The huge volume of data collected routinely by most businesses today presents a real challenge. It's not so much the costs of storing the data that keeps the CIO awake at night but the complexity of generating competitive advantage from the data.
A simple four-step process would help most businesses, no matter how small or large, to generate value from the data they already have.
Step 1: Get your data in order
Ask around in any company and you will generally find plenty of customer and other data. The trouble is it will be all over the company. Sitting in unconnected databases. In Excel spreadsheets. Even on paper. The first step in using customer data as an asset is to understand exactly what data you have and to bring key data together. Consultants have made much ado about developing a "360-degree view of the customer." The idea is sound; without all that customer data, it is hard to have a complete picture of the customer. But you don't need all that data for targeting a marketing campaign. You only need a subset. My own experience is that you typically need only 10 percent of what is available for most tasks. But which 10 percent?
One of the best ways to find this out is to construct a simple analytical data mart together with the data's business owners. By pulling a sample of customer data into the data mart, you can quickly develop a feel for what data is available, what state it is in and how it relates to other data. As the name suggests, the analytical data mart is the foundation of future analysis.
Step 2: Explore the data
Having built the analytical data mart, your first task is to explore the data using simple exploratory data analysis tools. This will help you understand key relationships between different types of data. Much of the analysis can be carried out with simple SQL queries, extracts to Excel spreadsheets and Excel's graphical capabilities.
For example, Toyota in Germany developed an analytical data mart as part of a customer loyalty project. Exploring the data, Toyota found that most customers repurchased new vehicles within one month of the purchase birthday of their current vehicle. More importantly, the peak repurchase month was Month 36, exactly three years after the vehicle was bought. Just knowing this simple fact enabled Toyota to replan the timing of repurchase mailings to take place four months before the 36-month birthday, rather four months after, as had been the case.
‘ Just knowing this simple fact enabled Toyota to replan the timing of repurchase mailings. ’
Step 3: Build simple predictive models
Now that you have explored the ins and outs of the data, your next stage is to develop simple statistical models to predict customer behavior at key points in the lifecycle. Of course, you cannot actually predict individual customer behavior, no matter how much customer data you have. But you can identify groups of customers with a higher probability to behave in a certain way.
For example, Toyota took customer data from previous repurchase campaigns and developed simple statistical models to predict which customers are more likely to repurchase in the future. The models identified the characteristics of customers who repurchased before and used this insight to identify, say, the 30 percent of customers most likely to repurchase in any given month in the future. The models were used to target customers in a number of repurchase campaigns to great success: The campaigns doubled the repurchase rate but at only one half of the volume (and cost) of mailings, a net 400 percent increase in campaign ROI.
Step 4: Bring customer events into the models
By now, you will probably have taken the data you need from the larger analytical data mart into a focused marketing data mart. Now you can enrich the data in the marketing data mart with carefully selected data to increase the accuracy of campaign targeting and to increase the responsiveness to campaigns. Some of the best data is about discrete customer events: events in the life of the customer that drive key sales behavior in customers.
Further analysis of repurchase data by Toyota identified that the birthday periodicity in repurchase behavior was driven in part by customers having a finance contract with Toyota's finance arm. In particular, Toyota found that the end-of-contract date was strongly related to repurchasing a new vehicle. Toyota developed a campaign targeting customers prior to the end of contract, to persuade them to repurchase another Toyota and to finance it through Toyota's finance arm. Dealers were heavily involved in designing and operating the campaign. This way, Toyota built on an existing relationship between customers and dealers, strengthened its own relationship with dealers and sold more vehicles, too. The results were spectacular, with more than 30 percent of customers repurchasing a new vehicle within a month of the end of contract.
None of the examples above required complex analytical capabilities to develop and run. Indeed, many of the campaign were arguably the result of plain common sense. But simple analytical tools increased the campaigns' accuracy, reduced the volume and cost of mailings, and greatly increased their response rate. That is the power of data-driven marketing.
Before you run your next marketing campaign, ask yourself a few simple questions:
- Do you know what drives your customers' behavior?
- Do you have all the key data to develop and run effective campaigns?
- Can you predict which customers are in the market at any given time
- Can you identify the right customers to target, with the right offer, at the right time?
- Have you involved all the right people in planning, developing and executing the campaign?
If you answer "no" to any of these questions, you are in danger of wasting much of your direct marketing budget. We all know what that can mean in these difficult times for marketers.
