A successful B2B marketing plan uses the power of predictive analytics for lead generation, to get more value from existing leads, determine which leads are most likely to convert, predict which customers are likely to churn, and more. In this post we’ll explore some of the ways predictive analytics can boost your marketing efforts and drive more ROI from your marketing technology investment.
Editor’s Note: This post is excerpted from The Buyer’s Guide to Predictive Analytics for B2B Sales and Marketing, by David Raab. David has more than thirty years experience as a marketer, consultant, author and analyst. His firm, Raab Associates, Inc., helps marketers select the best marketing technologies and service vendors. You can download the report free of charge here.
Re-energize marketing efforts with new prospects
Prospect discovery includes identifying “net new” names of potential customers for your product. This has traditionally involved building models against compiled lists enhanced with demographic data such as title and department or business size and industry. More recently, marketers have been able to access additional information such as interests and intent, generally derived from Web behaviors such as content consumption and social media postings.
The general approach has been to start with a list of existing customers, enhance them with external data, build a model with the external data, and then apply that model to select prospects with similar characteristics from a universe the vendor has identified by assessing all firms and, in some cases, individuals on the open and social Web.
Determine which leads are most worth your time
Lead scoring means estimating which current leads are most likely to become customers. This is used to prioritize marketing and sales activities to achieve the best results. In practical terms, this lets salespeople concentrate on the most promising leads while other leads are managed through lower-cost marketing contacts such as email nurture campaigns.
The process in this case is to start with a database of past leads and build a model that identifies those who eventually reached a goal such as sales acceptance or making a purchase. Scores can be assigned to individuals or companies, can be based on both behaviors and attributes, and can use external as well as company-owned data. Behavior-based scoring measures “engagement” and changes over time; attribute-based scoring measures “fit” against an ideal customer profile and is usually stable. Some systems assign separate fit and engagement scores.
Enrich your leads with better, fresher data
When we say enrichment, we mean appending data to existing customer profiles to provide insights about needs and interests. This data is derived from both structured as well as open and social Web sources.
Some data can be read directly from those sources, such as company, title, address and telephone number. Other data may need advanced technology to extract into a usable format, such as technologies used (based on job requirements), expertise (derived from past jobs and experience), important events (taken from press releases, blogs and news articles), and growth rate (based on changes over time).
Open doors with the right message
Predictive analytics can drive recommendations: predictions such as products a company is most likely to purchase, messages they are most likely to respond to, or channels they prefer to communicate through. The predictions are based on detailed data such as purchases of individual products or responses to different types of offers.
They often apply collaborative filtering techniques that depend more on correlations among product purchases than on correlations between individual attributes and results. They may also be based on enrichment information, such as opening a new office or launching a new product, that implies a certain set of needs.
Download The Buyer’s Guide to Predictive Analytics for B2B Sales and Marketing to learn more, including objective criteria for selecting the right predictive analytics platform for your business.
Determine which leads have the most potential
Lead intelligence derived from predictive analytics provides guidance to sales and marketing teams about how to treat individual leads. The key element here is predicting which leads are close to initiating a buying project, making a purchase decision, or otherwise responding to a contact.
Models are built by correlating behaviors such as Web searches or page views and social media comments with important outcomes such as requesting a proposal or making a purchase. The goal is to help internal teams apply their time to the most pressing needs. Lead intelligence can also be informed by the product, offer, and channel recommendations provided by content models.
Know which customers are likely to churn
Predictive analytics can help drive customer satisfaction and retention, by helping customer success, account, service, and support teams anticipate the needs and actions of existing customers. This relies largely on internal data to predict several types of customer behaviors: how they’ll use the system, what kinds of help they’ll need, what additional products or modules they might buy, and whether they’ll continue the relationship.
The core application is churn prediction, although there are of plenty others. The required models are broadly similar in terms of correlating system usage with behaviors by previous customers. Some models may also incorporate external data, such as social comments or Web searches that indicate a customer is dissatisfied and considering switching to a new vendor or, more positively, has expanded needs that offer an opportunity for additional purchases.
Predict the impact of marketing activities on revenue
Predictive analytics helps marketers predict the impact of marketing interactions and estimate customer value. This requires data that ties interactions to customers and customers to revenue. Assembling complete views of the interactions and revenues for each customer can be challenging, and takes sophisticated analytics to accurately estimate the incremental impact of individual interactions on long-term value. But such models are increasingly important as marketers move towards automated approaches, such as programmatic media buying, which rely on accurate predictions of a marketing interaction’s impact.