b2b lead predictive analytics
 

In this guest article, Dr Niladri Syam shares tips for successful B2B forecasting models.

The undeniable truth is that every company in B2B has only limited sales and marketing resources. The most significant expense of selling organisations is the cost of their sales force. Companies should always be on the lookout for methods that can potentially enhance the productivity of their sales operations. Salespeople often act on leads, generated either by marketing or by themselves.

Not all leads are alike – some are much more likely to result in final sales than others are. Even though it seems self-evident that sending only the high-quality leads helps sales teams close deals more efficiently, actual practice lags far behind.

According to a recent study by MarketingSherpa, “61 % of B2B marketers send all leads directly to sales; however, only 27 % of those leads will be qualified” (Doyle, 2011). Of the 27% qualified leads, the lead qualification process is often ad hoc and does not incorporate the latest available methodologies.

 
Two of the most well-established principles in marketing are segmenting and targeting. After dividing the market into several homogeneous segments, marketers identify a subset of customer segments to target. The targeted segments that are chosen are the most attractive ones in the company’s estimation. This approach ensures that the marketer achieves the most efficient allocation of marketing resources by only selectively targeting high potential segments.

 

An analogous approach is also likely to prove beneficial for the selling organisation when it comes to lead management. The idea is that the sales organisation should rely on data-based, analytical methodologies by which they can sort their leads into ‘bins’ based on the attractiveness of the leads. Targeting the most attractive sales leads avoids a situation where resources are wasted in pursuing the low percentage customers. This is the concept of ‘Lead Targeting’.

 

Effective Lead Targeting rests on a few well-established and battle-tested principles:

1. Analysis of past behavior is the strongest predictor of future behavior
2. Inferred importance weights are often more valid than self-reports
3. Resources should be spent on areas where the ‘marginal’ returns are highest

 

The good news is that social scientists have developed a whole toolkit of analytical tools that accomplish all of the above: Predictive Modeling. Simply stated, Predictive Models are designed to quantify the probability that a given customer/lead will purchase in the future. Data on past behaviour along with consumers’ preferences are inputs to these models, and the outputs are inferred importance weights. These importance weights along with the choices yield the required probabilities.

 

Predictive Models are designed to quantify the probability that a given customer/lead will purchase in the future. Click To Tweet

 

This Management Science approach to Lead Targeting is not merely theoretical. It has been applied in the field by a few companies and has met with tremendous success whenever it has been applied.

For example, Professor Dennis Gensch at the University of Wisconsin-Milwaukee, along with a team of executives at ABB, has implemented a Logit model, a type of predictive model, to obtain the probability that a customer of ABB will buy in the future (Gensch et al. 1990).

They then divided the customers into four groups based on these probabilities and identified two groups where additional spending on marketing and sales would be most useful.

 

Customers who have a very high probability or a very low likelihood of purchasing in the future are not as attractive for receiving additional resources as clients who are likely to switch to a competitor, Click To Tweet

 

Customers who have a very high probability and customers who have a very low likelihood of purchasing in the future are not as attractive for receiving additional resources as clients of ABB who are likely to switch to a competitor, say GE, and current customers of GE who are likely to be attracted to ABB. The latter two groups of customers are in play and should be targeted by ABB’s salespeople, consistent with the idea that resources should go to leads whose marginal productivity is highest.

 

Accurate forecasting of these probabilities is essential for proper Lead Targeting. Another advantage of the Management Science approach is that it is possible to do useful counterfactuals. Gensch et al. (1990) perform such a counterfactual analysis to demonstrate that their approach improves profits significantly compared to not following it. These methods have stood the test of time, and it is clear that many more companies could benefit from such a scientific approach to Lead Targeting.

 

References:
1. Doyle, Jen, 2011, “Funnel Optimization: Why marketers must embrace change”, http://sherpablog.marketingsherpa.com/b2b-marketing/lead-gen/b2b-funnel-optimization/
2. Gensch, Dennis, Nicola Aversa and Stephen Moore, 1990, “A Choice-Modeling Marketing Information System that Enabled ABB Electric to Increase its Market Share”’ Interfaces, 20 (1), 6-25.

 

syam_niladri_web
About Niladri Syam:
Before joining the faculty as an Associate Professor of Marketing at the University of Missouri, Professor Syam taught at the Bauer College of Business, University of Houston. Prof. Syam also spent two years at Tilburg University in the Netherlands. His research has appeared in Marketing Science and Review of Marketing Science. He has served on the editorial board of Marketing Science. In 2008, he was one of two professors in the Bauer College of Business to receive the Melcher Award for Excellence in Research. Bio