15: Lifecycle: A Martech Saga part 4: Picking the right MQL model
You need a good MQL model so that marketing leads make it to sales and get followed up. There are a lot of ways to define MQLs and pass them over. It’s very common to have a lead scoring model, and it’s the best way to get to build a scalable, highly automated lifecycle. But let’s actually examine the journey that gets us to a scoring model and benefit from critically thinking about why it’s the best model.
We’re Martech geeks -- of course you’re going to say to deploy a lead scoring model -- but why is it important to imagine a universe without one?
Same goes for MQLS - we’ve accepted scoring as the definition of MQLs without always thinking it through. For me, an marketing qualified lead is a lead that marketing has qualified.
When marketing qualifies a lead, it’s passed to sales, sales follows up with it, and you make more money.
Exactly. We get stuck on the how and what too often. Why is this important?
Marketing is casting the net -- they build personas, execute on strategy to fill the funnel, often even own the automation systems. Marketing also deals with leads at scale -- one to many communications. It makes a lot of sense organizationally that marketing helps filter leads to sales.
By recentering on the why, we can now talk about the how and the what. Let’s start with the what:
Marketing could define an MQL as any of the following:
- A direct response to a marketing campaign through a form or offer acceptance
- Hand-bombing leads over from a list, for example from a conference booth
- Automated scoring!
- Numeric scoring
- Grade Score
- Fancy AI algorithm
Most common is numeric.
Good start and familiar toolset. Evaluate properties like country, industry, job title, etc. Evaluate behaviour like web and email interactions. Don’t want to get lost here but some amazing touch points that lead to purchase intent like what pages they viewed, pricing page counter, integration pages, where they started they trials.
Pros -> Super easy to implement, easy to maintain, easy to understand (and therefore trust).
Cons -> Harder to extract insights from, a bit basic in some cases, and sometimes you want more sophistication.
Data enrichment tools like Clearbit, not 100% match rate but help you figure out what matters, then you can ask that question instead of inferring it.
Grading model: Two axes: Fit & Engagement (or whatever). Get your 1-4 and your A-D. Matrix to plot out where leads land. Lots of precision and predictability.
Pros -> Precise, easy to understand, easier to extract insights.
Cons -> Harder to implement, harder to train folks on, more technical stuff
AI algorithm: Usually you plug in list of best customers, AI looks up common attributes and then sets up predictive model based on those attributes. Usually pretty black box.
Pros -> Easy to set up, sophisticated, and uses latest tech.
Cons -> Expensive, requires trust.
Thanks for listening homies.
If you absolutely can't wait 7 days for our finale, part 5, we'll give you a super secret link to the unpublished episode if you sign up for new episode notifications here humansofmartech.com. :)