The dilemma
Illustrative scenario
Priya is brand director at a regional sparkling-water company stocked in about a third of the supermarkets in its home state. Its buyers are vocal: reviews are glowing, and a customer survey says most of them "love" the brand.
Next year's budget has room for one big bet. A CRM agency pitches a points-based loyalty app with push reminders, and its deck shows that members of a small pilot bought more often than non-members. The sales team wants the same money spent on winning shelf space in two more grocery chains and on in-store sampling, which would put the brand in front of shoppers who have never tried it.
The conflicting metric: the pilot's repeat rate looks impressive, but the brand's penetration (the share of households buying it at least once a year) has barely moved in three years. Which lever is more likely to grow the brand?
The short answer
Double Jeopardy is the pattern in which smaller brands in a category suffer twice: they have far fewer buyers than larger brands, and those buyers also buy them slightly less often. The name is generally traced to the sociologist William McPhee, who described it in research on popularity in the early 1960s; Andrew Ehrenberg and colleagues later documented it extensively in brand purchasing data.
How strong is the evidence?
Empirical regularity
A pattern repeatedly observed in market data. It describes what usually happens; the exact numbers vary.
This is an empirical regularity in buying behavior, not a psychological mechanism. Across many repeat-purchase categories, when brands are ranked by market share, most of the difference in share comes from penetration (how many people buy the brand in a period). Loyalty measures, such as purchase frequency or share of category requirements, step down only slightly as brands get smaller.
Ehrenberg, Goodhardt and Barwise (1990) reviewed the pattern across product categories and explained why it is to be expected when brands are broadly similar substitutes competing for the same buyers: it does not require small brands to be doing anything special wrong. Researchers at the Ehrenberg-Bass Institute have since reported the same pattern among B2B brands.
Conditions: the regularity is clearest in stable, competitive categories. Deviations are informative rather than embarrassing. A brand with much higher loyalty than its size predicts may be a niche brand serving a distinct need, or may benefit from lock-in such as contracts, switching costs or rivals being hard to find.
What it does not say: that loyalty is irrelevant or that loyalty programs never pay. It says a brand's loyalty is usually about what its size predicts, so the larger growth gap is normally in the number of buyers.
Investigating an example
Documented study
Double Jeopardy across product categories
Ehrenberg, Goodhardt and Barwise brought together purchase-panel evidence from a range of product categories and found the same shape repeatedly: brands with lower market share had lower penetration and somewhat lower average purchase frequency among their buyers. The loyalty gap between big and small brands was consistently much smaller than the penetration gap. Their reading is that a small brand rarely stays small because its customers are unusually disloyal; it stays small because fewer people buy it.
Limits: the evidence comes mainly from established repeat-purchase categories in fairly stable markets, measured over a fixed period. New launches, fast-changing categories, strongly differentiated niche brands and contract-based markets can deviate. The paper describes a pattern in buying data; it does not evaluate any specific loyalty program.
A published study or record, linked below. Its limits are stated.
Choose your move
You are in the dilemma above. Which move would you make? Open each one to see its trade-offs.
A Fund the loyalty app for existing buyers.
Members may well buy more often, but pilots tend to attract the heaviest buyers, who would have bought anyway, so part of the uplift is selection. Double Jeopardy suggests a brand this size has limited room to lift loyalty far above what its penetration predicts, and the cost of rewards lands mostly on sales you already had.
B Spend on new distribution and sampling, and track penetration as the main goal.
Most consistent with the pattern, because penetration is where small brands lag most. But new listings cost fees and margin, sampling may produce one-off trial, and the payoff takes longer to show up than a loyalty pilot's repeat rate.
C Split the budget: mostly availability, plus a small loyalty test against a randomized holdout.
Hedges the bet and produces real evidence on the program. The risk is spreading money too thin for either to show a measurable effect, so agree in advance which metric decides next year's split.
No move wins in every business. The law tells you what to check, not what to choose.
How to apply it
- Benchmark against rivals. Collect penetration and average purchase frequency (or share of category requirements) for your brand and its main competitors, from the same data source and the same period.
- Plot loyalty against penetration. Brands usually line up on a gently rising curve. Note where your brand sits compared with brands of similar size.
- Investigate deviations. Unusually high loyalty for your size may signal a niche or a lock-in; unusually low loyalty may point to availability problems or a brand people use only occasionally for variety.
- Write growth targets in buyers. Express the plan as additional buyers per year, and treat loyalty metrics as something to monitor rather than the main lever.
- Test retention spend properly. Compare a randomly invited group with a randomly held-out group on incremental purchases, not members with non-members.
What to measure
Annual penetration, purchases per buyer, share of category requirements and the number of first-time buyers. Guardrail: judge any loyalty initiative on incremental purchases against a randomized holdout, never on a comparison of members with non-members.
Worked example
| Brand | Market share | Annual penetration | Purchases per buyer per year |
|---|---|---|---|
| Brand A (leader) | 48% | 40% | 4.0 |
| Brand B | 27% | 25% | 3.6 |
| Brand C | 15% | 15% | 3.3 |
| Brand D (ours) | 7% | 8% | 3.0 |
| Brand E | 3% | 4% | 2.8 |
Illustrative numbers, not data from a real category. Penetration varies tenfold across brands while purchases per buyer vary by less than half. For Brand D, lifting purchase frequency all the way to the leader's level would add about a third to its volume; doubling its buyers would roughly double it.
Where it breaks
- Category matters. The pattern is clearest in stable, competitive repeat-purchase categories and weaker in new, fast-changing or highly fragmented ones.
- Niche and lock-in exceptions. Brands that serve a distinct need, or that rely on contracts and switching costs, can show more loyalty than their size predicts.
- Not a verdict on loyalty programs. The law sets expectations for how far loyalty can move; a program can still be worthwhile if a holdout test shows incremental profit.
- Penetration is not free. Distribution, sampling and reach cost money, and extra buyers still have to be profitable buyers.
- Ethics: chasing penetration with misleading trial offers or hard-to-cancel subscriptions buys short-term customers and long-term distrust.
What if you apply it incorrectly?
A team reads Double Jeopardy as "loyalty doesn't matter" and cuts product and service quality to fund a reach campaign. Trial rises, but disappointed first-time buyers don't come back, and the brand ends the year with more lapsed buyers rather than more regular ones. The law describes how loyalty usually scales with size; it is not permission to neglect the experience that turns trial into repeat purchase.
Myth versus evidence
The social-video version
Small brands have the most loyal fans, so a loyalty program is the fastest way for them to grow.
What the evidence supports
Smaller brands typically have fewer buyers who are also slightly less loyal, so growth usually depends more on reaching new buyers.
Five-minute field assignment
From your sales, panel or CRM data, write down two numbers for your brand and two rivals (or, without rival data, for your own brand in each of the last three years): the share of the market that bought at least once, and the average purchases per buyer. Which of the two numbers differs more between brands, or has moved more over time?
Checkpoint
Two questions: one about the idea, one about a decision. Answer both correctly to master this law. Questions are drawn from a larger bank and graded on our server; a retry brings different questions.
Evidence ledger
Every factual claim in this lesson, where it comes from, and how far it has been checked.
-
Smaller brands have fewer buyers who also buy them somewhat less often; the loyalty gap is much smaller than the penetration gap.
Limit: Mainly established repeat-purchase categories in fairly stable markets.
-
The term Double Jeopardy is traced to the sociologist William McPhee in the early 1960s.
Limit: Attribution as reported in the marketing literature; McPhee's original work is not cited directly here.
-
The Double Jeopardy pattern has also been reported among B2B brands.
Limit: Institute summary article; check which B2B categories and data it covers.
-
The sparkling-water company and the five-brand table.
Limit: Teaching scenario; no real company or category data.
Further reading
- Ehrenberg, A. S. C., Goodhardt, G. J. & Barwise, T. P. (1990). Double jeopardy revisited. Journal of Marketing, 54(3).
- Ehrenberg-Bass Institute for Marketing Science. The double jeopardy law in B2B shows the way to grow.
Which other law could explain this decision?
Editorial review: Ali Dezhami, November 8, 2026. Last substantive update: November 8, 2026. Our methodology