Data defines the model by dint of genetic programming, producing the best decile table.

Predicting Share of Wallet without Survey Data
Bruce Ratner, Ph.D.

Share of wallet (SOW) – the percentage of a customer's spending that a company captures from the customer – is the prime measure used in planning sales and marketing strategies. Targeted sales and marketing efforts to large SOW customers yield increase in SOW with greater efficiency than the second important measure, share of market. (Share of market is the percentage of the total available market that is being capture by a company; i.e., a company's sales revenue/units divided by total sales revenue/units in the market.) SOW is important: It allows monitoring a company vital signs, such as, customer loyalty, levels in attrition, retention and profit. Total SOW is all-important: It allows monitoring a company vitals with respect to its competitors, such as, customer loyalty, trends in attrition, retention and profit Calculating Total SOW is difficult because competitors’ customer-level spending data are not available. There are two approaches for obtaining Total SOW estimates.

    1. Aggregated Data – are public, macro-level industry-aggregated data complied from market research studies, which include industry metrics, such as, industry revenue, competitor analysis and market share, and product and customer segmentation. Most important, it also includes major spending categories (in units and dollars), and many two-dimensional tables of a major spending category by a socio-demographic variable. The data analyst seeks to derive Total SOW from its company SOW by unfolding the macro-level spending results into its summands, and thereby calculating weights. The weights make possible the interpolation of a company Total SOW estimate from a company SOW. Total SOW estimates are arguably unreliably, as suggested by often derogatorily reference of the weights as fudge factors. The weights are needed to force calculating micro-level (i.e., some degree of unfolded customer metrics). This approach is often attempted first because it is quick and inexpensive to complete; but, it typically yields poor results.
    2.  Survey Data – are expensive market research survey of a random sample of company customers to obtain each respondent’s complete spending behavior (e.g., transactions, and the like). Although caveated with the adverse effects of incentives (usually monetary), and self-reporting (bias due to the nature of the information sought, a person’s financial goings-on), survey data provide close-to-true findings, from which customer-level, survey-based reliable weights are obtained. The survey, costly and time-consuming, is the approach of choice for honest Total SOW estimates. 
The literature apparently is void of any proposed method to predict SOW without survey data. The purpose of this article is to present a modeling approach – using the GenIQ Model© – for predicting SOW, and predicting the unobservable Total SOW without survey data

For more information about this article, call me at 516.791.3544, or e-mail,
My publisher owns the copyright of the article, about which this abstract addresses. The article will appear in my forthcoming book.
My publisher has granted me permission to discuss orally the article's content, but by no means provide an outline, a draft or proof-ready of the article.

Sign-up for a free GenIQ webcast: Click here.