Recommended Ratio between Behavior and Demographic Scores

[Anonymous]

Hi All,

Does anyone have a recommendation on the ratio of the "maximum" demographic score to behavior score ratio should be? For example, if the maximum demographic score possible is 100, what you recommend the maximum behavioral score should be - and how this related to total score?

Thank you!

[Guy_Goldstein6]

Hey Alex,

consider that the function of each of the components is distinct and that the ideal result may be more closely related to the relationship between the 2 scores rather than a ratio.

Ive done some very extensive scoring projects in the past and, after relearning SPSS, the data generally showed that demographic score, when limited to a fixed range (say 100) and classified into clusters (A,B,C,D) acted as a relatively good predicted of propensity to buy (whether or not the lead would ever buy) while behavior score (with an unlimited range) acted as a better indicator of readiness to buy (the likelihood that they would buy NOW).

the best performance was achieved not by focusing on the ratio between the 2, but in adjusting thresholds of how much behavior each demographic class required in order to indicate that a response was necessary. Or in English, better demographic leads need lower behavior scores to be ready for sales attention.

hope this helps

- Guy

[Anonymous]

Hi Guy,

Yes that makes a lot of sense. Do you have any recommended literature or reading related to what you just described?

The sales reps are swearing to me that they want to talk to every lead who meets our demographic score and they don't care at all about behavioral score (they want a score of 0). I'm nodding and smiling because I know that it will take some form of engagement to build up that demographic score to begin with.

[Guy_Goldstein6]

Hey Alex,

I haven't found the time to write it yet (Sadly I've never found any solid literature about scoring, but I happened to have studied statistics, had a love for number crunching, and had just come out of a job working with a team of academics - it was all just a perfect storm)

Sales people can be trying at times, but if they are asking for all of the leads your best bet is to give them what they ask for while collecting the data to justify the case you're trying to present to them in a way that makes it attractive to them.

One way that you might go about doing this is by passing them all leads with a demographic score of X or higher, but in the background start collecting data on behavior scores and measuring the behavior score when they either qualify or disqualify the lead. That way, you'll start to get a picture that should (hopefully) show that on average qualified leads have a higher behavior score at the time that sales qualify them (note the effort to make the qualification/disqualification the focus of your statistics). In addition you may also want to track qualification based on hitting certain scoring thresholds so that you can present the "quality" of behavior score categories - eg. Once a lead hits 20 points in behavior (with your defined demographic profile of X) it has a 10% chance of being qualified, once it hits 40 points that chance goes up to 15% etc. By doing this you present sales with the choice of having to work harder for the qualified leads (ie calling those with a lower likelihood to qualify/lower score) or enjoying the fruits of your labor and getting better quality leads, and you can back it all up with numbers.

One thing that it's important for you to note is that they may well be happy even with the 5% chance that leads with a 0 behavior score have because they consider that to be worthwhile, and that's totally OK. The goal of scoring is in part to get the right leads to sales at the right time, and in part to get sales the leads they want. As the number of leads that sales have to handle grows, they will have more incentive to ask for more qualified leads, but it may be that right now they simply need the funnel to be driving leads to them and they don't care how good they are because they are "lead starved".

Don't take it personally, it's not a criticism of your scoring or your efforts, sales being willing to accept lower quality leads is generally a function of sales being desperate to hit quotas and feeling like they aren't doing enough.

Hope this helps,

Guy

[Anonymous]

Thanks so much! I'm definitely going to take up your idea on analyzing behavior score in conjunction with success rate. I'm thinking of some basic logistic regressionv to show the predictive power of behavior score.

Yay statistics! I'm hoping to get my masters in business analytics some day

thanks again for your recommendations!

[Guy_Goldstein6 · Accepted solution]

Any time!

Let it me know how it goes, and if you have any issues ping me at guy@g8technological.com

[JD_Nelson]

A lot of good input on this thread already. I definitely think there is benefit to what your reps are saying... but it likely differs from company to company. We, for example, use demographic scoring (0-50) and pass anything 30+ to our reps (with a behavior score of 10+ which is basically one or two responses) and then further qualify anything demographic from 20-30 before passing it. Anything under 20 basically goes into our personalized nurture program that will get email returned to the rep (ultimately passing it to them if they respond)...

We are still early in our adoption, but that has been our latest game plan (always changing)...

I like Marketo's Sales Insights for their ranking of urgency (behavior over time) and score (demographic-ish). Since it's a ranking system of all available leads  it's a good basic ratio to use as your most likely prospects. Beyond that I think it gets very customized to your LOB and your buyer's profile.