Customer Retention with Predictive Analytics:
How Profitable Is It?
by Eric Siegel, Ph.D.
This article describes and links to the forecast model spreadsheet introduced in the DMReview article, Predictive Analytics' Killer App: Retaining New Customers.
How much higher would your revenue climb if you could predict which customers are likely to churn? Predictive analytics targets customer retention campaigns to ensure bottom-line ROI. By predicting which customers are at risk for defection, campaign dollars are applied effectively -- without predictive targeting, a retention campaign may cost more than it gains.
As detailed in this author's DMReview article, Predictive Analytics' Killer App: Retaining New Customers, if we predict which first-time customers won't return, we can target them with an aggressive conversion campaign, enticing those that would otherwise leave with a heavy discount or free trial. In this way, we turn the first date into a long-term relationship.
A predictive model tells us which new customers are likely to return and which are probably one-timers. The model is created with data mining methods that "learn" from the collective experience of your company recorded in your sales records. The model then applies what's been learned to produce a predictive score for each new customer, in real time. With this in place, new customers we would otherwise never see again are targeted and enticed to stay. Since we don't waste the retention offer on new customers likely to return, the numbers work out very well: The growth rate and medium-term profits potentially skyrocket, and immediate-term profits are not put at risk.
For an overview of this killer app of predictive analytics, see the published article. But that article doesn't tell you how to forecast expected profit -- it links back to this webpage to show you how.
To calculate how much profit a predictive retention campaign will generate, you need a forecast model. That's right, it takes another kind of model to tell you the profitability of a predictive model. This is because profit depends on more than just how accurately the model predicts -- it also depends on how it will be used:
- The total number of customers
- How much the retention campaign will cost per contact
- How much profit will be gained by each customer successfully retained
- The response rate of the campaign, which depends on the design and execution of each customer contact
The spreadsheet forecast model below combines these factors to calculate forecasted revenue for a predictively targeted retention campaign. It also forecasts the revenue for the same campaign deployed to all customers, i.e., without predictive targeting. The second value, which often shows a loss in revenue, serves as a basis for comparison.
If the retention campaign is targeted only to those customers predicted to not return, the overall profit increases. Assume our business currently retains 20% of new customers. Predictive modeling may discover a series of sub-segments that together identify 60% of new customers, each of whom are 40% more likely than average to return. This is a realistic result, and a valuable one at that, enough to overcome the loss in revenue that comes from offering a rebate. The spreadsheet is set up with these values, and assumes a high response rate of 20% to an aggressive campaign that offers a hefty coupon of 50% the customer's next purchase.
With this forecast model, you can plug in your own numbers and try different "what-if" scenarios. The numbers you plug in can be based on the results of true modeling and test campaign results, or can be based on business objectives you intend to achieve.
Download the forecast model spreadsheet
You can download the spreadsheet in two ways:
- Excel spreadsheet version. Forecast model with a graphical bar chart of results.
- Google spreadsheet version. This may save you several seconds of downloading and starting Excel, but you must have or create a free Google account, and it does not support the graphical bar chart.
About the Author
Eric Siegel, Ph.D., is President of San Francisco-based Prediction Impact, providing analytics training and services to medium through Fortune 100 companies. He is an expert in data mining and predictive analytics and an award-winning teacher of graduate-level courses in those areas. Siegel served as a computer science professor at Columbia University, where he forwarded data mining technology in the realms of machine learning performance optimization, text mining and data visualization. He cofounded two New York City-based software companies for customer/user profiling and data mining. With data mining, Eric has also solved problems in computer security, fraud detection, computational linguistics and information retrieval. You can reach him at firstname.lastname@example.org or (415) 683-1146.