by Byron Sharp
In short, the aim of growing category purchases through presenting new uses often turns out to be a pipe-dream. That's why marketers who want to increase sales need to win more market share and/or enter new markets.
Penetration the winning target
One sound principle that is taught at business school is that you are far more likely to succeed if you have the right goal. A meta-analysis of 207 US split-cable advertising weight tests concluded that only one of the measured strategy variables was correlated with larger sales effects – i.e. having the objective of increasing penetration (Lodish et al. 1995 p.130).
While an analysis (Binet & Field 2007) of 880 entries to the IPA's Effectiveness Awards showed that winners were far more likely to have set targets to increase market penetration. Each entrant stated their primary business purpose: for 178 entrants, this was 'customer retention/increase loyalty', while for 79 entrants it was 'new customer acquisition/penetration'. Here's how the awards played out:
Table 2.6: Analysis of the IPA's Effectiveness Awards
Target to increase: Penetration Loyalty
Gold winners 21% 2%
Silver winners 20% 6%
Bronze winners 18% 3%
No medal 41% 89%
100% 100%
Source: Binet & Field 2007.
Campaigns that, unfashionably, aimed to increase penetration were more than twice as likely to report very large improvements in all hard measures of effectiveness including sales and profits. This pattern was strongest for durables and services. Yet only half the number of submissions aimed to increase penetration as aimed for loyalty/retention improvements or to change attitudes. Therefore, marketers are using the wrong targets.
UPDATE: Since the first version of “How Brands Grow” was published, Binet & Field have updated their analysis in a 2013 IPA report titled “The Long & the Short of It: balancing short and long term marketing strategies”. This time with a larger database covering 30 years of marketing campaigns, and 996 campaigns, they again report “across the board, campaigns targeting new customers outperform campaigns targeting existing customers. In terms of the average number of business effects reported, the former are three times as effective as those targeting existing customers” (p.24).
Les Binet confided to me that he once wrote an effectiveness awards submission where he and his client aimed to increase penetration in the north and loyalty in the south. Annoyingly, while loyalty did improve in the south, penetration there improved even more. Try as one might, if you are successful in gaining sales it's unlikely you'll break the double jeopardy law.
And the IPA Effectiveness Awards results remind us that it is a much better strategy to go with the law than against it. Jim Nyce, previously Insights Director at Kraft, describes this as 'swimming downstream'. An analysis (led by Frank Cotignola) by his Consumer Insight & Strategy department showed that 56% of their brand plans were trying to 'swim upstream' by raising purchase frequency, while an internal study of the growth and decline patterns of 67 Kraft brands showed that penetration was the dominant driver of sales and share, in line with the double jeopardy law.
What about niche brands?
The term 'niche' is used loosely in marketing; often it is used simply to mean small.15 This turns out to be appropriate because most niche brands are small. Technically, a niche brand in a category should, for its market share, have an unusually small base of buyers who are unusually loyal.16 If every category had a number of these brands, and/or their differences truly were substantial then the double jeopardy pattern wouldn't exist. There are far fewer niche brands than people expect, and they are less niche than we think.
Nor is being niche necessarily a good thing. Rather than thinking of niche brands as having excess loyalty more are best thought of as having less penetration (and market share) than we’d expect given their loyalty levels. This penetration deficit is usually due to some lack of availability, such as being missing from a particular distribution channel (e.g. retail private labels are completely missing from rival store chains), or missing in a particular geographical area. This penetration deficit may also be due to lack of mental availability amongst part of the buying population, for instance, because of their lack of advertising some private labels are simply not thought of or noticed by some customers – more people than expected given their market share. Similarly differentiated brands may be rejected by part of the buying population while the rest value the differentiation – Spanish language TV channels in the US are an example. They are not watched by those who don’t speak Spanish but watched for many hours by those that do. Whatever the reason a brand’s lack of penetration is rarely something to boast about.
What about cross-selling?
Another route to sales growth is to encourage current customers to buy different products. Customer relationship management (CRM) systems and loyalty programs often promise this benefit. It's widely viewed that this is an easy path to growth: that is, you already have a relationship with these buyers, so if they are made a good offer they should take it up. Yet many large corporations make good offers to their own employees and are disappointed by the response rate. If selling new products to your own employees is difficult, then cross-selling to existing customers might not be so easy after all.
Cross-selling metrics are another measurement of loyalty, so once again we find the double jeopardy law applies. There is very little difference in cross-selling metrics between competing brands, and the small differences that do exist tend to reflect market share – not whether or not they have dedicated cross-selling programs.
Tables 2.7 and 2.8 provide data on cross-selling metrics for insurance and banking, respectively.
Table 2.7: Cross-selling metrics for insurance
Insurance providers
(Australia)
Market penetration
(%)
Average number of products held by customers
RAA 16 1.5
CGU 14 1.4
SGIC 13 1.5
AAMI 9 1.5
APIA 6 1.4
Average 12 1.5
Source: Mundt, Kerry, John Dawes, and Byron Sharp (2006) "Can a brand outperform competitors on cross-category loyalty? An examination of cross-selling metrics in two financial services markets", Journal of Consumer Marketing, Vol. 23, p.465-569
Table 2.8: Cross-selling metrics for banking
Banks
(Australia)
Average number of banking products held by customers with that bank
Commonwealth Bank 2
ANZ Bank 2.1
Westpac Bank 2.2
National Australia Bank 2.3
Average 2.1
Source: Mundt, Kerry, John Dawes, and Byron Sharp (2006) "Can a brand outperform competitors on cross-category loyalty? An examination of cross-selling metrics in two financial services markets", Journal of Consumer Marketing, Vol. 23, p.465-569
The insurance data shows very little difference in the success of cross-selling by the competing brands. In banking there is a little more variation, but oddly there appears to be a slight reverse of the double jeopardy pattern where the two smaller brands have marginally higher customer loyalty. This is due to the two smaller brands being skewed towards business banking and having a consumer base of wealthier customers who hold more financial products overall, while the two larger brands are particularly successful in one product (e.g. credit cards), and this success gives them slightly more customers than expected – customers who hold just one product with them. So their deficient cross-selling metric is actually due to their unusual success in one part of the market. In other words, we see no evidence of excellence in cross-selling, though we do see some brands excelling at selling a particular product to additional customers.
In both categories there is very little difference between brands in cross-sell metrics. Every insurance brand has a customer base that, on average, buys one and a half services from them; each banking brand's customers buy two services.<
br />
There are dubious case studies of firms that have achieved cross-selling success. One of the problems is that the metric ‘numbers of products’ is open to manipulation. For example, you must have a separate EFTPOS card even though the credit card we gave you can already access your savings account, oh and we've added a small overdraft to your savings account – voila the count of ‘products per customer’ goes up. When managers are set targets to increase ‘number of products’ per customer there is a strong incentive to be inventive and manipulate the loyalty metric. Wells Fargo is the market leader in retail banking in the US, it also claims to be the leader in cross-selling. It says it has reached six products per household (out of a potential of 16). Of course, double jeopardy says Wells Fargo has to have the highest loyalty metrics because of its market share. We also know that the difference between brands is unlikely to be great, and this appears to be the case. The Wall Street Journal (Smith 2011) reports that financial analysts are doubtful about Wells Fargo’s claims: "It's always been a bone of contention, one of those things that Wall Street tended to view with a jaundiced eye," says bank analyst Nancy Bush of NAB Research LLC in Annandale, New Jersey. "If I have a checking account and get a debit card, that's two products which another bank might only count as one”, she added. Unsurprisingly, other banks have disputed Well Fargo’s claim of leadership saying that their own cross-sell metrics are just as good.
None of this says that cross-selling is impossible, but it does imply that it’s nowhere near as easy as textbooks make out, and certainly not a route to huge sales gains. The lack of difference among the brands shown in Tables 2.6 and 2.7 suggests that dramatically changing cross-selling metrics is difficult and expensive (some brands have tried, but you wouldn't know it from looking at their results).
One reason it is harder than expected to cross-sell products to existing buyers is that those who haven't bought these products from you don't need them; for example, you can't sell car insurance to someone who drives a company car. Another neglected reason is that service brands often already enjoy high loyalty; for example, most of your customers who needed a home loan came to you already (and you either gave them one or turned them down). So improving this already high loyalty is difficult.
And profits?
It’s often naively hoped that investing in current customers might yield greater net returns than seeking to expand the customer base. The simplest idea is to spend the most on heaviest customers and the least on the lightest. It assumes that heavy customers (i.e. those that buy more of the brand) are the best candidates for investment. I call this the ‘Heavy Buyer Fallacy’ because the misguided logic confuses past/current buying with growth potential – when it’s more reasonable to think that heavy customers are already buying all that they ever will. It seems logical that a firm can afford to spend more on a heavy customer than a light one; certainly direct marketing managers find it easier to justify sending expensive letters to heavier customers, and very hard to justify spending several dollars on customers who deliver less than this in sales revenue. However the key issue is what this marketing expenditure will return (What change will it evoke?), so even customers who are currently worth nothing may be worth marketing to in order to win their future custom.
Another fallacy is that customers who stay longer will become more profitable as they somehow become less price sensitive. Rienartz & Kumar (2000, 2002) investigated profitability and tenure using data from a US catalogue retailer, a French retail food business, and a German direct brokerage, and found that a customer’s tenure did a surprisingly poor job of predicting their lifetime profitability, because even short-term customers could generate a good deal of profit. Their cohort analysis partly explains this by showing that customers did not become more profitable over time, nor did long tenure clients pay higher prices. In industrial settings this may due to long-standing customers extracting better terms and discounts. In consumer settings long-standing customers may learn to pick the very best value items from a firm’s portfolio, or learn when price discounts are offered.
Further reading on the double jeopardy law
Extraordinary claims require extraordinary evidence, hence the many data sets in this chapter. Those interested in delving deeper, or simply seeing the wide range of conditions under which the double jeopardy law has been documented (e.g. attitudes, behaviour, industrial, services, durables, retail, voting, media) may wish to read the following peer-reviewed journal articles:
Bennett, Dag & Graham, Charles 2010. 'Is loyalty driving growth for the brand in front? A two-purchase analysis of car category dynamics in Thailand.' Journal of Strategic Marketing, 18:7, 573-85.
Donthu, N. 1994. 'Double jeopardy in television program choice.' Journal of the Academy of Marketing Science, 22:2, 180-85.
Ehrenberg, Andrew S. C. 1972. Repeat Buying: Theory and Applications. New York: American Elsevier.
Ehrenberg, Andrew, Goodhardt, Gerald, Barwise, Patrick 1990, 'Double jeopardy revisited', Journal of Marketing, vol. 54 (July), pp. 82-91.
Ehrenberg, Andrew & Goodhardt, Gerald 2002. 'Double Jeopardy revisited, again.' Marketing Insights, Marketing Research, Spring 2002, 40-42.
Ehrenberg, Andrew 1991, 'Politicians' double jeopardy: a pattern and exceptions', Journal of the Market Research Society, vol. 33, no. 1, pp. 347-53.
Bhat, S, Fox, R 1996, 'An investigation of jeopardy effects in store choice', Journal of Retailing and Consumer Services, vol 3, no. 3, pp. 129-33.
Martin, C., Jr 1973. 'The theory of double jeopardy.' Journal of the Academy of Marketing Science, 1:2, 148-56.
Michael, JH, Smith, PM 1999, 'The theory of double jeopardy: an example from a forest products industry', Forest Products Journal, vol. 49, no. 3, pp. 21-6.
McDowell, W. S. & Dick, S. J. 2001. 'Using TV Daypart 'Double Jeopardy Effects' to Boost Advertising Efficiency.' Journal of Advertising Research, 41:6, 43-51.
McDowell, WS, Dick, SJ 2005, 'Revealing a double jeopardy effect in radio station audience behavior', Journal of Media Economics, vol. 18, no. 4,pp. 271-84.
Sharp, Byron, Riebe, Erica 2005, 'Does triple jeopardy exist for retail chains?', Journal of Empirical Generalisations in Marketing Science, vol. 9.
Solgaard, H, Smith, D Schmidt, M 1998, 'Double jeopardy patterns for political parties', International Journal of Public Opinion Research, vol. 10, no. 2.
Uncles, Mark & Lee, D. 2006. 'Brand purchasing by older consumers: an investigation using the Juster scale and the Dirichlet model.' Marketing Letters, 17:1, 17-29.
Wright, Malcolm, Sharp, Anne, Sharp, Byron 1998, 'Are Australasian brands different?', Journal of Brand and Product Management, vol. 7, no. 6, pp. 465-80.
A full reference list is at the end of this book.
Chapter 3:
How To Grow Your Customer Base
Byron Sharp
What happens when brands grow or decline?
The double jeopardy law tells us that when brands improve their market share their buyer base enlarges. This growth in customer numbers could be due to improvements in customer acquisition, but it could also be the result of reduced customer defection. It's a fact of marketing life that each year you lose buyers. If a brand can improve its retention levels then it should grow its customer base.
So in theory it's possible to grow your customer base by improving either retention or acquisition, or a combination of both. We'd expect that making customers more satisfied might bring about both, especially retention. There is now a large body of literature based on this assumption – in fact one could say that the marketing literature for more than a decade has championed retention management.
This raises the strategic question of whether marketers should emphasise retention or acquisition. Modern marketing ideology says retention is cheaper than acquisition. But is it? And what returns are possible? How much emphasis should be placed on retention versus acquisition?
Is retention cheaper?
A widely read Harvard Business Re
view article by Reichheld and Sasser (1990) states that customer defections can have a “surprisingly powerful impact on the bottom line ... companies can boost profits by almost 100% by retaining just 5% more of their customers”. This is a surprising, fantastic claim, and it turns out to be just that – pure fantasy.
It's reasonably presumed that this claim is based on empirical research, but Reichheld and Sasser's (1990) statement is merely based on a thought experiment; and it goes like this:
Suppose a credit card company loses 10% of its customers each year, then the average customer life would be 10 years. Now if that firm were able to reduce its annual customer defection to 5% then the average customer tenure would double to 20 years. Given that a customer delivers some profits each year, now they stay for more years they must each give more.
This is an analytical tautological 'finding', not a real-world discovery based on observing the results of retention efforts. It's similar to saying that if you win the lottery then you will be richer: true by definition and hardly surprising.
Reichheld and Sasser (1990) presented their logic in a misleading manner because:
•Their fictional 5% drop in defection is actually a drop of 5 percentage points, i.e. from 10% to 5%, which is a 50% decrease; a halving of customer defection (doubling of loyalty).17
•Their thought experiment wasn't about company profitability, it was about 'customer profitability', which is different (it’s a rather theoretical accounting concept). Essentially all they revealed is that if a customer stays (i.e. buys) for longer, then they give you more money over this longer period.