One theme that has been consistent in my conversations with B2B marketers over the past few months is this: Lead scoring is one of the greatest opportunities and challenges when it comes to implementing and tuning their marketing automation processes and systems. Marketers tend to get the basics -- especially when it comes to core ideas such as scoring against target demographics and applying 'BANT' analysis (i.e., budget, authority, needs and timing). But B2B marketers seem to really stumble in taking their scoring and subsequent routing and nurturing to the next level. In fact, in research Silverpop is releasing this week, my colleagues found that 53% of B2B marketers still don't score, and 69% don't nurture (which requires some level of scoring and is a casualty of low scoring rates).

This issue is especially front and center for me, given I'm in the midst of developing my presentation for a lead scoring Webinar I'm giving with the folks at Target Marketing on Thursday. My presentation will be rooted in the basics, but I'm increasingly of the mindset that such a Webinar also needs to cover the 'advanced' issues, too. After all, this seems to be where scoring falls down.

What are these advanced 'lead scoring secrets'?

So while my Webinar will go into more details -- and also ground these 'advanced' topics in context and frameworks -- I thought I'd present some of the working insights I think can really help B2B marketers take their games to the next level ... and find real success with scoring. So here goes:

> Build the concepts of dialogue and momentum into your model: Given an environment in which B2B buyers have more information power than ever, and sales teams are being pulled into the buyer dialogue later and later, a new opportunity has emerged for B2B marketing organizations to be the organizational 'point person' on engaging with, managing and providing continuity in the pre-sale dialogue with buyers. The goal is simple -- nurturing a lead until it is sales ready -- but to do this marketers have to operate like a tenured sales professional, carefully managing the nurturing process in response to the buyer's signals.

"Industry statistics show that up to 40% of leads may make their first purchase after having been in the 'remarketing database' [i.e., nurturing pool] for 18 months or longer," notes David Taber in a recent Computerworld piece. "This is the whole purpose of marketing automation systems ..."

Yet the sales concepts of building a dialogue with a prospect and of understanding when a buyer gains momentum too often are not at the heart of scoring -- especially when there is over-reliance on demographic and BANT data. Moreover, behavioral score components should distinguish between activity that is increasing versus activity that is decreasing.

SiriusDecisions explains in a recent research brief that your scoring must understand "... the 'arc of activity' that buyers tend to use." Combinations of activity that build on each other -- as a consistent 'dialogue' and that demonstrate momentum in propensity to buy -- should increase the score. Similarly, lead scores should 'decay' after periods of inactivity -- demonstrating declining momentum.

> Leverage insights from the communication channel and the nature of the information 'consumed' by the prospect to better assess a buyer's relative maturity: This builds on the previous point and is perhaps one of the most mis-understood of the factors that go into 'great' lead scoring. Research into integrated marketing communication programs and buyers' information search patterns show that different communication channels and types of information are sought at different stages in the buying process. (Note that this pattern will differ by company, product and industry.) Observing this activity can indicate the relative maturity (and momentum) of the buyer in their search process; thus, it should be a key factor in increasing and decreasing a lead's score.

> Make sure your lead score captures insights from both online AND offline activity: This also builds on the previous point and is another area that B2B marketers fail to fully integrate into their scoring methodology. If your score only takes into account online activity, it is not a complete picture. Make sure that event attendance, inbound calls and other offline behaviors that are integral to the buyer's process -- and that also indicate relative maturity and momentum -- are baked into your scoring methodology.

> Expect your scoring model to change: Before you can even build your scoring model, you will have to examine past campaigns and historical data and conduct conversations with both marketing and sales team members. You will need to look for correlations that exist in your core business logic between marketing/sales actions and propensity to buy. You will make initial assumptions about relationships between factors and build initial score models ... and yet your model still won't be right.

"All successful [marketing] processes are ongoing in nature," explains Steve Gershik in a past post on his blog, The Innovative Marketer. "Tweak your programs, tweak your scores, change the metrics you look at to analyze the scores of your leads. Be open and flexible when you get started and you'll find you have a program that your whole team, marketing and sales, buy into."

It is only through constant testing and monitoring that your lead model will mature. But this makes sense. After all, what you are building in the lead score model is the heart of an ongoing set of demand-generation process -- a lead factory -- that requires care and maintenance. Silverpop's Lead Management Workbook adds: "A solid lead-scoring approach not only helps you to rank prospects against each other, but can smooth the lead flow and help you build a more powerful and accountable marketing organization based on rigorous analysis and testing, rather than intuition and educated guesswork."

> Constantly re-assess lead score data: Underlying your changing model is a constantly-changing set of behavioral data from your buyers. Maintaining accurate scores requires constantly re-assessing and updating the data. "Allow scores to be updated with third-party information such as data-appends or data entry by your sales force," suggests the Lead Management Workbook (cited above). "Automatically add new data as it is gathered over time and re-score leads."

> Work on tuning the relative 'elasticity' of variables in your model: Perhaps the most fundamental 'integrity' issue for lead score models -- firmly rooted in the disciplines of economics and of linear regression -- is the idea of relative elasticities. I.e., different elements of your score will have varying degrees of impact on a prospect's 'propensity to buy.' So make sure your scoring model reflects this.

"The actual score doesn't matter," explains Steve Gershik (cited above). "The important thing is that the point value is relative to other activities so in the end, the higher the score, the more actionable the lead is."

SiriusDecisions discussed this issue in a recent brief, "When Good Lead Scoring Models Go Bad." In one section of their brief, they call out the importance of tuning elasticity in the overall weighting of variables in a scoring model:

While it may be tempting to take five variables in an overall scoring model and weight them all equally at 20 percent of a prospect's overall viability, we advise you to resist. Based on your ideal customer profiling (combining this with activity and BANT variables if both are possible/intuitive to include), choose the variables you believe to be most predictive of a prospect's viability and prioritize your weighting there.

What do you think?

While it is not an exact science, lead scoring is critical to successful nurturing and lead management. This requires that as marketers we get under the covers a bit and address some of these 'advanced' issues -- which is what I wanted to do here. I hope that these ideas were helpful. What thoughts/ideas would you add to this list?

Posted by Adam Needles at 6:22 AM | Permalink | Comments (1)

Uncontrollable data (non structured): At times marketing will not have control over the way certain data is collected and brought into the marketing database. This is typically the case for tradeshows, list acquisition, and previously collected data. Your challenge then is to find a way to score this manually entered data without being able to exactly match it. The solution depends on the capabilities of your LSM.

Does your lead scoring model support wild card matches?

Yes (congratulations): For each field you plan to score, create a list of phrases that your LSM will search.

No: Try to find the most common occurrences of manually entered data and add them to your LSM. This approach is severely limited and might result in a high amount of score-able data not being identified.

Copied below is a sample lead score model using some of the ideas I've discussed thus far.

Closing points:

• Use cookies, de dup on submit, or other syncing and authentication methods which will ensure that data is appended to the existing lead's record rather than creating duplicates each time they submit a form (your demand generation software should do this for you)
• Use progressive scoring tactics by asking additional LSM questions with each interaction of your lead (1st form - 3 questions, 2nd form - next 2 questions, 3rd form - all 8 questions)
• Pre-populate the questions they have previously answered to increase their confidence in the relationship and save them time
• Structure data when possible for more accurate scoring

These simple steps will increase the simplicity and accuracy of your lead scoring model allowing marketers to send qualified leads to sales.

Posted by Bryan Brown at 2:35 PM | Permalink

Data collection is the start of an effective lead scoring model (LSM). This is a given for the advanced lead scoring guru, but for everyone else this is the often overlooked beginning point to develop a LSM that actually works. If you've recently implemented a LSM or are getting ready to start, here are a few pointers that will set you on the right path:

• Begin with a simple evaluation of the data entry points to your marketing database. Your list will likely include the following: tradeshow lists, Website forms, CRM data entry, landing pages, and third party sources.
• Group the data entry points into two categories: controllable data and uncontrollable data.

Controllable (structured) data: If marketing has the ability to control the way data is collected, they should seek to structure that data for the purpose of their lead scoring model. This is most often the case with Web forms, landing pages, and CRM data entry.

• Evaluate the questions on these lead forms. Most likely you have already been asking questions that are part of your manual lead evaluation process.
• Identify the basic questions that can be standardized across all your forms. Not that the same questions need to appear on every form, but when they appear they need to collect the data in standardized formats.
• Ensure that the data points you will score on are not open ended questions. For example: instead of asking people to type in their job title, include a drop down list of roles you would expect from leads/visitors to your site so you'll capture consistent responses.
• Follow best practices when creating forms. The number of questions should increase or decrease in harmony with the perceived value of the offer.

The image below shows an example of standardized questions one might use to score leads.

In my next blog I will share more information about handling the data you can't control and other "gotchas" that can destroy your lead-scoring model.

Posted by Bryan Brown at 10:07 AM | Permalink