Customer value (see jargon buster, page 26) is rightly regarded as a key component of CRM. However, it is only one of a range of measures required to develop and optimise your customer management strategies.
In highly fragmented markets, such as retailing or financial services, where the customer may have a wide portfolio of suppliers, the customer value alone can only provide part of the picture. You need instead to understand the value of the customer to the market place (or customer market value - see jargon buster) and your share of their wallet (see jargon buster). After all, it's difficult to develop a customer when you already have a large share of their wallet.
Understanding share of wallet helps you decide which customers may be worthy of further investment and what marketing strategies may be appropriate for the different groups of customers. For example, 'high spending, high share of wallet' customers are obvious candidates for tender loving care with the emphasis on retention. Their counterparts, the 'high spending, low share of wallet' customers, are important to develop as prospects.
In this article we look at how to use predictive modelling to estimate customer market value and calculate share of wallet.
The principle of understanding customer market value applies equally to prospective customers as to existing ones. Understanding a prospect's market value helps you manage your customer acquisition programme. The ideal is to balance investment in acquiring a customer with knowledge of their potential value.
The process we describe of predicting individual-level share of wallet has been used in various industry sectors. It relies on surveying a representative sample of customers (and/or prospects) to understand their overall spend in a market, and then uses that data to create predictive models to forecast the market value of all the other customers or prospects on a marketing database.
Define the customer
It is important in the first instance to define the customer, in terms of individual or household, and the market for which the spend is to be predicted. For example, in fashion retailing an individual's market value is often the target variable, whereas in financial services, where many of the product decisions are shared, it is often better to target the household market value.
Construct a sample design
The survey process will create data with which to build the predictive models which project customer market value. It is vital, therefore, to construct a survey sample which is representative of the whole customer or prospect base. To construct the models will require a sample of about 2,000 customers and you can expect between 12 per cent and 20 per cent response to the survey. (The case study on page 25 achieved 15.9 per cent).
It will also be important at this stage to weight the sample to ensure you get responses from a representative sample of the whole customer base.
You will find that customers with weak brand relationships will generally respond less well than the loyal and committed customer. Typically, the best customers respond four times better than the worst.
Questionnaire design
Asking customers to help you calculate their spend in a given market is not easy. The choice of survey channel and the construction of the survey are critical. In financial services customers often have a record of their spend in each product area. In this case you will find it easier to use a postal questionnaire or e-mail survey and to ask a series of questions, such as: Do you have an ordinary building society account?
If so, who is it with and what's the current balance?
For fashion products it is less easy. Here it's often better to use a telephone survey and to break the market down into product areas and ask a series of questions regarding each product in which you are interested.
For example, a telephone survey can ask the customer: when did you last buy a suit?, where did you buy it from?, how much did you spend?
This way you build up a detailed understanding of their expenditure patterns.
The results from the survey can be combined with internal transactional data of customers' spend with you to provide a fascinating insight, not only into their relationship with you, but also their value and relationship with your competitors.
Apply the customer market value models to your database
The next step is to create a customer market value model which uses internal data, and sometimes bought-in lifestyle data, to predict customer (or prospect) market value for each individual on the database.
The models can be built using any of the multi-variate techniques employed by predictive modellers which are then applied to your database. You can now estimate market value and share of wallet.
On a simple box grid (see above) you can describe what your customers' or prospects' value is to you, their value to the market, and your average share of wallet. It's possible to draw an interesting map of your customers and their value to you and to the market as a whole.
- Julian Berry is a partner at The Customer Partnership
CASE STUDY: SHARE OF WALLET
A top 10 building society with 4.5 million customers was seeking to understand share of wallet as part of a marketing strategy review.
We mailed 65,000 customers with a questionnaire (un-incentivised) which sought to understand the value of their relationships with other financial service providers (with questions relating to demographics and household composition).
The questionnaire reviewed their savings, lending and insurance products and for each product area asked which supplier(s) the customer was using and current total balance or premium paid.
In total, 9,000 replies were received and were merged with an extract of the main marketing database. The merged file was used to create a series of models to predict customer value for each product area.
The output allowed the society to plot the average segment value and share of wallet.
We used the information to paint a portrait of each segment and the society's share of wallet position for each product area.
The main outcome of the project was that the marketing team were able to understand the untapped value of the customer base and target those customers with high opportunity value.
Over the next three years marketing investment was gradually switched from customer acquisition to customer development and for the savings team greater emphasis was made on targeting high-value, low-share, wealthy investors.
JARGON BUSTER
Customer value
The value of the customer to you (usually expressed as an annual revenue or contribution figure).
Customer market value
The value of the customer in the market place, for example, the value of the customer to you and your competitors in your specific market place (usually expressed as a revenue figure).
Share of wallet
Your share of a customer's expenditure in your market, expressed as a percentage (customer value/customer market value) x 100).