The Promise Of Predictive Lead Scoring: 7 Experts Sound Off

The promise of predictive lead scoring for B2B marketers is huge: marrying customer behavioral data with additional predictive insight to zero in on only the very best leads. Using data modeling rather than guesswork to drive marketing programs is incredibly appealing to marketers with limited resources.

Marketers today need a comprehensive view of customers based on interests and buying intent, and predictive lead scoring provides that precision and more. That technology is now readily available to scores of marketers without the need for data science experts.

The challenge is that predictive lead scoring is difficult to implement, requiring alignment throughout the organization and buy-in from the C-suite. Demand Gen Report assembled a “virtual” roundtable of some of the best minds in the predictive business to address the challenges and opportunities inherent in executing predictive lead scoring.

Here they discuss how predictive analytics can help forge alignment between marketing and sales teams, the difference between traditional and predictive lead scoring, and why predictive lead scoring remains an aspiration rather than a reality in B2B marketing today.

Demand Gen Report: What are the major trends you’ve witnessed in predictive lead scoring this year?

Leslie Fine: 2015 has been the year of predictive analytics. While predictive modeling has been around for quite some time, businesses have adopted predictive analytics into their marketing strategies this year in ways that were previously unprecedented.

In the past, utilizing predictive scoring involved hiring a team of data scientists to build a modeling system, analyze the data, and report on insights gleaned from it. Advances in technology have enabled businesses to implement predictive scoring without the need for that team of data scientists.

Justin Gray: The biggest trend … is predictive scoring as a function of demand generation. Ultimately, marketers know what their ideal leads look like and therefore they can use non-traditional data sources to go find more of those “A” leads. Like most technologies in marketing, new tools always are mistaken for lead generation at first, and in their second maturity cycle they are adopted as a more attributive strategy. Predictive lead scoring (PLS) is no different. PLS and the entire notion of automation-qualified leads will ultimately help marketers uncover their best customers, but they will need to use core tenets like marketing automation and account-based marketing to engage those buyers in a hyper-personalized way.

Ruth Stevens: [We’ve seen] an increased understanding among B2B marketers of the practical value of data-driven tactics and an appreciation of the power of analytics to identify higher-value prospects.

Liz Osborn: Most predictive lead scoring vendors have a “black box” approach to predictive scoring. Many vendors we’ve looked at for this space are customizing their software, approach and algorithms, and it’s very difficult, if not impossible, to understand what goes into their scoring model. A major trend I’ve witnessed is these vendors and their products are maturing. Many vendors are able to lay out their approaches more openly,and share the elements of their scoring models: what they’re based on, and how they end up at a final score.

Jerry Rackley: We’re really on the front-end of awareness of predictive analytics. For example, predictive lead scoring has great application in the area of account-based marketing (ABM). We just completed a study that shows just 18% of ABM programs are in the “mature” category, and just about 40% of this mature segment is using predictive to help them select accounts to target. It’s a small percentage of the total, and I suspect this ABM adoption of predictive mirrors broader adoption. But, the key is that predictive in the case of ABM — and other marketing strategies as well – is a best practice, associated with the most mature use cases. It’s aspirational right now, but predictive is the target at which more marketing organizations are aiming.

Matt Heinz: It’s not really about big data, but rather about fast data and the right data.The best marketing technologists inside organizations today are cutting through the clutter and identifying the trends, data and insights that most directly and precisely lead to buyer momentum and velocity. The data needs to lead to action and results, period.

DGR: What are the upcoming trends marketers should be aware of regarding predictive lead scoring?

Rackley: I think the trend to pay attention to … in the new year has little to do with technology, but everything to do with culture. There are three broad categories of culture when it comes to data-driven marketing, characterized by organizations who “fly blind,”those who use “pseudo-analytics,” and those who are executing the “real deal” — and this last category is the smallest. The reason an organization lands in one of these three categories is entirely cultural. The organizations that are the “real deal” will have no trouble embracing predictive lead scoring. They’ve already embraced the underlying concepts,they have or will get the enabling technology, and don’t have to be sold on the value. It is the other two organizations that will struggle. As organizations look into the new year and plan on how to achieve objectives enabled by predictive lead scoring, they need to assess their cultural readiness to do so.

Gray: The ABM craze is upon us, with the same fury as PLS. Both tools, along with marketing automation, are building the pillars of the tech stack. There’s going to be a good degree of consolidation from a technology perspective, and it speaks to the need to create a highly process-oriented buyer engagement strategy with all of these critical tools optimized. The No. 1 area I would recommend executives invest in is education for the marketing department. No other area will yield as high a result. We are putting some very complex technology in front of today’s marketing departments, and we are asking them to learn on the job. It’s costing hundreds of millions of dollars, not only in time and materials, but in opportunity costs.

DGR: The differences between traditional and predictive lead scoring are significant. Are there cases in some companies where traditional lead scoring is enough?

Rackley: Not all organizations need to rush to implement predictive lead scoring. Those marketers whose solutions are highly specialized, that are aimed at very narrow markets,and where a heavy reliance on relationship marketing is necessary, may not see benefits.But these organizations are the exception. The traditional B2B lead generation function that is focused on producing a constant stream of leads of increasingly high quality will benefit the most. Predictive lead scoring helps to better align the sales and marketing efforts, and it makes the overall process more efficient.

Heinz: No. Traditional lead scoring is good and is still more than many companies are already doing today. Predictive lead scoring means translating more data into better insights, impact and results.

Nick Ezzo: If your prospect base is individuals or very small businesses, there may not be enough data to build an accurate scoring model. Predictive lead scoring works when you sell to certain categories of companies that have similar traits. If you don’t have that,behavioral scoring is really your only option.

Stevens: Many B2B companies are targeting a limited number of prospective accounts where they already know who they need to talk to. For example, I had a student in class recently from a Unisys division that was selling to five target accounts globally. For that kind of situation, it’s really a matter of message content, media and cadence. Predictive lead scoring is most valuable when casting a larger prospect net.

Gray: Let’s face it: if you have a great product that your buyers love, you’ve got a huge leg up on most organizations. Predictive lead scoring (PLS) lets you zero in on what we call the ICP, or Ideal Customer Profile. Some organizations have that dialed in. Look at Slack. Slack never marketed to me, never reached out in any way. But our organization fit their ICP to a T, we heard about them through their presence in companies like us, and we bought the solution even though they have [competitors] that are free. If that’s where your organization is, PLS may not be your highest priority. But if that’s where your organization is, you probably are looking for scalable solutions to create even more efficiency through that hyper growth and will likely implement PLS. Working smarter is never a bad idea.

DGR: Sometimes lead scoring efforts fail internally for marketers. What is your advice to companies interested in re-evaluating their approach?

Gray: Lead Scoring fails [for] three reasons. 1. It’s a complete guess and therefore has no significance. 2. It was rolled out in a poor manner where it wasn’t adopted due to lack of education. 3. Both of the above. If your efforts have failed, it’s time to bring validity to the table and respect lead scoring for what it is, a science that requires a huge amount of organizational change.

Osborn: The most important factor is getting sales on board with the scoring model and process early, getting buy-in from all levels — from the top on down — and educate, educate, educate! The sales team needs to be educated on how predictive lead scoring works, what goes into it and why it’s been scored that way. If the team has a set idea on what leads have been proven to work in the past and what they know will close, they will continue to stick to those sources, ignoring the predictive model. Credibility of the model and education of sales is key.

Ezzo: If you’re going to use predictive lead scoring as the scoreboard to judge good and bad leads, you have to implement it with both sales and marketing working together to understand and tweak the model. You also need to get a handshake from both parties that the model will be the judge.

DGR: What elements are crucial to a well-considered predictive lead scoring strategy?

Stevens: The most important matter is identifying the right qualification criteria. If the data is available and of good quality, automated tools can generate a basic field of predictive variables and weightings. But you can’t just set it and forget it. Marketers must also take the time to validate the results, using their judgment and experience. They must collaborate with their sales counterparts to continuously refine the process to ensure that the leads are meeting their needs.

Predictive lead scoring is only beneficial if the organization is able to put insights into action. For example, if the marketing and sales teams within an organization are not aligned — if they do not have shared objectives, or if there is no buy-in on the qualities that differentiate a high quality lead from a low quality lead — it will be difficult to reap the full benefits that predictive lead scoring provides.

Gray: Data, data and data. You have to be using CRM and marketing automation, and you have to be using them religiously. You also need to have a pretty consistent definition of a great customer. We may say that all “won” deals are positive signals we want to base our model around, but the real magic happens when you can also bring into the picture the health of a customer. Can your model predict a customer who is less likely to churn? That only comes from reliably capturing that data in your systems of record. It’s frankly breathing new life back into CRM and data stewardship.

DGR: Is there anything marketers are getting wrong in relation to predictive lead scoring?

Ezzo: I think it’s a mistake to replace behavior scoring with demographic/firmographic scoring. It works best when you have a tic-tac-toe matrix with behavior on one axis and demographic/firmographic on the other axis. That way, you can zero in on the leads that are high in both categories — those that show both buying signals and fit.

You should also be aware that predictive scoring creates a feedback loop of similar customers and prospects, and will eventually result in a self-fulfilling prophecy. By that, I mean that you will only sell to prospects that look like your customers. If you want to break into new markets, you’ll need to take that into account. Case in point, one of our largest customers came in as a “D Lead.”

Stevens: Relying 100% on online factors. The Internet provides extraordinarily valuable information about prospect characteristics and behavior. But it’s not everything. It’s important to integrate the online data with offline behaviors for a more robust view.

Gray: Marketing sometimes gets distracted by shiny objects. The demand generation side of PLS is the shiny object, and it’s easy to get overwhelmed by it. The fact is you should be using modeling to not only predict your best customers, but also to assemble the right marketing and sales plays around those buyers, as well as engaging the buying committee. If all we get out of this complex data science is the same as a well-targeted lead buy, we’re doing something wrong. The data will unlock a ton of great information on what creates your best buyer and what interests them — don’t get distracted by the first piece of low-hanging fruit.