Skip to content
Primores AI wiki

Double Jeopardy Law — Why Smaller Brands Get Hit Twice

Double Jeopardy Law

TL;DR: Smaller brands suffer twice: they have fewer buyers AND those buyers buy slightly less often. The empirical pattern is remarkably consistent across categories, countries, and decades. The implication is uncomfortable for “build loyalty” strategies: loyalty doesn’t vary much across competing brands of similar size — penetration does. Brand growth tracks customer-base expansion, not loyalty deepening.

What It Is

The Double Jeopardy Law is an empirical regularity first observed in the 1960s and formalized by Andrew Ehrenberg and colleagues. It states that brands with smaller market share have:

  1. Fewer buyers (lower market penetration), AND
  2. Slightly less loyal buyers (lower purchase frequency among those who do buy)

A small brand is “hit twice” — losing on both width and depth. The two effects compound: market share = penetration × purchase frequency, so smaller brands lose ground on both factors simultaneously.

The pattern has been observed across:

  • Consumer packaged goods (washing powder, shampoo, toothpaste, breakfast cereal, biscuits)
  • Durables (cars, appliances)
  • Services (banks, insurance, telcos)
  • B2B / industrial (ready-mix concrete, office supplies, business services)
  • Media (newspapers, TV networks, TV programs, radio stations)
  • Politics (party affiliation, voting behaviors)
  • Soap operas (literally — same pattern)

This is what makes it a “law” rather than a heuristic: Ehrenberg and his colleagues documented the same pattern across hundreds of categories, multiple countries, and decades of data. It’s about as close to a fundamental empirical law as marketing has.

The Empirical Anchor

From Sharp’s How Brands Grow, the washing powder example (UK 2005, Kantar WorldPanel data):

BrandMarket share (%)Annual penetration (%)Purchase frequency
Persil22413.9
Ariel14263.9
Bold10193.8
Daz9173.7
Surf8173.4

The pattern: penetration varies a lot (17%-41%, ratio of ~2.4x); purchase frequency varies very little (3.4-3.9, ratio of ~1.15x). Market share follows penetration almost entirely; loyalty barely moves.

The same shape holds in shampoo (UK 2010, US 2005), ready-mix concrete (West Yorkshire UK), and every category Ehrenberg-Bass has tested. A specific quote from Sharp’s analysis: “82% [of UK Advertising Effectiveness Award submissions] reported large penetration growth, 6% reported both penetration and loyalty growth, and only 2% reported loyalty growth alone.”

Why It Holds

The mechanism Sharp argues for, supported by behavioral data:

  1. Light buyers dominate any category. Most buyers in any category are light buyers who buy occasionally. Heavy loyalists exist but are fewer than commonly assumed.
  2. Light buyers are roughly equally available to any brand. Their purchase decisions are driven by what comes to mind first (glossary/mental-availability) and what’s easy to acquire (physical availability), not by deep brand commitment.
  3. Therefore market share depends on capturing light buyers — and the brand with broader reach captures more of them.
  4. Heavy buyers do exist, but they’re roughly proportional to market share — bigger brands have proportionally more heavy buyers, but heavy buyers don’t cause the size; they’re a result of larger penetration distributing across the loyalty curve.

This means “loyalty marketing” — trying to deepen the relationship with existing buyers — runs into a mathematical ceiling. You can’t out-loyalty your way past Double Jeopardy. The lever for growth is reaching new buyers, most of whom will be light occasional purchasers.

Implications That Cut Against Marketing Conventional Wisdom

Several common marketing claims become questionable under Double Jeopardy:

  • “Build a loyal core and cherish them” — your loyalty levels are mostly determined by your size, not by your loyalty programs. The lever isn’t there.
  • “Heavy users are 80% of revenue, focus on them” — they’re not 80%; the 80/20 doesn’t hold the way the cliché claims, and even where heavy users matter, they’re not where growth comes from.
  • “Differentiate to stand out” — Sharp argues differentiation doesn’t drive Double Jeopardy patterns; penetration does. See glossary/distinctive-assets for the alternative.
  • “Loyalty programs drive growth” — Sharp dedicates a whole chapter (Ch 11) to arguing they don’t, and the data supports him.

The reframe: growth comes from making the brand easy to think of and easy to buy, for as broad an audience as possible. Penetration is the lever; loyalty follows mathematically.

Implications for Content and AI-Era Marketing

Several useful translations:

For content strategy: designing content for engagement depth (heavy-user retention) hits the same mathematical ceiling. Reach-broadening content that recruits new audience members is the growth lever, not depth content for existing followers. This connects to glossary/weak-ties — light-buyer recruitment requires inter-cluster diffusion, which requires bridge-traversal (in pre-algorithmic networks via real social weak ties; in algorithmic-feed environments via synthetic weak-tie bridges).

For AI visibility: Double Jeopardy’s analog in AI-search is real — a brand that’s only well-represented in a small slice of training data and a few queries will surface less often and with less context. Brands with broad mention density across authoritative sources surface more often AND with richer context. Same “hit twice” pattern, different substrate.

For TRIPS: picking AI use cases by where time is spent (penetration of pain) tends to produce broader, more-frequent value than picking by where impact is theatrical. Same penetration-over-depth principle.

Where the Law Doesn’t Apply (Cleanly)

Honest limits:

  • Niche brands that have deliberately limited distribution — Sharp argues most “niche brands” are actually small brands that haven’t grown rather than purposefully-niche brands. But genuine niche brands (luxury, regulated, geographically-bounded) can deviate.
  • Solus loyalty contexts — categories where buyers are genuinely committed to one brand for structural reasons (e.g., enterprise software with high switching cost, certain B2B contracts). Penetration math still operates but loyalty effects are stronger.
  • Subscription / membership categories — where the underlying purchase is renewal, the loyalty dimension is the dominant axis by definition. Sharp’s Part 2 (Romaniuk & Sharp 2016) addresses this.
  • Very early-stage categories — before Double Jeopardy can stabilize, niches and outliers may dominate. The law manifests once a category has matured.

Reconciliation with Primores Worldview

Double Jeopardy reinforces several Primores priors:

  • Prior #5 (substance > popularity) — engagement metrics within strong-tie clusters reflect cluster reception, not the broad penetration that drives Double Jeopardy growth. Sharp’s data confirms that loyalty within a base is roughly determined by the base’s size, not by content quality optimization for that base.
  • Prior #6 (most AI value is unglamorous optimization) — Double Jeopardy is itself an unglamorous truth: growth is mostly about being thought of more often by more people, not about hero campaigns or deep emotional commitment. Same flavor of “the boring lever beats the glamorous lever.”

It partially complicates prior #4 (niche authority > broad TOFU) at the brand-building layer — see marketing/brand-vs-content-layers for the explicit reconciliation. The short version: niche authority wins for AI-citation, broad reach wins for mental availability; both layers matter.

Key Takeaways

  • Smaller brands lose twice: fewer buyers AND those buyers buy slightly less often.
  • Penetration varies a lot; purchase frequency varies very little within a category.
  • Loyalty programs and “build a loyal core” strategies hit a mathematical ceiling — you can’t out-loyalty your way past your size.
  • The lever for growth is broader reach to light buyers, not deeper relationships with heavy buyers.
  • Empirically observed across hundreds of categories, multiple countries, and decades.
  • Reframes “differentiation drives growth” — Sharp argues penetration does, distinctiveness aids penetration via glossary/mental-availability.

Sources

  • Sharp, B. (2010). How Brands Grow: What Marketers Don’t Know. Oxford University Press. — Chapter 2 (“How Do Brands Grow?”) is the canonical chapter; the empirical tables in this entry are drawn from there.
  • Ehrenberg, A. S. C., Goodhardt, G. J., & Barwise, T. P. (1990). Double jeopardy revisited. Journal of Marketing, 54(3), 82–91. — The pre-Sharp empirical synthesis.
  • McPhee, W. N. (1963). Formal Theories of Mass Behavior. Free Press. — The original observation in mass-media data.
  • Romaniuk, J. & Sharp, B. (2016). How Brands Grow: Part 2. Oxford University Press. — Extends the law to emerging markets, services, durables, B2B, and luxury.