<img height="1" width="1" style="display:none" src="https://www.facebook.com/tr?id=204513679968251&amp;ev=PageView&amp;noscript=1">

Do You Even Data

A data-driven marketing blog

Want to learn how you can translate incredible data list information into killer marketing campaigns? Want to better understand how data research and models can enhance the data you already have?
All Posts

Combine Market Research and Predictive Models to Improve Cross-Selling


Executive Summary

Today's member-based organizations (or any direct marketing company) can combine predictive analytics with research to efficiently identify which members will respond to cross-selling, which members won't, and which sales messages will be most effective.  The result is optimum sales, reduced direct marketing costs, and an ROI that is higher and demonstrable.

Editor's Note: This blog topic applies to any company that relies on direct marketing via digital, social, mobile, email, direct mail and tele-marketing.

Cross-Selling Challenge

If your member organization relies upon a revenue stream from training, certifications, resource materials, and other value-added benefits, then it is critical to aim your marketing at the exact members who will respond while skipping the ones who won't. Otherwise, the marketing ROI you can expect from your social, digital, email and direct mail efforts is at risk. 

 This waste can be avoided if you only knew how to predict who is likely to respond and who isn't. In addition, you can further improve your marketing performance with precise messages that resonate with each unique member (or customer).

The Power of Predictive Modeling

Member organizations are turning to predictive analytics to boost their marketing results. Without further delay, I will lay out four key steps to achieve analytics-driven cross-selling success.

1 - Gather Your Internal Data

Assemble your member data and catalogue the data points captured for each member.  This usually includes sign-up date, membership level, and transaction data about past purchases. In business-to-business applications, just as important are demographic information and firmographics such as company size, sector, tenure, and title.

2 - Develop Your Predictive Model

Using responder and non-responder data, develop a predictive model that uses the available member (or prospect) data points to accurately identify which members (or prospects) are most likely to purchase you resources and takes your courses. This is often referred to as a lookalike model, a cross-selling model, or a response model.

3 - Enhance Your Member Data

In order to score and rank each member by their likelihood to buy (or respond), you must first enhance your member file with the data used by the predictive model. This typically includes third-party data, such as our DecisionPointsTM data source. This data includes behaviors, demographics, lifesytyles, channel preferences (social, digital, email, direct mail) and other data that has predictive value in the model.

4 - Deploy Your Model

Deploy your cross-selling model by scoring and ranking each member by their likelihood to buy (or respond). You now know your best prospects, in rank order, to respond to your marketing, sign up for your courses, apply for certifications, and purchase your resources. You can invest direct marketing dollars in these members with confidence that you have minimized waste, optimized purchase, and increased your ROI. 

Cross-Selling Case Study — Predictive Modeling In Action


We've just described a process that we know works. Here is how it worked for one company.

Challenged by the Competition

A large association competes with a private company to certify members' knowledge levels in a professional discipline.  It made business sense for our client to reach their non-certified members with information about the value and benefits of their certifications before they lost them to the competitor.  But with a large member base, the direct marketing costs would be enormous.

Competing with Predictive Analytics

The association looked to us to help them reduce their direct marketing costs by precisely identifying their members who had the highest likelihood of responding to the professional certification offer.  We helped by analyzing their member database to identify and profile those who were already certified and in this way, created a predictive model of Likely Purchasers.  We then applied this model to the database of members who hadn't applied for certifications and scored the non-purchasers on their likelihood to apply.  To help them create targeting priorities, the modeling broke the group into deciles based upon response likelihood.

Challenge Met

The results of our predictive modeling made their direct markets extraordinarily effective and greatly improved the ROI. We helped our client target their marketing more directly, reduce their direct marketing costs, and increase their ROI.


The success is illustrated above in Figure 1, which shows how predictive modeling improves the response rates over a non-modeled direct marketing.  For example, the response rate of Decile 1, representing most likely to respond, indexes at 175—75 percentage points higher than a non-modeled (Random Lift) Decile 1.

Marketing Research Improves Cross-Selling Results

Want to make cross-selling even more efficient?  Of course!  This is where research and predictive modeling are used in combination to match likely buyers to sales messages that are most compelling to them and their situation.  Here are the steps: 

1 - Recruit a Random Sample of Members

Starting is easy—pull names and emails from the Likely Purchasers list that predictive modeling identified up above.  You will use this as the sample source for a survey.

2 - Identify Pain Points and Hot Buttons

Use survey research to identify the Likely Purchasers' expectations, preferences, pain points and other data that can help marketing create pin-point messaging. Then, take these additional steps:

Create Segments: Analysis identifies segments based upon similar pain points, hot buttons, demographics, and firmographics.

Create Predictive Models: Analyze each Likely Purchaser segment to identify those tell-tale attributes that distinguish them from the other purchasers to make a predictive model for each group.

3 - Segment Likely Buyers

Now take your entire list of Likely Purchasing members and score them with the models to assign them to the proper Likely Purchaser segment. Then, use the pin-point messaging that Step 6 helped to customize a campaign for each segment. 

4 - Launch

Now your organization can launch a direct marketing campaign to Likely Purchasers with confidence that you’re addressing their particular needs most effectively. 


Associations and member organizations can increase their marketing ROI by using predictive modeling. Intelligently used, predictive modeling can identify the members who are most likely to respond to direct marketing and avoid the cost of targeting the less likely.  The result is increased revenue with a much better ROI. 

Predictive modeling can also be combined with research to create customized campaigns for different types of purchasers.  The research determines which type of purchaser a member is and predictive modeling scores each member with the hot buttons that are most compelling to them.  Armed with this knowledge, marketing can proactively create custom messages to each segment.

Learn More >>

Contact us today to discuss your cross-selling needs and how to combine predictive analytics and marketing research to get the best results.

Request Info

David W. Wilson
David W. Wilson
David W. Wilson

David Wilson has over 25 years of experience helping leading companies improve their marketing results using digital marketing, direct marketing, database marketing, consumer data, predictive analytics and marketing research.