Lead scoring is a hot topic these days, and there are some awfully sharp minds trying to figure it out. Vtrenz has a white paper on the subject that has good information. We’ve got our own work on multidimensional lead scoring. The Pedowitz Group even tried to boil lead scoring down into “7 Easy Steps” (there’s some good information there, but I don’t entirely agree with the proposed approach). And, recently, SiriusDecisions put out a research brief on “When Good Lead Scoring Models Go Bad.”

The SiriusDecisions brief identifies three “attribute families” for lead scoring models: demographic data (both for the individual and for the company), BANT (budget, authority, need, and timeline), and activity. The brief points out how easy it is to focus on what you have historically used and what you have the easiest access to when developing a lead score…and points out the flaws in this approach.

The brief goes on to identify five common traps that they see companies fall into when developing their models. What is striking is that the first four traps are all, basically, about taking shortcuts and oversimplifying the lead score. They’re all valid — over-emphasizing BANT when it’s not appropriate to do so, under-emphasizing activity and demographics data, not putting enough thought into the different possible values for each variable used in the score, and weighting the variables equally. After reading through those, I audibly chuckled when I hit the fifth trap: “Overcomplicating the model.” The whole gist of that section is to keep the lead scoring model simple!

This is where SiriusDecisions falls short. On the one hand, they are wayyyyyy better than many analysts when it comes to laying out practical, pragmatic approaches to address business challenges. On the other hand, they’re still analysts, and they sometimes get caught up in words that sound good…but that are not necessarily grounded in reality. I wrote on one of my other blogs about the complexity involved in a simple, two-variable lead score situation. And, in that case, I was barely touching on the common mistakes SiriusDecisions outlines in their brief.

Which gets me to two pretty serious gaps in SiriusDecisions’s brief.

Gap No. 1 — Focusing on “What” Instead of “Why”

Their three attribute families — demographics, BANT, and activity — are families of what is measured, as opposed to why it is measured. Multidimensional lead scoring focusses more on decomposing the key attributes of your leads into “why” dimensions: their profile (who they are), their engagement (how mentally interested they are in your company already), and their position in the buying cycle. Each of these dimensions may include variables from multiple of the attribute families identified by SiriusDecisions.

This isn’t to say that SiriusDecisions doesn’t make excellent points. They absolutely do. But, I would be leery of using their implied structure as a lead scoring approach.

Gap No. 2 — Making Complexity Simple

One of the key points that I have to make whenever I speak or write about multidimensional lead scoring is that, while the name sounds complex, the fundamental approach is geared towards making lead scoring simpler. People are complex animals. Your leads are people. Lead scoring is inherently complex, especially in a B2B environment. I have yet to find a company where there are 2-3 magic variables that are both practical to expect your leads to answer truthfully (or that you can get without asking the lead a question) and that can genuinely assess the quality of the lead.

However, by deconstructing your leads into multiple dimensions, you can find a subset of attributes that are a good measure of their quality for each dimension. When it comes to measuring engagement, there are a number of measurable behaviors (there is a bias towards “activity” for the engagement dimension) that are indications of engagement. Any mix/combination of these behaviors can indicate a lead is sufficiently engaged to be sales ready on that dimension. In the case of the lead’s profile, there are typically some “must-have” attributes, be they BANT, demographic, or some combination thereof. Leads who don’t meet these minimum requirements should not be considered sales ready.

To me, if you are determined to stick with a single lead score — rather than multiple lead scores, and a requirement that a lead exceed thresholds for each one before they are passed to Sales — you will always wind up with a model that looks like it’s overcomplicated.