A little bit of history

The Double Jeopardy Law is one of the heroes of marketing science. It was first discovered in the 1960’s (see McPhee, 1963) and since then has popped up in a wide range of categories, contexts and countries. Examples include packaged goods, retail banking, insurance, luxury products, political voting, automobile buying. At the Ehrenberg-Bass Institute, we find even newly disrupted categories such as music listening, ride-share or vacation accommodation all display Double Jeopardy patterns, and in all countries where data is available.

The Double Jeopardy law matters because it reveals the path to brand growth. There are those who claim there are multiple paths of grow to choose between: that a brand can grow via acquisition, or cross/up-selling/retention, or some preselected mix of the two. The Double Jeopardy law shows us this is an imagined world, a world dreamt up by fanciful consultants. We don’t observe sustainable1 growth coming from a range of different paths. Nor do we see brands of similar market share, but very different penetration and loyalty metrics. The real world shows one clear path to growth: Through getting more customers to buy from you in any time period.


The Double Jeopardy law matters because it reveals the path to brand growth.


Why might B2B marketing be different?

We’ve been asked many times over the years, “Surely B2B is different to consumer marketing, will your empirical laws like Double Jeopardy hold up there?”. We asked marketers why they thought B2B might be completely different to B2C. The answers included:

  • (Sometimes) there is a limited number of potential customers
  • Big value of purchases means buyers are more involved
  • There is a structured purchase cycle, so less reliance on memory
  • People’s jobs and KPIs are about procurement so they are financially motivated to get the ‘right’ brand/company
  • Often involve long term contracts so chance to develop deeper relationships
  • You get advance notice on contract renewal times so can effectively plan retention efforts There are more switching barriers, and switching costs
  • Longer purchase cycles
  • And so on…

But we also need to consider the similarities between the two. What do B2B markets have in common with B2C markets?

  • Often there is a wide range of options for customers to buy
  • Decisions are made by humans, who have an inability to process all the information available
  • While buyers form memories about B2B brands, not all memories are equally accessible, some are fresher than others
  • Any single buying decision is only a small part of the total buying that any company (or person within that company) has to do
  • Sometimes something new catches a buyers’ eye and they decide to try it, which can mean switching even though the original supplier/brand has done nothing wrong
  • The more important a purchase is, the more a customer might want to spread the risk by using multiple suppliers
  • And so on…

It is easy (and common) to focus on the differences rather than the similarities. But the B2B customer is also buying toothpaste, cars, vacations, chocolate, whisky and home insurance, with the same brain. So it need not be too surprising that there are similarities in how brand choices are made, even when the buying context differs dramatically. Indeed we could have written similar/different lists comparing packaged goods with services; services with durables; luxury with non-luxury and so on – all of which where the Double Jeopardy Law holds. So we should not assume that just because the context is different, the path to growth will also differ, otherwise we are failing to learn from science and history.


Types of B2B markets

B2B markets or categories are many and varied, and it’s useful to be able to classify them. A broad way to classify categories comes from work by Professors Sharp, Wright and Goodhardt (2002) which talks about repertoire and subscription categories. The type of category affects the relevant loyalty metrics:


Repertoire categories 

Where people can satisfy their category requirements with more than one brand (called the ‘repertoire’) on an ongoing basis. So if you see someone buying widgets from Company X and then on the next purchase buying the same or similar widgets from Company Y, this is not necessarily an indicator they have stopped using Company X, just that they are cycling through the multiple brands on their repertoire (on B2B markets this often gets the fancier name of ‘preferred suppliers’).

Each purchase is a transaction that requires a distinct choice, timing of purchases is often sporadic, and split loyalty across multiple brands for category buying is normal. Loyalty metrics in these categories are about purchase frequency, share of category requirements and percentage of solely loyal buyers.


Subscription categories

This is where people typically use one brand at a time for the category due to its nature. So if someone took out General Liability insurance with Allianz, they would only take out a General Liability policy with Zurich Insurance if they cancelled the Allianz policy.

The timing of purchases is cyclical and using one company to satisfy category requirements is normal. Therefore loyalty metrics are about retention/defection and tenure.

This means there are a wide range of loyalty metrics for B2B categories. Fortunately, brand penetration, or the number of customers a company has, is a single metric that is constant across both types of categories and systematically related to the other metrics. The only complexity for calculating brand penetration is clearly specifying the customer universe (‘Who is a category buyer?’). As long as this specification is not too narrow in either type of customer or time frame, and is consistent for all companies and over time, penetration can be measured and used in analysis.

In the next section we explain the background of Double Jeopardy, then show several examples of it in B2B categories. We explain how you can measure it, and summarise how you can use the resultant knowledge to grow.


Knowing if your B2B category is repertoire or subscription helps define the metrics that matter.


What does the law of Double Jeopardy look like in B2B?

The Double Jeopardy Law shows that brands with fewer customers have lower loyalty/ higher defection levels than brands with more customers. Here we show some recent examples of Double Jeopardy in B2B markets – UK Business Banking, and US Business Insurance.

For Banking we show three loyalty metrics: # of banking products with the bank (out of 17 possible products); % of customers solely loyal (just use one bank for all banking products) and attitude to the brand (% of customers saying it is their favourite).

For Insurance, we focus on a defection metric, which is the probability of switching away from the current brand at next renewal (average score on an 11-point Juster Scale, where 0 is little or no probability of switching while 10 is certain/practically certain to switch Insurance suppliers). In this instance, a higher score equals greater chance of defection, therefore lower loyalty.

Of course we see some variation in some instances (eg, Nationwide lower on Solely loyal or TSB higher on % of customers saying it is their favourite bank). Before placing too much importance on these, we recommend checking it is not simply sampling variation, and these are persistent deviations worth investigating further.

The first finding is that the Double Jeopardy Law is evident – brands with fewer customers have lower loyalty/higher defection levels.


The Double Jeopardy Law shows that brands with fewer customers have lower loyalty/ higher defection levels than brands with more customers.


Table 1. UK Business Banking, data collected by the Ehrenberg-Bass Institute



In Table 2, we see the relationship between loyalty and penetration in action for specific Insurance products. In Accident and Health Insurance, Progressive Commercial is a smaller brand and has a higher defection rate (3.1 compared to 2.1 for the biggest brand in the product category, Allstate).

However when we look at Commercial Property Insurance, Progressive Commercial now has the lower defection rate of a big brand (2.1) while Allstate has the higher defection rate of a smaller brand (3.0). To re-iterate: Loyalty is largely a function of the size of your customer base.


Table 2. US Business insurance brand defection rates, data collected by the Ehrenberg-Bass Institute



But the Double Jeopardy pattern is not just for B2B services, In Table 3 are two ‘durable’ B2B examples: Concrete Suppliers and Airplane buying taken from other published sources (Pickford & Goodhardt 2000 and Bennett, Anesbury & Graham 2018). Again the Double Jeopardy pattern is evident.


Table 3. Double Jeopardy in Concrete and Airlines markets



Some other evidence, in case you still need convincing…

Table 4 lists a range of studies from a diverse set of categories, countries, data collection methods and researchers. All show evidence of Double Jeopardy.


Table 4. Overview of published studies with evidence of Double Jeopardy in the B2B sector



“OK, you’ve almost convinced me, but my market is a bit different so I need that final bit of evidence to confirm this really does apply to my company/brand….?”


One of the challenges of the B2B sector is that good quality, customer centric data across the category is difficult to acquire. Indeed, lack of data has held back research in the sector for a while. How can you overcome this challenge?

A well-constructed survey can obtain all the information you need to examine Double Jeopardy in your market. To do this you need a way to reach a good cross-section of business buyers in your market, not just your customer base. There is no one single sample size recommendation, but just remember that loyalty metrics are taken from each company/brands customer base only, so you need a large enough sample to make sure you have robust loyalty metrics for all competitive brands you are interested in.

The key questions you need to ask are quite easy to construct. Let’s use a simplified example of business insurance: and say the main products are professional liability, property, worker’s compensation, product liability, vehicle, and business interruption insurance.


1. Find out if the business buys the product / service

So, your survey would ask along the lines of “which types of business insurance does your business have: does it have – Professional liability insurance …. Property insurance …” and so on. To identify the category buyer universe.


2. Brand Penetration

Next, your survey would ask which providers the business uses, for each of the insurance products that it buys. As this is not designed to be a test of memory, we recommend providing a list of the major providers and then space for someone to add any that are not on the list. From this information you can calculate each brand’s penetration: the proportion of category buyers who use each brand.


3. Loyalty metrics

There are several different options you can use for loyalty.
For multi-product categories, examples include:

  • Number of products with each brand – sum of how many times the brand is bought across all products by its customers
  • % solely loyals – the % of a brands customer base that only use one brand across all products
  • % main/lead supplier – % of a brand’s customer base that consider it their main/lead supplier

For single product categories (such as professional liability insurance only), examples include:

  • Past Defection rate – for annual renewal products ask about current supplier and supplier this time last year, and calculate the % who switched on last renewal occasion
  • Future Probability of switching/renewing contract – the Juster probability scale provides an future mean average renewal/defection rate2
  • Purchase frequency – the number of times the brand is bought in repeat purchase categories, where companies buy on different occasions and each purchase is a separate transaction.

Now you have the key information you need for your brand and competitors. Armed with this category wide information, all you need to do is order the brands by penetration, as in Tables 1-3, to see the Double Jeopardy law in action.


“What if I see Exceptions/Deviations to the Law?”

So you collect the data and sort the brands in market share order, and perhaps you notice something odd: a brand that has higher or lower loyalty than its similarly-sized neighbours. What does that mean?

Well, first remember, it is the Double Jeopardy Law that allows you to see this exception. An exception doesn’t invalidate the law: some brands are a bit different but there is usually a simple reason.

Let’s take the example in Table 5 for Commercial Auto Insurance. If you look at the defection metrics you see that Progressive Commercial has a much lower defection rate than other brands that have similar penetration. Therefore it would be useful to know why this has occurred3.

The Double Jeopardy Law won’t explain exceptions, but it will identify them, and help you prioritise what to investigate further. Otherwise you might spend a disproportionate amount of time trying to understand why Geico or AIG have higher defection rates for this product, when it is simply due to them being smaller brands in the category.


Table 5. Exceptions to the rule: One brand with unusual loyalty




This Ehrenberg-Bass Institute report shows Double Jeopardy Law holds for different B2B categories, and for the majority of brands within those categories. Therefore, despite the differences that exist between B2B to B2C sectors, a fundamental aspect of company performance is the same across them: the Double Jeopardy law holds in both.

This is a step forward for B2B marketing as it points to a clear path to growth, which in turn can help you focus your strategic and tactical options to ones more likely to succeed.

There are some deviations, but these are the exceptions that ‘prove’ rather than refute the law. The presence of Double Jeopardy allows us to identify the exceptions and investigate them further. Given the rarity of these deviations, your B2B company/brand is more likely to be normal than exceptional. And that is a good thing for growth, as it suggests no barriers to growth.

At the Ehrenberg-Bass Institute, we have also spent a considerable amount of time studying exceptions to the Double Jeopardy Law. A company with what looks like ‘excess loyalty’ is most likely suffering from a deficit in penetration, which means it is neglecting or cutting some category buyers out of its potential acquisition base. This might be due to the firm’s distribution, or its portfolio of offerings. Such a company has a ceiling on its growth potential.


…a clear path to growth, which in turn can help you focus your strategic and tactical options to ones more likely to succeed.


Four steps to use this knowledge to grow your business

Recognise that in order to grow, your business needs to expand the size of its customer base – this is not optional.

Double Jeopardy tells us that growth comes primarily from new customer acquisition. Just marketing to your own customer base will not achieve this objective. Review your marketing plans – how much effort is directed at (only) your customer base, versus reaching the wider base of category buyers?
Is that allocation indicative of a company that is going to grow by getting more customers?

It could be easy to interpret this as saying the loyalty from existing customers isn’t important.

We’re not saying that. Of course the fact that you have existing customers who buy from you repetitively over time is important – you need to look after your existing clients – that is necessary, but not sufficient to grow. Instead of being preoccupied with loyalty, look to removing barriers to penetration that will ease the path for acquisition, as this is more likely provide opportunities to grow. Barriers to penetration can be mental (what you are salient for offering) or physical (where, how or what potential customers can buy from you).


Set evidence based KPIs

The results tell us that there are natural ceilings to any loyalty metric. Senior managers cannot just say, we want to increase loyalty by 50% or, we could grow a lot by simply selling all our buyers one more product.

An easily missed (and misinterpreted) point about Double Jeopardy is that while big brands get some more loyalty, it’s not really about the brand, but the brand’s market share. If certain big brands got substantially more loyalty than their littler competition, this could imply there are some that are just much better – higher levels of product quality, greater expertise, for example. But they don’t. Product quality is not unimportant, it is just not all-important such that you can neglect other areas of marketing and expect to win simply with the best product.

This take-out is surprising for many senior managers who assume that the reason some companies are so successful is to do with being better than competitors (better product, better service, better technology etc). But the ubiquity of Double Jeopardy says this isn’t the reason. You have more loyalty because you are big, rather than you are big because you have more loyalty.

Double Jeopardy tells you what your loyalty should be (which is usually what it actually is). Not just now, but in the future too, should your company grow or decline in share. This allows you to set realistic, evidence based, KPIs and forecasts for growth goals.


Put yourself (and your marketing) in the mind of your non-customers

You need to reach and build the brand among businesses who are category buyers but who do not buy your brand currently, for them to start buying from you. The ways you can do this range from wide reaching advertising and/or setting KPIs for the sales force to talking to prospects. Think about all of the mechanisms you have to connect with non-customers, and then look at how they can be changed to reach more/different non-customers.

For example, advertising in an industry magazine might reach a different set of non-customers than a mail out from your existing marketing database. Also switching out one in four calls to existing customers to be to non-customers will expand your reach to non-customers. Now, you might be concerned that this reduced contact will cost you customers, so to reduce this risk you might space out the contact you have with existing customers so there are no large gaps, or replace the phone call with an email.

There is no single formula for this, it’s simply about working out how to use the resources you have in the next quarter to reach more non-customers than you did last quarter.

It’s crucial you don’t waste that reach because of poor branding quality. Its vital to have clear and prominent branding, so your company is easily identifiable to the most disinterested potential customer. Building strong Distinctive Brand Assets is important here (see Romaniuk, 2018). These are the visual and/or audio elements that automatically trigger the brand, when the brand is not present These help your company’s branding to cut through in any environment.


Prioritise Mental and Physical Availability

Overall, the business implication is that to grow, one needs to invest more in making the brand easier to think of in buying situations (Mental Availability), and easier to buy (Physical Availability). There will be more on these two levers in upcoming reports, but essentially they are about:


Mental availability  
Being thought of, by more customers, in more buying situations. This is a function of the media, branding and message quality of your customer interactions via media or in person via your sales force.


Physical availability

Being easy to find and buy from. This is a function of the brand’s presence and prominence in buying channels, (including sales force coverage) as well as the product portfolio on offer.

Each amplify the effect of the other, and are essential for brand growth in all categories, including B2B.



1 We are separating out longer term growth from spikes due to short term activation activities such as price promotions. These spikes are rarely cost effective. For more on this see Dawes, John and John Scriven (2010).
What price promotions really do. How Brands Grow. Byron Sharp. Australia, Oxford University Press: 153-170.
2 Danenberg, Nick J. (1998). Predicting Customer Loyalty: A Probabilistic Approach. School of Marketing. Adelaide, University of South Australia
3 The most likely cause for instances of excess loyalty like this is that the firm has restricted distribution, such as operating in a specific geography or sub-market.




Bennett, Dag, Anesbury, Zachary & Graham, Charles 2018. Buying in a heavy industrial context. Report 88 for Corporate Sponsors. Adelaide: Ehrenberg-Bass Institute for Marketing Science.

Bowman, Douglas & Lele-Pingle, Shilpa 1997. Buyer Behavior in Business-to-Business Services: The Case of Foreign Exchange. International Journal of Research in Marketing, 14, 499-508.

Ehrenberg, Andrew 1975. The Structure of an Industrial Market: Aviation Fuel Contracts. Industrial Marketing Management, 4, 273-285.

Ehrenberg, Andrew, Goodhardt, Gerald & Barwise, Patrick 1990. Double Jeopardy revisited. Journal of Marketing, 54, 82-91.

McCabe, James, Stern, Philip & Dacko, Scott G. 2013. Purposeful empiricism: How stochastic modeling informs industrial marketing research. Industrial Marketing Management, 42, 421-432.

McPhee, William N. 1963. Formal Theories of Mass Behavior, New York, United States, Free Press. Michael, Judd H. & Smith, Paul M. 1999. The Theory of Double Jeopardy: An Example from a Forest Products Industry. Forest Products Journal, 49, 21-26.

Pickford, Chris & Goodhardt, Gerald 2000. An empirical study of buying behaviour in an industrial market. AM2000, Academy of Marketing Annual Conference, University of Derby, Derby, University of Derby.

Romaniuk, Jenni 2018. Building Distinctive Brand Assets. Victoria, Oxford University Press. Romaniuk, Jenni & Sharp, Byron 2016. How Brands Grow: Part 2. Victoria, Oxford University Press.

Romaniuk, Jenni 2021. Getting down to business in How brands grow: Part 2 (2nd edition). Victoria, Oxford University Press. Forthcoming.

Sharp, Byron 2010. How Brands Grow. Victoria, Oxford University Press.

Sharp, Byron, Wright, Malcolm & Goodhardt, Gerald 2002. Purchase loyalty is polarised into either repertoire or subscription patterns. Australasian Marketing Journal, 10, 7-20.

Stern, Philip 1994. Prescriptions for branded and generic pharmaceuticals. Journal of Brand Management, 2, 177-183.

Uncles, Mark D. & Ehrenberg, Andrew 1990. Industrial Buying Behavior: Aviation Fuel Contracts. International Journal of Research in Marketing, 7, 56-68.

Wright, Malcolm & Riebe, Erica 2010. Double jeopardy in brand defection. European Journal of Marketing, 44, 860-873.



Now available as an eBook exclusively to Apple iBooks


The Ehrenberg-Bass Institute for Marketing Science is the world’s largest centre for research into marketing. Our team of market research experts can help you grow your brand and develop a culture of evidence-based marketing.