5 Replies Latest reply on Nov 7, 2013 2:05 AM by 5be2827a6d30249e286cfe869203c119ae9b0cda

    Lead Scoring/MQL Threshold

      I'm redoing my lead scoring model and am have been trying to determine what the best scoring threshold would be. I've run numerous reports that compare the old scoring model with my new scoring model and wonder if I'm overthinking this one question. So I'll open it up to the community and ask my question vaguely.

      What's the best way to determine the score threshold at which a lead should become an MQL?
        • Re: Lead Scoring/MQL Threshold
          Is there any reason for making a change? Are the leads that sales are receiving qualified?

          We have a MQL threshold of 75 and we have been statisfied with the decision thru the first 6 months. What scoring threshold do you currently use?

          tm
          • Re: Lead Scoring/MQL Threshold
            We're changing our model because we currently use the lead score field and were getting feedback from sales that our leads' scores were inflated due to demographic information - it's relevant information but we were sending over leads that hadn't taken enough behavioral activity. So I'm finally turning on the scoring model I built months ago (I had it running in the background so I could predict the change that it was going to make): http://community.marketo.com/MarketoDiscussionDetail?id=90650000000PeNtAAK

            Basically I'm splitting out demographic and behavior scores and am only sending leads to sales that have taken meaningful behavior.

            Long story short, my new model was going to send over too many leads so I need to tighten it up.

            A number like 75 is pretty relative, since I don't know the score values of individual conversion activities. Our conversion activities are scored fairly low, so our threshold is also very low. Let me rephrase the question?

            I want to know how to determine my threshold:
            •           Pull a report of all leads that have taken any activity on our site and find the average behavior score for them? Then pick a number above the average?
            •           Pull a report of all leads that have converted into an opportunity and find the average behavior score for them? Then pick a number above the average?
            •           Find the number that would represent a few form fillouts and pick that number?
            These methods don't seem scientific enough for my analytical mind. I can figure out conversion to opportunity rates for different conversion types on my website, but not conversion to opportunity rates based on score since that value is constantly changing (we decrease due to inactivity.)

            Is there a more scientific method then pick a number and adjust based on sales feedback?

            I fear I may be overthinking this.
            • Re: Lead Scoring/MQL Threshold
              Have you thought about asking some of your mor experienced sales reps to identify the combination of actions that they think represent a very qualified lead (behaviorally speaking), and then backing out what the average point total would be for those combinations?

              From the 3 options you listed for determining your threshold, I would probably start by looking at average scores for leads that have been associated with closed-won opportunities.
              • Re: Lead Scoring/MQL Threshold
                Josh Perry
                I would suggest doing what Kate talked about.  We did a half day pow wow with a small number of our experienced reps to come up with our scoring.  
                • Re: Lead Scoring/MQL Threshold
                  I don't think you're overthinking it Kim - you want the most objective metric you can obtain to ensure your lead scoring is more accurate second time around!

                  I'd do pretty much what Kate suggested, but make sure you've been generating lead scores for a long enough period (i.e. big enough sample size) so that the average score has weight to it. Also, I'd suggest that you implement your behavioural/demographic split scoring & retrospectively apply this before calculating the average, so you can utilise averages for behavioural/demographic scores separately.