Sign up for our newsletter and get the latest big data news and analysis.

B2B Predictive Analytics in the Big Data Era

Keke_WuIn this special guest feature, Keke Wu of Avention, outlines the need for B2B marketing predictive models to catch up to their B2C counterparts as well as methods for doing so. Keke Wu is Director of Analytics at Avention, a leader in business information solutions. Keke was previously director of analytics at Monster Worldwide where she provided data-driven insights to a wide audience from C-level decision-makers to marketing, sales and pricing teams. She has a deep background in marketing analytics, data mining, predictive modeling and business intelligence. Keke holds an MBA from the Tuck School of Business at Dartmouth College.

B2C marketing is often perceived as being creative and cool, while its counterpart B2B marketing is being labeled as boring and dull. In the B2C world, companies market to consumers with campaigns full of personality, emotion and opinion. The best examples of those campaigns are designed around the vast amount of available information. When companies leverage that information in creative, personalized, well-timed ways, consumers are far more receptive to the marketing messages – and far more likely to convert. In fact, a study from MyBuys found that 48 percent of consumers buy more from retailers who personalize the shopping experience across channels.

In the B2B world, marketers too often communicate to faceless entities. Although some large B2B companies have long been leveraging predictive analytics in their marketing and sales efforts just like the B2C companies, they face a unique set of challenges. There are two major hurdles on the journey from building good-enough predictive models to razor-sharp models:

The missing pieces of the puzzle: In the old days, B2B companies had a limited number of attributes to put into predictive models. These included firmographics, past purchase behavior, marketing interactions, customer service interactions and product usage. With those data elements, companies tried to identify the best targets for acquisition, retention, cross-sell, upsell and share-of-wallet growth. When companies first implement these attributes into predictive models, they usually see a big initial uplift. After all, predictive analytics done right can yield substantial payoff over random targeting. Encouraged by the early results, modelers often regularly refresh or rebuild the models seeking to make the models better and better. But then diminishing return hits. There just isn’t enough information about the targeted companies. It is like trying to put a big puzzle together, but there are many missing pieces.

Timing: In the B2B world, company attributes don’t change frequently. That stagnation, coupled with the limited amount of information, makes it difficult for marketers to predict the timing of an event. For some companies in the B2B world, “conversion” is rare. While B2B marketers know the ideal profile of their customers, that alone isn’t actionable. Such profiles don’t allow for precision targeting.

These challenges don’t have to hold back B2B companies anymore. The big data era has opened the door to a new world of possibilities. There are thousands of new data sources. We are now able to know so much more about businesses. Those faceless entities have taken on personalities and are constantly sending digital signals to potential suitors. It is time for B2B marketers and data scientists to catch up to their B2C counterparts. The B2B world can be just as creative as the B2C world. After all, B2B and B2C marketing are not that different, and all marketing efforts are person to person. B2C companies have been able to tap into the following types of data for behavioral targeting and real-time engagement, and now B2B companies are getting on board with these, too:

Demographic/firmographics: In the B2C world, these are age, gender, race, and similar attributes, while the equivalent in the B2B world is industry, size, and company ownership type.

Psychographic and behavioral: Related to consumers, psychographic and behavioral attributes include personality, values, opinions, attitudes, interests and lifestyles. Related to businesses, these are characteristics and business dispositions, such as whether a company is a technology adopter, is risk resilient, is women- or minority-friendly, or promotes work-life balance.

Real-time activities: These are the time-sensitive trigger events. A family just bought a new house. A young woman tweeted that she was just engaged. A retailer just announced a new store plan. A company hired a new CMO who was known for implementing marketing automation at her last company. Employees at a company have downloaded a white paper on compliance management.

If a retailer is able to identify an ideal customer based on the demographics and the behavioral footprints left in the digital world and send a real-time personalized offer when the person is in the proximity of its store, why shouldn’t a B2B company do the same? Imagine a company trying to promote its next-generation IT security solutions. The company identifies its ideal customers as mid-sized retail and health care companies that are technology adopters. If the company also knows which companies had recent IT executive changes, which companies just announced big data initiatives and which companies are surging on IT security intent on the Web, then it can combine this information and launch targeted marketing campaigns.

The time has come for the marketers and data scientists in the B2B world to unleash their creativity. Finally, it is possible to build not just good-enough models but awesome, mind-blowing models. Powered by the new-world predictive and prescriptive data, there will be more and more of those just-in-time marketing offers and sales calls so targeted and relevant that decision-makers in targeted organizations would want to click through and entertain.

 

Sign up for the free insideBIGDATA newsletter.

Leave a Comment

*

Resource Links: