November 18, 2014 | By Kerry Cunningham
In the mid-1980s, Ireland’s Industrial Development Authority famously used print and billboard advertising to proclaim “Missing the Industrial Revolution Was the Best Thing That Ever Happened to Ireland.”
The intent of the campaign was to attract investment by positioning Ireland as a place not mired in the grimy past of industrial Northern Europe, but instead poised – with its well-educated, under-employed populace – to leapfrog its industrial neighbors into the information age. Indeed, the ploy worked, as global businesses in need of bright, educated information workers began investing in the development of Irish operations.
So, what’s that got to do with lead scoring?
Well, missing the initial “lead scoring revolution” might be the best thing to happen to organizations that have yet to adopt traditional (i.e., non-predictive) lead scoring based on a marketing automation platform (MAP).
Traditional lead scoring is based on assumptions about the meaning and value of prospect behaviors and attributes, and is limited to data collected in marketing and sales force automation systems. Predictive lead scoring uses statistical procedures to determine the value of attributes and behaviors and, in most cases, augments data in MAPs and sales force automation systems with data collected from third-party providers and Web crawling/scraping (big data).
Consider this: At Sirius, we regularly talk with marketers about how to improve lead conversion through the demand waterfall. If these marketers are in tech companies, they probably have marketing automation in place (about 70 percent of these organizations do, according to our research). But the adoption of marketing automation drops rapidly outside of tech.
And only about half of the companies that have adopted a MAP also do some sort of MAP-based lead scoring today. Taken together, that means that most marketers are not yet scoring leads, predictively or otherwise.
That is unfortunate. Furthermore, when I describe the benefits of using statistical methods for scoring leads, I am frequently met with the objection that because the organization has not yet done MAP-based lead scoring, they are not in the position to take on the (seemingly) more complicated job of predictive lead scoring.
Don’t believe it! Missing traditional lead scoring may be the best thing that ever happened to your organization. That isn’t to say that some organizations don’t do traditional lead scoring very effectively – some do. But it is not a prerequisite for doing predictive lead scoring.
In fact, because most organizations stumble and fall numerous times with MAP-based lead scoring before hitting on something sales likes (our research suggests that most salespeople – approximately 60 percent – do not think MAP-based lead scoring is valuable), starting from scratch with predictive lead scoring may make your path to lead scoring success less fraught with sales skepticism than if you do MAP-based scoring first.
The things you need to do to be successful with predictive are the same as you would do with MAP-based lead scoring:
- Don’t assume alignment with sales. MAP-based scoring is based on assumptions. One of the major killers of any type of lead scoring is the assumption that sales will like the leads produced by scoring, just because they should be better. If you want to get buy in from the sales organization, you need to involve them. Do you know what tele and sales think makes a good lead? Are they involved in the development of the lead scoring schema?
- Implement service-level agreements (SLAs). No matter what kind of lead scoring you do, if the recipients of sales leads are left to make their own decisions about whether to follow up and how much effort to apply, then their decisions are the only lead “scoring” that really matters. Strong SLAs are an absolute prerequisite for success.
- Continuous feedback. All lead scoring needs to be iterative; it will improve over time with feedback. This feedback should come from downstream conversion results, and those results should control for the effort applied. Statistical techniques should be used to rule out the variability in effort that inevitably applies to lead follow-up (even when there are SLAs) in order to see how accurate the scoring really was.
So, if you know lead conversion could be better and you’re thinking of lead scoring as a solution, excellent. Consider all options, including predictive lead scoring, and keep in mind the importance of alignment, service level-agreements and iteration.