Predictive Lead Scoring — The Future is Here for B2B Marketers


Share on LinkedIn

“The future is here. It’s just not evenly distributed yet.”
–William Gibson.

The market for Predictive Lead Scoring (PLS) solutions is starting to take shape. This according to SiriusDecisions, a B2B research and advisory firm focused on optimizing sales and marketing.

“Optimizing” is exactly the point of predictive analytics. As I pointed out in a previous post, “complexity is the Achilles’ heel” of the Marketing Automation (MA) industry. B2B marketers have to think like programmers to figure out lead scores, campaign logic and more.

Lead scoring is largely a manual process, based on human judgement. Nothing wrong with that — certainly better than just sending all leads to the rep — but hardly optimized. Ideally marketers should be scoring based on how certain lead attributes — demographics, firmographics, social signals and more — actually related to business outcomes like qualified leads or closed sales.

But help is at hand, says Kerry Cunningham, Research Director at SiriusDecisions. Here are a few highlights from a recent study, published in a Field Guide.

First, marketers aren’t all that happy with the current MA solutions. While about two-thirds of marketers say they have implemented MA-based leading scoring, only 40% report that their sales team agree or strongly agree that the scores add value. A full 90% thought that PLS has more value than traditional scoring.

This is critical because if the sales organization doesn’t trust the scores, they won’t act on the leads in a, um, predictable way. Every experienced rep has a mental map of the “good” leads — those likely to be worth pursuing. Just like marketers, it’s based on their experience and judgement. So marketers need to not only score the leads well, but also convince the sales reps to give them an honest try.

Second, the vendor community is rapidly taking shape. Although they use a variety of different techniques, Cunningham says he was pleasantly surprised at their progress in “productizing” predictive analytics, with solutions that can be implemented in a couple of weeks. Anyone who has tried to create predictive models and then operationalize them knows this is a huge accomplishment.

Here are a few snippets from the Field Guide to give a sense of what the vendors do:

  • 6Sense is a strong fit for high-tech, manufacturing and other b-to-b enterprises focused on identifying new leads in early buying cycle stages.
  • Fliptop targets SMBs and larger organizations that use with a predictive solution that analyzes contacts, accounts and opportunities.
  • Infer emphasizes scoring existing contacts from an SFA or MAP platform vs. discovering new leads.
  • Lattice offers its customers a comprehensive and broad set of demographic, firmagraphic and behavioral attributes.
  • Leadspace takes an original approach to predictive lead scoring with a solution that matches products to specific buyer personas.
  • Mintigo provides data remediation services during customer onboarding, … for organizations with known or suspected data hygiene problems.
  • Salesfusion is a MAP that offers predictive lead scoring as a feature.

The technology is cool, and for sure you’ll want to select tools that fit your situation well.

But the real key to success, according to Cunningham, is to create a Service Level Agreement (SLA) between marketing and sales. Marketing should agree to provide that volume and quantity of leads needed, and sales should give leads the attention they deserve (based on the scores) and provide feedback. Shared measurements can also help to keep marketing and sales aligned, acting like one team and not two.

PLS solves a big problem in B2B marketing and sales. With buyers engaging more online to self-educate, more leads will be created by marketers serving up content. It’s critical that leads passed to sales really are ready for the human touch.

How to get the Field Guide? Good question. SiriusDecisions doesn’t sell reports separately, but if you want more details you can try this link.

Further reading: Predictive Analytics is the Future of B2B Digital Marketing


  1. Bob – if the sales organization doesn’t trust the scores, they won’t act on the leads in a predictable way. True! , , , kind of. At least, the sales force won’t act in a favorable way. But there’s another wrinkle that I’m sure gives everyone on the marketing team grey hair. The way sales responds to leads depends highly on situation and context. If a lead is ranked a ‘6’ on a 10-point scale, with ten being the best, a given sales rep might be highly responsive if it’s in the middle of the quarter, and his or her back is not up against the quota wall. But at the end of the quarter, when pipeline opportunities must be closed, that same six might not get much attention at all.

    I’m not sure whether this conundrum can be fixed – or if it needs to be, but it should be recognized in light of predicting how a sales force will respond to lead qualification scores.

  2. Good point. However, time of quarter is another variable that could be part of a predictive model.

    I don’t think that “favorable” is the right word for how reps handle marketing-generated leads. Perhaps “consistent” is better.

    My point is that if marketing is going to go to the expense of generating leads, they should be passed along to reps with the best information possible. The score is one key indicator of whether a lead is “worthy.”

    If reps consistently give leads the treatment they deserve (based on the score and any other factors) then the business outcomes will eventually help to tune the model, which will generate leads with better scores, … and so on.

  3. I agree that there’s diminished value for marketing-generated leads if the information passed to sales is vague or fuzzy. The problem with encouraging/motivating/demanding that sales personnel handle leads using consistent methods and processes rep-by-rep boils down to a major issue: compensation at risk. The moment a company implements a pay plan that shares risk between the company and the salesperson, that company must accept that the employee will discriminate about how he or she expends personal resources (notably, time). The same occurs a little tiny bit with salaried employees, but it’s de riguer when comp is on the line.

    Take two extremes:

    1) No commission – pay plan is salary only. Company can specify which leads get handled, and how they get handled. Salesperson complies because there is no risk for lost compensation. Company has full leverage over what to ask, and what to expect.
    2) Full commission. Salesperson has near-complete discretion over which opportunities he or she accepts, and how much time to devote.

    Most companies have established pay pay plans that reside somewhere in the middle of these extremes, and many skew toward a smaller ratio of compensation “at risk” for the employee. Therefore, a company can reasonably exercise some leverage, but with compensation in play, it’s not realistic to expect salespeople to hold a consistent view of the WIFM (What’s In It For Me?) for marketing’s numerical tabulations.

    The reason is perceived risk and context. Take a specific hypothetical lead with a score of ’60’ on a 100-point scale. It’s early Q4, and Rep A has had a successful year and has already over-achieved his annual quota. Rep B is in a much more tenuous situation, and not only that, he has financial problems at home. Assume that both reps have 33% of their total compensation at risk. No matter how sophisticated the lead scoring, or how insistent the company’s management is about consistent handling for marketing-generated leads, it’s highly unlikely these reps will treat that ’60’ lead the same way.

    There are two possible remedies to the consistency problem. 1) trust in the scoring algorithm must be sufficiently high that salespeople recognize that opportunities given a score of X threshold must be worthwhile time investments, regardless of their personal income at risk or personal consequence of failure. Or, 2) modify the compensation plan so that there is little to no income at risk, which allows the company to prescribe and enforce how each opportunity is managed.

    In my experience, the first approach is exceptionally rare. Only by increasing the ratio of fixed to variable compensation can a company exert more leverage over how the sales force must qualify and manage leads.

  4. Andy, even in cases where 100% of compensation is at risk (I’ve had such a job), management has influence. Reps have discretion, but they also work for a company that provides facilities, tools, marketing, and yes, leads.

    Regardless of the “system” in place, if leads are being generated and worked, with some becoming qualified, deals, etc., that creates data which can be used to generate a predictive model which should improve the quality of leads scored and passed to sales.

    Reps that have ignored the “score” and continue to do so won’t see any improvement (assuming leads continue to be passed along) but they won’t be hurt either. Eventually the model could be tuned to realize it’s a waste of time to score leads for rep X.

    But I suspect a more common scenario is that reps will look at the info provided with the lead by marketing, make his/her own personal assessment and then decide whether to invest the time. No change in compensation plan required. If the scored leads are better quality, the result should improve. Again, this assumes the rep gives at least some weight to the score.

    I don’t agree that companies need to overhaul their compensation systems to make lead scoring, predictive or otherwise, to work. If marketing and sales have a cooperative relationship (a big “if” I realize), I should think that explaining how the scores have been improved will motivate at least some reps to give these leads a try.

    In the end, if PLS improves results, reps will figure it out and adopt, because it’s in their own best interest. Whether they are are on 100% commission or 100% salary.

  5. Bob –
    Thanks for the mention. We are quite familiar with the marketing/sales misalignment and go a long way to help our customers solve this issue.

    First, since 6Sense ties activity data from the B2B web (blogs, forums, research sites) to company attribute and our customers’ CRM and MAP data, we get a FULL picture of the buying intent of each account and contact. The net result is that our score shows what the buying stage is of that account and how active their “buying committee” is. Both marketing and sales can see the same scores and the same insights. Second, 6Sense exposes the “why” behind intent score. For example, what did that company do (or the members of their buying committee do) to indicate a propensity to buy? What whitepapers did they download? Which sites did they visit? Both sales and marketing can see those insights. Finally, we show all the descriptive indicators (company size, revenues etc.) so there’s NO question about the validity of that prospect. All this information is available to both teams, either on our platform, or is pushed into CRM or MAP and becomes the basis for a very clear and actionable SLA.

    Alison Murdock​


Please use comments to add value to the discussion. Maximum one link to an educational blog post or article. We will NOT PUBLISH brief comments like "good post," comments that mainly promote links, or comments with links to companies, products, or services.

Please enter your comment!
Please enter your name here