Hitting the Bull's-Eye: Using Uplift Modeling to Improve Marketing Campaigns

Source: Inside Direct Mail

By Mark Smith

In the game of marketing, relevancy is king. Ensuring your message is targeted to your customers' needs should always be a top priority, and how well you achieve this relevancy will have an immense impact on the success of your campaigns. Equally as important, however, is ensuring your audience is correctly targeted.

Even if the message is on target, your campaign will be wasteful and less successful if you're reaching out to the wrong people. Blanket, untargeted marketing campaigns are costly, ineffective and can even be detrimental to sales. Conversely, more targeted campaigns, fueled by intelligent analytics, can help companies reduce marketing costs while also increasing results.

Today's marketers must equate marketing dollars spent to revenue gained through targeting customers. This is sometimes a struggle because traditional modeling does not account well for the customers that may respond negatively to marketing campaigns.

Understanding how customers truly respond to a campaign is the key to success. This insight will ensure that marketers only target customers who are most likely to react positively to the marketing message, avoiding all others.

Traditional response models effectively assume that all purchases during, or slightly after, a campaign are incremental, i.e., would not have happened if the campaign had not been carried out. They also implicitly assume that no purchases are lost as a result of the campaign. History has taught us that this is simply not true - conventional response models can be misleading and new; more sophisticated uplift models (incremental models) are shown to perform much better.

Uplift modeling is an analytic-based approach to marketing that predicts the difference a marketer's actions will make on the behavior of his or her customers. This approach divides your audience into segments based on the predicted difference in response to a marketing campaign when compared to a control group. This allows marketers to focus efforts only on "the persuadables," those who are likely to respond to marketing by buying (or renewing), but wouldn't have if you hadn't contacted them.

Similarly, uplift modeling prevents businesses from wasting time and money targeting those customers that are already "sure things" and will buy regardless, as well as those that are "lost causes" and will never buy. The "sure things" would be targeted by a traditional response model, significantly decreasing the return on marketing investment of the campaign.

Uplift modeling also prevents companies from mistargeting customers completely, sending a marketing piece that could inadvertently irritate them and provoke them to opt out of email lists or join "do not call" lists, making further contact more of a challenge and reducing future selling opportunities. In the worst-case scenario, an irritated customer may even leave the company and take his or her business elsewhere. In all cases, unnecessary marketing contact is detrimental to the customer experience, and to the long-term profitability of the business.

Uplift modeling identifies these critical customer response segments before you run a campaign, so you can target "the persuadables" and leave everyone else alone. Companies can therefore reduce the resources invested in the campaign while improving campaign results, significantly boosting overall ROI.

U.S. Bank, for example, has benefitted significantly from incorporating uplift modeling into its marketing plan — improving net revenue from its cross-sell campaigns by more than 300 percent, while simultaneously reducing program costs by 40 percent. Incorporating uplift modeling into marketing campaigns is beneficial for environmental reasons as well. Blanket marketing campaigns can be incredibly wasteful in terms of environmental resources. Uplift modeling helps marketers cut down on direct mail and exemplify "green marketing."

In summary, uplift modeling allows companies to:

• Avoid marketing costs for customers who will not respond to a campaign (and for those who may be negatively affected by it).
• Avoid giving unnecessary incentives and discounts to customers who are likely to purchase regardless.
• Correctly prioritize those customers most likely to respond positively to a message or offer.
• More accurately estimate the profitability of campaigns.

We've found that uplift modeling is a powerful tool that enhances the marketing experience — for both the marketer and the customer — by providing valuable insight into customers' specific needs and helping predict their behaviors in response to marketing. By employing uplift modeling, marketers can structure their campaigns with only relevant messages and targets to reduce costs and improve the ROI and overall marketing success.

Mark Smith is executive vice president of sales and marketing at Portrait Software, a leading provider of customer interaction management software. Portrait enables organizations to engage with their customers as individuals, resulting in improved customer profitability, increased retention, reduced risk and outstanding customer experiences.