Most companies have reams of data they could be using to impact future customer messages, offers and calls to action. But most of those companies either don’t use that data, or don’t realize it exists in the first place.
That’s part of the problem Lattice Engines is trying to solve. Brian Kardon runs marketing at Lattice, and was also the CMO at Eloqua in it’s early hockey-stick-growth days.
I sat down earlier this week to talk with Brian about what predictive analytics really means for B2B marketers, and how more companies can start to leverage its full potential.
I’m not sure a lot of people really understand what we mean by “predictive analytics.” Can you give a quick layman’s definition of what it is, and why it’s important?
The best example is Amazon. Its recommendations are predictive analytics in action. Its recommendations are based on a number of simple elements: what a customer has bought, items they’ve rated and liked, where they live, and what other customers have viewed and purchased. Amazon uses all that data in a model that predicts what you will likely buy next.
For Amazon, predictive analytics is a big deal. More than 30 percent of its business comes through its recommendation engine. It is analyzing current and historical data to make predictions about the future … with stunning results.
For businesses, predictive analytics can be used to analyze patterns in historical, transactional, social and other data to identify which prospects are most likely to buy and when they are likely to buy.
Most B2B marketing organizations aren’t executing on the predictive analytics opportunity, but the potential is enormous. How can you help make a business case for more companies to focus here in 2014?
First, more B2B companies are embracing predictive analytics than you might think. The marketing teams at companies as diverse as Adobe, GE, Mindjet, Bank of America and Time Warner are using predictive analytics to find, prioritize and close their next customer.
Marketers have been great at transformation, from analog to digital, from embracing social, content marketing, organic search, and, of course, marketing automation. We are seeing that the use of predictive analytics is the next thing that will transform marketing performance.
Predictive analytics lets you stop guessing. You can predict which lead characteristics indicate likely buying signals. For example: job postings, news stories, credit ratings, financial history, purchase transactions, litigation, patents, contracts, locations, growth, executive management changes and beyond could be indicative of a purchase decision.
A predictive model can scour a host of web, social and internal sources (think CRM and marketing automation) to evaluate thousands of attributes and provide a score that is predictive of buying intent. With predictive scoring, marketers can confidently push high quality leads to sales and keep earlier-stage leads in marketing automation for further nurturing.
The business case for predictive analytics is clear: more pipeline, higher conversion rates and faster revenue growth. Our customers are seeing 75 percent more pipeline and 2x win rates in a matter of weeks.
Can you share a success story of a client with before/after data on how predictive analytics impacted their business?
Dell is a Lattice customer. The Wall Street Journal recently profiled Dell’s success with predictive analytics. The story said: “The marketing department was able to cut the number of leads it sent to the sales organization by 50 percent – to only the most promising prospects – and revenue results went up by almost double.”
One of the myths I hear often about predictive analytics is that you need to start capturing a ton of new data. Can you talk about how predictive analytics can have an impact immediately using the data companies may already have – via Web activity, Marketo, Eloqua, etc.?
The data coming out of CRM, Marketo and Eloqua only represent less than one percent of what is knowable about a prospect. Often, the “trigger” is something that you are not capturing – but is discoverable.
For example, one of our customers sells foreign exchange software. The key trigger for them is whether or not the prospect has recently opened an office outside the US. For a customer selling CAD-CAM software, the trigger is the number of design engineers the prospect is hiring and the number of workstations in use. These elements are discoverable and highly predictive.
What’s the best “step 1” for a company that wants to start “crawling before they walk” with predictive analytics?
A good starting point if for marketers to find their triggers! What are the attributes, if knowable, that would be predictive of a new customer? For a lab equipment manufacturer it might be newly funded labs. For a company selling switches and routers it might be a new real estate lease, indicating that the prospect is moving to a new office. These kinds of triggers can be invaluable in starting the process of applying data science to marketing.
