Share of Wallet Analysis

Market share is an aggregate, firm-level metric: your total sales as a fraction of total category sales across all customers. Share of wallet is a customer-level metric: for an individual customer, the fraction of their category spending that goes to you. The two can diverge sharply. A firm can have a respectable market share built from many customers who each give it only a small slice of their spending, leaving large untapped potential inside the existing base. Conversely, a niche brand might have low market share but dominate the wallets of its few customers. Because share of wallet is measured per customer relative to that customer's full category spending, it reveals loyalty and growth opportunities that aggregate market share hides, which is precisely why it is treated as a loyalty metric rather than a size metric.

The Wallet Allocation Rule predicts a customer's share of wallet from how the brand ranks among the brands that customer actually uses and how many brands they use, rather than from the brand's absolute satisfaction or recommendation score. Its logic is that customers split their money according to relative preference: being ranked first among three brands earns a large, predictable share, while being ranked third earns little, regardless of a high standalone satisfaction rating. Traditional satisfaction and Net Promoter scores are typically measured for one brand in isolation and ignore that the customer also rates competitors, which is why they often correlate weakly with actual spending. Keiningham and colleagues showed empirically that the rank-based rule explains and predicts share of wallet far better, because spending is a zero-sum competition for the customer's budget.

Measuring the denominator, the customer's full category spending across all brands, is the central challenge of share of wallet analysis, since a firm's internal records only show its own sales. Common approaches include consumer or household panels that track purchases across brands, shared or aggregated transaction and account data (for example financial-account aggregation in banking), syndicated data from third parties, and customer surveys that ask respondents what fraction or amount of the category they buy elsewhere or how they rank the brands they use. Each method has trade-offs between cost, coverage and accuracy, and survey-based estimates in particular are subject to recall and social-desirability bias. Because the denominator is usually estimated rather than observed, careful category definition, validation against any available behavioral data, and treating the results as estimates are essential.