consistent pattern, documented across hundreds of product categories, in multiple countries, over decades of consumer-panel data. Wherever researchers have looked — groceries, durables, services, financial products, media consumption — the law holds.

Where does the law come from?

The pattern was first observed in mass-media audience research by William McPhee in the 1960s: less popular radio programmes were not only listened to by fewer people, but those people also listened less often. McPhee called it "double jeopardy."

Andrew Ehrenberg and his colleagues — later institutionalised as the Ehrenberg-Bass Institute at the University of South Australia — then demonstrated that the same law governs brand buying. They formalised it mathematically through the Dirichlet model (Goodhardt, Ehrenberg & Chatfield, 1984), which predicts a brand's penetration, purchase frequency, repeat rate, and other metrics from its market share alone, with remarkable accuracy. The Dirichlet is the formal backbone of the law: it shows that the patterns we see in real markets are exactly what a statistical model of mostly-light, polygamous buyers would produce.

Byron Sharp made the law central to a wider audience in How Brands Grow (2010), where it underpins the argument that growth is a penetration game, not a loyalty game.

Why is this so counterintuitive?

Because it contradicts the story most marketers were taught. The dominant paradigm for decades was the loyalty model: grow by deepening the commitment of your existing customers, turning them into advocates, increasing their share-of-wallet, and locking out competitors through CRM, loyalty programmes, and retention mechanics. In that model, the big brand is big because its customers are fiercely loyal, and the small brand is small because its customers are fickle.

The double jeopardy law says the opposite. Loyalty differences between competing brands are small and largely fixed. Big-brand customers are only marginally more loyal than small-brand customers. What actually separates a big brand from a small one is penetration — the sheer number of people who buy it, even occasionally. The big brand's edge is breadth, not depth.

This is hard to accept because it demotes much of what marketing teams spend their time on. Loyalty programmes, CRM deepening, "customer love" campaigns — these feel like the high-impact work. The evidence says they have a low ceiling, because loyalty is not where the variation is. The variation — and therefore the growth — is in how many buyers you reach.

What does the evidence actually show?

Three lines of evidence converge on the same conclusion:

  1. The double jeopardy pattern itself. Across categories, lower-share brands consistently show both lower penetration and lower purchase frequency than the law predicts they should, were loyalty the differentiator.
  2. The Dirichlet model's predictive power. Given only a brand's market share, the model predicts its full set of behavioural metrics with high accuracy. This is powerful corroboration: the structure of real markets matches a model built on the assumption that buyers are polygamous and lightly loyal.
  3. The failure of the "niche premium" hypothesis. Marketers often believe small, premium niche brands defy the law — that they have few but fiercely loyal customers. Under scrutiny, these apparent exceptions usually dissolve. When you account for category and the "subscription" versus " repertoire" distinction, niche-brand loyalty lands close to what the law predicts for their size. They are not super-loyal; they are simply small.

A related pattern, the duplication of purchase law, reinforces this: the customers of any one brand are disproportionately the customers of every other brand too. Buyers are not exclusive. They spread their purchases across the category, which is why reaching the shared pool of light buyers — rather than defending an imaginary fortress of loyalists — is what moves share.

Where people get it wrong

  • "I'll beat it with a great loyalty programme." Loyalty programmes rarely shift purchase frequency enough to break the pattern. They can be useful for data and retention, but they are not a growth engine.
  • "Our niche brand has super-loyal customers." Mostly a myth. Measure loyalty properly and it almost always lands near what the law predicts for the brand's size.
  • "So loyalty doesn't matter at all." It matters — retaining customers is cheaper than acquiring new ones, and defection is costly. The point is that loyalty is not where the growth lever sits. You cannot grow your way out of smallness by loving your few customers harder.
  • "The law means I shouldn't bother with existing customers." No — it means your growth strategy should prioritise reach and penetration, while you still manage retention as table-stakes.
  • Treating it as a licence to ignore the brand. Reaching more buyers requires mental availability and distinctive assets. The law tells you what drives growth (penetration); it does not excuse dull, unrecognisable marketing.

In practice — what the law means for strategy

  1. If you're a smaller brand: compete on reach, not loyalty. Do not try to out-devote the leader. Get into the consideration set of more category buyers — build distinctive assets, grow mental availability across category entry points, and expand penetration. That is your only real path up.
  2. Track the right metrics. Make penetration and share of search your growth scoreboard, not retention rate or customer-satisfaction scores. Retention is a cost to manage, not a growth lever to pull.
  3. Reach the light and non-buyers. Your heaviest buyers are already yours. The people who buy your category rarely are where the growth is — and they are the hardest to reach, which is why continuous, broad reach matters.
  4. If you're the leader: defend your penetration. The law means competitors cannot easily out-loyal you — but you can lose penetration if you let mental availability decay or go dark. Your risk is complacency, not disloyalty.
  5. Use the Dirichlet as a benchmark. Model what your metrics should look like for your share. Deviations — higher frequency than predicted — are worth investigating, but they are rare and usually small.

So what?

The double jeopardy law is not a curiosity; it is the central empirical fact about how brands grow. It tells you that the difference between the brand that wins and the brand that loses is not the intensity of its customers' love, but the breadth of its customer base. That single insight reorients an entire marketing strategy away from loyalty theatre and toward the unglamorous, proven work of reaching more buyers, more often, with brands they can actually remember. It is, in short, the fundamentals — which is the whole point.

Evidence & sources

  • Sharp, B. (2010). How Brands Grow. Oxford University Press — the double jeopardy law and its implications for growth strategy.
  • Goodhardt, G.J., Ehrenberg, A.S.C. & Chatfield, C. (1984). "The Dirichlet: A Comprehensive Model of Buying Behaviour." Journal of the Royal Statistical Society, Series A, 147(5), 621–655 — the mathematical model predicting brand metrics from share.
  • Romaniuk, J. & Sharp, B. (2016). How Brands Grow 2. Oxford University Press — further evidence on duplication of purchase and the stability of loyalty.
  • Ehrenberg-Bass Institute — foundational and ongoing research on the law across categories and markets.
  • McPhee, W. (1960s) — the original observation of double jeopardy in mass-media audience behaviour.

Definition type: DefinedTerm. Reviewed: 2026-07-09. Written by Andrei Tarnovski for Boringly Effective. Fundamentals over fashion.