Your MQL Definition Is a Sales Problem, Not a Marketing One

The argument that never resolves

Every B2B company I've worked with has some version of this standoff. Marketing reports MQLs are up. Sales says the leads are rubbish. Marketing points at the scoring model. Sales points at their calendar.

The usual response is technical: rebuild the scoring model. Add behavioural weighting, add firmographic fit, add a decay function, integrate an enrichment tool.

I've done this. It doesn't work, and the reason is that the disagreement was never about the model. It's about who gets to decide what counts as a real opportunity — and that's a negotiation, not a data problem.

---

Why a better model doesn't fix it

A scoring model encodes a definition. If marketing builds the model alone, marketing has unilaterally defined what sales should spend their time on.

Sales did not agree to that definition. So they do what any rational team does when handed a queue they didn't help design: they work the leads that look promising by their own criteria and ignore the rest, while the reported MQL number keeps climbing.

Now you have two definitions running in parallel — one in the CRM, one in the sales team's heads — and the gap between them is the thing everyone is arguing about. No amount of model refinement closes a gap that exists because one side wasn't in the room.

---

The session that actually fixes it

This takes about two hours and it needs sales in it — ideally the person who runs the team, plus one or two reps who actually work the leads.

Step one: bring 20 real leads, not criteria.

Not a definitions document. Twenty actual leads from the last quarter, anonymised enough to be discussable, with everything you know about each: company, role, what they did, what happened next.

Step two: sort them into three piles.

Sales sorts. I would have called this immediately. I would have called this eventually. I would not have called this.

Don't debate during the sorting. Just sort.

Step three: ask why on the disagreements.

The interesting leads are the ones where the score and the pile disagree — high score, wouldn't call; low score, would call immediately.

This is where the actual criteria come out, and they are almost never the ones in the model. In one session the real signal turned out to be whether more than one person from the same company had visited in a fortnight. In another it was job title seniority combined with company headcount, and nothing behavioural mattered at all. In a third, sales were reacting to a signal marketing wasn't capturing anywhere: whether the company had recently posted a relevant job ad.

Step four: write the definition together, in one sentence.

Then get both sides to say it out loud. If either party hesitates, you haven't finished.

---

The part everyone skips: the reciprocal commitment

A joint definition only holds if both sides commit to something.

Marketing commits to only passing leads that meet the definition — including when that means the monthly number drops.

Sales commits to working every lead that meets it, within an agreed time, and logging a reason when they don't convert.

That second commitment is the one that gets negotiated away, and without it the whole thing collapses within a quarter. If sales can silently ignore qualifying leads, you're back to two definitions and no visibility into which one is right.

---

What to expect afterwards

MQL volume drops. Typically by a third to a half. This looks like failure on a dashboard and needs to be pre-agreed with whoever reads that dashboard, or you'll spend the next month defending it.

Conversion rate rises immediately. Partly because the leads are better, partly because sales now believes in the queue and works it differently. Both effects are real and you generally can't separate them.

The definition goes stale in about six months. The market moves, the product changes, the ICP shifts. Put a recurring session in the calendar rather than waiting for the argument to resurface.

---

The reframe

Stop treating lead quality as a measurement problem that marketing owns and start treating it as an agreement that two teams maintain.

The scoring model is just where you write the agreement down. It's worth exactly as much as the conversation that produced it — which is why the same model can work brilliantly at one company and get ignored entirely at another.