Share of wallet is not the key to growth - a loyalty myth
05/10/11 15:33
I was surprised and disappointed to see that the authors of the book “Loyalty Myths”, TIm Keiningham et al, themselves promote an old loyalty myth in the latest Harvard Business Review (Oct 2011).
The article reports a "discovery" of a relationship between verbal attitude reports and Share of Wallet/Purchases (measured in a period of panel data). This is re-discovered every so often, indeed Jan Hofmeyr even won an award for it a few years ago. This re-discovering (and misinterpretation) over and over says a lot about the sad state of our discipline given that the relationship was very well documented in the 1960s by Bird, Channon, and Ehrenberg (see references below).
Like Hofmeyr & colleagues (2008) the HBR team have done a rank order correlation which, in this case, guarantees high correlations of around 0.9. Correlation is not a good way of assessing predictive ability (which I illustrate here). Yet they report that their analysis is “to our knowledge the largest and most rigorous of its kind”.
The article goes on to imply that great financial returns can be gained from improving loyalty of existing customers. They recommend that managers move from a focus on the drivers of satisfaction to focus on share of wallet rank. This takes a leaf out of Reichheld’s book, when he said forget about satisfaction and focus on willingness to recommend (Net Promoter Score). In this case it’s true that larger market share brands do have higher share of wallet rank. What they fail to mention is that such loyalty metrics move only a little, and do so when a brand's penetration increases dramatically – the Double Jeopardy law.
In effect they rely on an old “wouldn’t it be great” argument that goes along the lines of “if those customers who give us very little of their spending became near 100% loyal, then we'd earn X more money”. It’s like saying “if you win the lottery then you’ll be richer”, all logically/analytically/theoretically true but it’s easy to make theoretical statements that clash with the empirical world, as this does. It's all beautifully academic, but not scientific nor practical. It’s not how brands grow.
Note how costs are also conveniently ignored. In the same way that Reichheld ignored them when promoting his retention and profitability myth.
The authors recommend finding out the motivational reasons why people give only a small amount of their spending to your brand and lots to others. But this ignores the weak link between claimed motivations and behaviour.
In the article they present a grocery store as an (fictional, hypothetical) example - why they didn’t (couldn’t?) show a real case is rather telling. In their hypothetical example the store drops prices on staple items to encourage low share of wallet customers to give them more of their supermarket shopping trips. Now supermarket repertoires have what is called high "1st store loyalty", in that the top store in anyone's repertoire tends to dominate, with the remainder of their repertoire being divided amongst several others. From the store's perspective then, it has a large group of customers who are very loyal in the sense that they do most of their grocery shopping there. And then another group of customers who largely do their shopping at another store. The former group tend to live close by, the latter further away (a fact that is ignored in the article).
So TIm Keiningham & co are on sound ground in recommending a focus on the people with low SCR - sales growth does indeed largely come from the latter group (since obviously it's hard to get more out of people who already give you nearly all their custom). But the real story for growth is, target your light and non-customers. Go for reach. Increase your mental and physical availability (e.g. open longer hours, open another store, make it easier for people to shop and shop faster...).
That's a very different story than the old-fashioned loyalty one (re)presented in the HBR article. And the promise of riches to be gained from a share of wallet focus…..that’s a chimera.
Professor Byron Sharp
October 2011
BIRD, M. & EHRENBERG, A. 1966. Non-Awareness and Non-Usage. Journal of Advertising Research, 6, 4-8.
BIRD, M. & EHRENBERG, A. 1966. Intentions-to-Buy and Claimed Brand Usage. Operational Research Quarterly, 17, 27-46.
BIRD, M. & EHRENBERG, A. 1970. Consumer Attitudes and Brand Usage. Journal of the Market Research Society, 12, 233-247.
BIRD, M., CHANNON, C. & EHRENBERG, A. S. C. 1970. Brand image and brand usage. Journal of Marketing Research, 7, 307-314.
BIRD, M. & CHANNON, C. 1969. Brand Usage, Brand Image, and Advertising Policy - Part I. Admap, 6, 27-46.
BIRD, M. & EHRENBERG, A. 1972. Consumer Attitudes and Brand Usage - Some Confirmations. Journal of the Market Research Society, 14, 57.
HOFMEYR, J., GOODALL, V., BONGERS, M. & HOLTZMAN, P. 2008. A new measure of brand attitudinal equity based on the Zipf distribution. International Journal of Market Research, 50, 181-202.
The article reports a "discovery" of a relationship between verbal attitude reports and Share of Wallet/Purchases (measured in a period of panel data). This is re-discovered every so often, indeed Jan Hofmeyr even won an award for it a few years ago. This re-discovering (and misinterpretation) over and over says a lot about the sad state of our discipline given that the relationship was very well documented in the 1960s by Bird, Channon, and Ehrenberg (see references below).
Like Hofmeyr & colleagues (2008) the HBR team have done a rank order correlation which, in this case, guarantees high correlations of around 0.9. Correlation is not a good way of assessing predictive ability (which I illustrate here). Yet they report that their analysis is “to our knowledge the largest and most rigorous of its kind”.
The article goes on to imply that great financial returns can be gained from improving loyalty of existing customers. They recommend that managers move from a focus on the drivers of satisfaction to focus on share of wallet rank. This takes a leaf out of Reichheld’s book, when he said forget about satisfaction and focus on willingness to recommend (Net Promoter Score). In this case it’s true that larger market share brands do have higher share of wallet rank. What they fail to mention is that such loyalty metrics move only a little, and do so when a brand's penetration increases dramatically – the Double Jeopardy law.
In effect they rely on an old “wouldn’t it be great” argument that goes along the lines of “if those customers who give us very little of their spending became near 100% loyal, then we'd earn X more money”. It’s like saying “if you win the lottery then you’ll be richer”, all logically/analytically/theoretically true but it’s easy to make theoretical statements that clash with the empirical world, as this does. It's all beautifully academic, but not scientific nor practical. It’s not how brands grow.
Note how costs are also conveniently ignored. In the same way that Reichheld ignored them when promoting his retention and profitability myth.
The authors recommend finding out the motivational reasons why people give only a small amount of their spending to your brand and lots to others. But this ignores the weak link between claimed motivations and behaviour.
In the article they present a grocery store as an (fictional, hypothetical) example - why they didn’t (couldn’t?) show a real case is rather telling. In their hypothetical example the store drops prices on staple items to encourage low share of wallet customers to give them more of their supermarket shopping trips. Now supermarket repertoires have what is called high "1st store loyalty", in that the top store in anyone's repertoire tends to dominate, with the remainder of their repertoire being divided amongst several others. From the store's perspective then, it has a large group of customers who are very loyal in the sense that they do most of their grocery shopping there. And then another group of customers who largely do their shopping at another store. The former group tend to live close by, the latter further away (a fact that is ignored in the article).
So TIm Keiningham & co are on sound ground in recommending a focus on the people with low SCR - sales growth does indeed largely come from the latter group (since obviously it's hard to get more out of people who already give you nearly all their custom). But the real story for growth is, target your light and non-customers. Go for reach. Increase your mental and physical availability (e.g. open longer hours, open another store, make it easier for people to shop and shop faster...).
That's a very different story than the old-fashioned loyalty one (re)presented in the HBR article. And the promise of riches to be gained from a share of wallet focus…..that’s a chimera.
Professor Byron Sharp
October 2011
BIRD, M. & EHRENBERG, A. 1966. Non-Awareness and Non-Usage. Journal of Advertising Research, 6, 4-8.
BIRD, M. & EHRENBERG, A. 1966. Intentions-to-Buy and Claimed Brand Usage. Operational Research Quarterly, 17, 27-46.
BIRD, M. & EHRENBERG, A. 1970. Consumer Attitudes and Brand Usage. Journal of the Market Research Society, 12, 233-247.
BIRD, M., CHANNON, C. & EHRENBERG, A. S. C. 1970. Brand image and brand usage. Journal of Marketing Research, 7, 307-314.
BIRD, M. & CHANNON, C. 1969. Brand Usage, Brand Image, and Advertising Policy - Part I. Admap, 6, 27-46.
BIRD, M. & EHRENBERG, A. 1972. Consumer Attitudes and Brand Usage - Some Confirmations. Journal of the Market Research Society, 14, 57.
HOFMEYR, J., GOODALL, V., BONGERS, M. & HOLTZMAN, P. 2008. A new measure of brand attitudinal equity based on the Zipf distribution. International Journal of Market Research, 50, 181-202.