By now, sales and marketing execs should be somewhat familiar with the term predictive lead scoring (PLS). Put simply, PLS shops take your existing CRM or marketing automation data, enrich it with a bunch of external data signals to form your ideal customer profile, then use this profile to rank your existing leads from best to worst.
Given that SiriusDecisions recently found that 94% of marketing qualified leads don’t convert, the idea of finding the magic 6% through data science has become quite tantalizing.
Predictive lead scoring today
Today, predictive still takes a back seat to traditional, two-dimensional lead scoring, but the landscape is quickly changing due to concerns over scoring inaccuracies and faulty lead qualification criteria.
The traditional lead scoring model ranks leads based on two main criteria: fit (demographic data like title, company, revenue etc.) and engagement (behavioral data like what they downloaded, pages they viewed etc.).
Unfortunately, this approach leaves much to be desired. The model scoring criteria and weights are assigned based on what everyone thinks is qualified, which often leads to uncomfortable powwows between marketing and sales over lead quality and prioritization.
The beauty of predictive modeling is that it takes the human element out of scoring by relying on data science and machine learning. Similar to how your credit score is calculated, predictive models combine historical win/loss data with thousands of public and private data points to calculate how likely each lead is to convert.
Today, PSL is still in the early adoption phase, but as the graph below suggests, B2B companies are catching on quickly with a 14x adoption increase since 2011.
With 20+ innovative vendors and dozens of new buying signals emerging every day, the future of PSL appears very bright. Here are a few areas in which predictive models are improving to give marketers the most accurate look into their pipeline.
Technology provider data gets dynamic
Given that 78% of PSL early adopters are technology companies, a crucial ingredient in any model is your ideal customer’s web technology stack. For example, if you sell to enterprise publishers, a company that uses WordPress, Google Analytics, and Facebook Comments is not as likely to convert as one that uses Adobe CQ, IBM Coremetrics, and Disqus, even if both companies are similar in size and revenue. This is great, but let’s take it a step further...
Imagine a predictive solution that dynamically bubbles up leads that have recently added or dropped a certain technology provider. If your product has an awesome integration with Marketo or Eloqua, being able to surface the leads that have added one of these providers the day it happens is sales gold. Conversely, being able to surface the leads that have dropped a competing provider empowers your sales reps to reach out when the timing is right.
Predictive scoring to generate net-new leads
Lately, it’s become a common practice for CEOs to give their marketing VPs a seemingly unattainable lead quota to hit each quarter, making lead generation the primary focus for most B2B marketing teams. To hit their lead quota, more marketers are relying on predictive solutions that not only score their existing leads, but source new prospects sales has yet to find.
The process is fairly simple—once an ideal customer profile has been built, PSL providers with a proprietary company and contact database can match this profile against millions of potential prospects to find and score the leads that are not currently in a company’s CRM or marketing database. This ensures that sales reps always have new leads to pursue that are a near perfect match to your ideal customer.
Question: What are your biggest lead scoring challenges? Where do you think predictive lead scoring is headed? Feel free to leave your thoughts in the comments below!