Your MQLs aren't converting because you're measuring the wrong signals

We see the same pattern constantly when auditing MQL-to-SQL dashboards. Either conversion rates are terrible (think 5-8%), or they're suspiciously high (80%+).

Both scenarios usually point to the same underlying challenge: the metrics we're tracking don't match buying reality.

Here's what the data tells us. The industry average MQL-to-SQL conversion rate sits around 13%. But companies that master advanced behavioural scoring? They hit closer to 40%.

That's not a marginal improvement. It's a fundamental difference in how we define lead quality. The gap exists because most teams are solving the wrong problem: chasing MQL volume when we should be identifying buying signals.

The Volume Trap

When we audit lead scoring models, the red flags often appear in the definition of "success".

Low MQL volume with high conversion rates often means the definition is too strict. We're likely leaving qualified leads on the table because the criteria are overly conservative. If we're generating plenty of contacts but few MQLs, we might be filtering out future champions too early.

High MQL volume with low conversion rates is the more common problem. We hit our lead targets, celebrate the numbers, but sales can't work with what we're sending them. In fact, research shows that only 27% of leads passed from marketing to sales are actually qualified.

That's not just a scoring problem. That's a definition problem.

Why the "Generic" ICP Fails Us

MQL definitions often break down due to poor ICP definitions.

Early-stage companies often build their ICP purely from sales feedback. The sales team says "we need enterprise leads," and marketing runs with it. But that high-level feedback hides crucial nuance.

Not all enterprises are created equally. A company with 5,000 employees in financial services has different needs, budget cycles, and buying committees than a 5,000-person manufacturing firm.

Instead of a broad "Enterprise" label, effective teams work from micro-ICPs. For example, distinguishing between a CEO at a smaller company versus a Head of Procurement at an enterprise. The pain points differ, so the scoring signals should too.

When digging into the data to find what actually predicts conversion, look for granular firmographic signals:

• Specific employee number bands (not just "enterprise")
• Market location and regulatory environment
• Technologies already in their stack
• Size of specific teams (like marketing ops or sales development)

The Intent Hierarchy Nobody Tracks

Here's where it gets interesting. We need to measure intent depth, not just engagement.

Intent platforms show keyword-level content consumption, telling us an account is in-market before they ever interact with our ads. But on-site behaviour creates an intent hierarchy that most teams treat as flat.

A newsletter signup is low intent. A demo request is high intent. But a demo request isn't automatically an SQL - it needs to match our persona and ICP .

Context is everything. Students fill out demo forms. Competitors research us. People from target accounts deserve different treatment than random inbound.

The Account Scoring Reality

A Head of Marketing Ops colleague, Jack Powell, recently reminded me of a crucial reality regarding account scoring: most B2B purchases involve multiple people in the buying group.

Historically, this has been complicated to manage because most marketing automation platforms were built for contact scoring. But if we only track individual contact scores, we miss the forest for the trees. Multiple people from the same account engaging with low-level content is often a stronger signal than one person downloading a single whitepaper.

Furthermore, up to 98% of website visitors stay completely anonymous. If we rely solely on form fills (first-party data), we're missing the vast majority of the actual buying journey.

What We Can Fix Today

Most teams are stuck with legacy scoring models that reward form fills and email opens. We can't always rip everything out and start over, but we can make tactical adjustments now.

1. Start with page-level tracking. If a contact in the MQL phase shows an uplift in viewing technical articles or pricing pages, that's a signal worth acting on immediately.
2. Review nurture sequences. Ensure content is appropriate to the contact's vertical and pain point. Tight targeting gets better engagement.
3. Differentiate by channel. SEO-generated leads convert to SQL at roughly 51%, whereas paid search leads convert closer to 26%. We should score them differently.

The Diagnostic Framework

If you're realising your MQL volume looks great but conversion is terrible, here is a practical way to diagnose the issue.

1. Audit Your Definitions Really understand your MQL criteria versus SQL criteria. Are they aligned with sales reality?

2. Deep-Dive Your "Lost" MQLs

• How well did they actually fit the ICP?
• What was the average time to convert?
• Did we miss account-level signals (e.g., three different people from one company visiting)?

3. Map the Content Journey Are we answering the questions they're actually asking at each stage? The point isn't to create a more complex scoring model, but to understand if our current definitions are fit for purpose.

What to Measure Instead

Stop celebrating MQL volume alone. Start tracking metrics that predict revenue:

• MQL-to-SQL conversion rate by channel (e.g., Organic vs. Paid)
• Account-level engagement scores (aggregating anonymous and known activity)
• Intent signal strength (from keyword-level content consumption)

These metrics tell us if we're attracting genuine buying intent. MQL volume just tells us if our forms work.

The difference between those two things is the difference between a pipeline that converts and a dashboard that lies to you.

Ready to fix your definitions? Start by auditing your last 10 "lost" MQLs against the diagnostic framework above. If the pattern is still unclear, let's look at the data together.

Photo by Luan de Oliveira Silva on Unsplash