Using Econometrics to Accurately Measure Share-of-Wallet and Improve Customer Strategy
By Jim Sprigg Director of Database Marketing & Analytics, Intercontinental Hotel Group | April 03, 2016
Co-authored by Matt Lindsay, President of Mather Economics
The rise of big data in an increasingly complex digital economy has economists and marketers constantly expanding what we want to know about individuals so that we can better analyze and target them with messages and offers. Yet, despite a rapid expansion in the types of customer data being offered by data vendors, the ability to know a customer's share-of-wallet remains just as desirable, and elusive, today as ever. Amidst the promises of big data, one company in the travel industry has found that traditional econometrics and store-level data is the most valuable for deriving share-of-wallet to inform customer strategy and investments.
Every marketing group in the hospitality industry would like to know which of their customers are spending a significant share of their nights with competitors, as those customers represent the largest opportunity for growth. For this reason, some companies try to estimate customer share-of-wallet, which is the percentage of a customer's spend at one company's hotels relative to all spend at all hotel brands. The two primary ways to do this are through surveys or third-party data partners. Each of these methods has strengths and limitations.
Two Common Ways to Measure Share-of-Wallet
In the case of surveys and focus groups, researchers can calculate share-of-wallet at an individual level for those customers who are willing to complete a survey, and this approach is relatively inexpensive and easy to complete with little need for data infrastructure and specialized analytical skills. This approach can be useful in tracking changes in share-of-wallet, similar to how companies track awareness, consideration and willingness to recommend. However, even if we assume that respondents can accurately recall and report their true hotel spending patterns in such surveys, which is debatable, there is a more basic limitation in using surveys for share-of-wallet.
Specifically, the customers who choose to respond to surveys are rarely representative of the broader customer base. In technical terms, respondents do not provide a statistically valid sample of the entire pool of customers; researchers call this self-selection bias. For example, if the customers who are most likely to respond to a hotel company's survey are also the ones who are most loyal to that company's brands, then the survey findings could underestimate the headroom opportunity with the broader audience of non-responding customers.