Shift the focus away from loyal customers

Too many marketing campaigns focus on customers who are already loyal. This gives high response rates but near-zero effect. Campaigns add real value only when they convince indecisive "swing customers" to make a purchase.

Kim Larsen, internationally recognized marketing analyst from US-based MarketShare Partners, references the US presidential election and the concept of "swing states" to introduce "swing customers" as the marketing department's most valuable target group.

Larsen is a number cruncher who is an expert in drawing knowledge from customer and market data. His message is clear: marketers must work much more analytically with data in order to identify the swing customers who increase a campaign activity's net worth.

"The vast majority of marketing campaigns are designed to achieve maximum response rates," says Larsen. "But a campaign's effectiveness can't be measured in the number of responses it generates. Instead, we should be looking at the value the campaign creates for the company. I often observe campaigns targeted toward "yes men" and those loyal customers who probably would have bought the product anyway. Aiming campaign content and messages at this group yields minimal effect."

Digging for gold
"The customers you want to reach are swing customers – the undecided ones who would not react unless they were exposed to the campaign. This is the group that delivers real value on marketing investments," Larsen told conference participants. He demonstrated his point with the case of a company that wanted to measure the effectiveness of sending a special offer to customers who visited the company website. The company ran a 90-day test with a test group that received the offer and a control group that did not.

"The initial test group results were quite convincing; 5.01 percent purchased the product after having received the special offer. The company broke out the champagne and was ready to launch the campaign. Then the results from the control group came in. These showed that 5 percent would have bought anyway – even if they did not receive any special offer. The actual effect of sending the special offer was a completely marginal increase in purchases of 0.1 percent, so the company decided to drop the campaign. The very premise of the campaign was wrong; customers who visit a company's website are, a priori, interested. So it makes no sense, from a profitability perspective, to invest campaign dollars there."

A needle in a haystack
According to Larsen, customers can be segmented into three groups:

• Those who are not interested, whether or not they are exposed to a campaign.
• Those who are loyal and would buy the product no matter what.
• Those who are on the fence – the swing customers.
  

The $64,000 question is how to isolate the swing customers as a target group. They are not immediately visible, since customers that react positively to a campaign can belong to either the loyal customer group or the swing customer group. The answer lies in the control group data. Larsen is a strong advocate of getting marketing departments to use this data.

"My recommendation is to expand campaign testing so that it includes a control group, just as pharmaceutical companies do in clinical trials. The purpose of the control group is to reveal purchase behavior among customers who were not exposed to the campaign; this contributes valuable data that can be used to isolate swing customers. The really interesting customers are those in the control group who do not make any purchase during the test. These are the ones we might be able to motivate toward purchase through exposure to the campaign. When control subjects do make a purchase, on the other hand, we should just be happy about it. But we shouldn't use our campaign budget on them since they would have bought anyway," explains Larsen.

Obama vs. McCain
Once swing customers have been identified within the control group, the next step is to use data analysis to identify other customers whose profile matches that of the swing customers. Here, relevant data can be demographic, macroeconomic, customer-based (which products does the customer already own?), behavioral (where does she shop?) or web-generated (what did the customer do on the company website if she looked at a product but did not buy it?). Another important source is data from social networks.

The resulting analysis method, Net Lift Modeling (NLM), is widespread among American businesses and the strategy behind NLM is widely used by political campaign organizations.

"The US presidential election was a textbook example of the principles behind NLM," Larsen recalled. "In San Francisco, where I live, we didn't see a single TV commercial for either Obama or McCain during the entire campaign. The explanation is simple: San Francisco is Obama country. Both Obama and McCain knew this, so why should either of them use their marketing budget there? When I was in Las Vegas, on the other hand, the situation was completely different and both candidates went all out. Nevada is a swing state and every single campaign activity can determine whether the vote will go one way or the other. Thus, there are real rewards to be won by investing heavily in this area."

Larsen prefers traditional probability models with test groups when researching campaigns for products with low market visibility. In these cases, control group reactions will be so limited that there will be insufficient data to analyze reactions to the campaign. As product awareness grows, including control groups in the testing becomes more and more relevant.

"I would always begin by considering the traditional method. But if this doesn't work – and this is more the rule than the exception in my experience – then I would switch to NLM," concluded Larsen.

The results illustrated in this article are specific to the particular situations, business models, data input, and computing environments described herein. Each SAS customer’s experience is unique based on business and technical variables and all statements must be considered non-typical. Actual savings, results, and performance characteristics will vary depending on individual customer configurations and conditions. SAS does not guarantee or represent that every customer will achieve similar results. The only warranties for SAS products and services are those that are set forth in the express warranty statements in the written agreement for such products and services. Nothing herein should be construed as constituting an additional warranty. Customers have shared their successes with SAS as part of an agreed-upon contractual exchange or project success summarization following a successful implementation of SAS software. Brand and product names are trademarks of their respective companies.

Copyright © SAS Institute Inc. All Rights Reserved.