June 14, 2013
By Ran Gishri,

Transforming Lead Scoring with Big Data

In the eyes of many, the marketing department is a service provider. Its primary goal is to generate demand and enable the sales team to meet their quota. As such, many organizations are putting service level agreements (SLAs) in place to ensure marketing delivers the required amount of leads. A common challenge is agreeing on the desired output—defining which leads are "sales-ready" and which marketing needs to nurture or qualify out, so salespeople don't waste time trying to sell to those unlikely to buy.

This is where lead scoring comes into play. Simply put, lead scoring is the process of ranking leads based on their sales-readiness.

The majority of lead scoring techniques are based on several data types:

Demographic—job title, role, geographic location

Firmographic—industry, revenue

Behavioral—Web site visits, downloads, searches

While lead scoring has become an important component in the lead management process, companies implementing a lead scoring program often run into issues that hamper accuracy:

People lie. According to MarketingSherpa, just 71 percent of site visitors provide accurate answers to questions on Web forms. Forty-five percent do not provide their real company name, and 47 percent lie about their job title.

We're living in a dynamic world. People are constantly on the move, changing roles and jobs.

The dark side of the moon. Lead scoring only relies on data points that are accessible through Web site analytic tools and company databases. What people do elsewhere (e.g., on social networks) goes under the radar.

The good news is that there is a solution in sight.

Every day, we create 2.5 (or 10³³) quintillion bytes of data. A large portion of this data comes from the "social Web"—social networks, content sharing sites, blogs, and Twitter. We call this big data.

The vast majority of the professional population leaves "digital footprints" online. People create social network profiles, blog, connect, follow, share statuses, and appear in other people's content. Following a lead's digital footprints can reveal a person's true role, topics she's interested in, business relations, tools she uses, and more. And there's a bonus—people care about their online presence, which means their online data is likely to be fresh and accurate.

But wait, there's a caveat. The data is unstructured.

Big data does not typically reside in database tables, which means you can't just import and push it into your marketing automation or CRM database. While it's easy for a marketer or a salesperson to read a press release and draw conclusions about a lead whose name is mentioned in it, doing so automatically, and in large scale, is anything but simple.

This is where a new breed of big data analytic tools comes into play. Advanced technologies use natural language processing, Web mining, and machine learning to structure the unstructured (i.e., turn the mountains of information in the social Web into data that is clean and actionable). With these new tools, companies can take lead scoring to the next level by fusing demographic, firmographic, and behavioral scores with a new type of score—the social score.

Social score is best used against an ideal buyer profile that takes into account all possible digital footprints. For example, a company might find that the ideal buyer for one of its products is a person who takes an interest in cloud computing, attends a Gartner or Forrester IT conference, follows certain industry analysts, and has knowledge about VMware and Amazon cloud. Matching new leads with this ideal buyer profile will provide a strong indicator about the buyer's relevancy, regardless of the information he or she provides.

Combining the behavioral score, which indicates a person's level of awareness and interest with the social score, which indicates a person's true relevancy to the business, produces an extremely accurate prediction of one's likelihood of making a purchase.

Analysis of big (social) data helps marketers beyond just lead scoring. As noted above, data in the social Web tends to be fresher and more accurate than the mostly static information that resides in marketing automation and CRM databases. Using this data, database marketing teams are now able to verify demographic and firmographic data they have on record. For example, one can verify that all leads still hold the same position, or that a company uses a certain product.

And adding a social data layer helps segment leads in ways never before possible. Marketers can now segment their lead database based on people's areas of interest, roles, social graphs, and more. Finally, social lead enhancement also plays a part in social selling (leveraging social networks and the associated tools in the overall sales function).

Eighty percent of the information generated daily is unstructured. Making sense of this data enables marketers to take a gigantic step forward, not only in how they score leads, but also in understanding their customers, making information-based decisions, and making predictions. Inch by inch, marketing is shifting from art to science with the help of big data.

Ran Gishri is the vice president of marketing at Leadspace. He has almost 20 years of software marketing and product management experience in start-ups and large corporations. Prior to joining Leadspace, he served as vice president of marketing at Zend Technologies and headed marketing and product management at BMC Software, Identify Software (acquired by BMC Software), Aladdin (acquired by SafeNet), Crystal, and Mercury (acquired by HP).