When 80% of Your MQLs Are Bad, More Marketing Makes the Problem Worse

An MQL count going up is not good news by itself. If eight of every ten marketing-qualified leads get disqualified before a rep will work them, the number that matters isn't MQL volume, it's the 80% you're throwing away, and pouring more spend into that same funnel just produces more of it, faster and at higher cost.

Here's a composite that will be uncomfortably familiar to anyone running demand gen or sitting in a pipeline review. Bridget Farrow ran marketing at a 140-person fintech infrastructure company I'll call Ledgerline, selling reconciliation software to regional banks and credit unions. Marketing was generating roughly 145 MQLs a month, and sales accepted 39 of them: a 73% disqualification rate everyone already knew was too high.

The board wanted more pipeline, so Bridget did what most marketing leaders are told to do. She doubled the budget: more paid search, a new webinar series, a looser scoring threshold to catch leads the old model was missing. MQLs climbed to 310 a month within a quarter. Sales accepted 42 of them. The disqualification rate didn't improve. It rose to 86%.

Nobody at Ledgerline had a lead-generation problem. They had a qualification-accuracy problem, and doubling the top of the funnel didn't touch it. It produced 165 more disqualified leads a month for someone in sales to open, read, and reject, which is its own cost even when the lead never becomes a conversation.

Key takeaways

• MQL volume and MQL quality move independently, and usually in opposite directions. Loosening a scoring threshold to hit a volume target predictably raises the disqualification rate, because the leads that clear a lower bar are, on average, worse fits than the ones that already cleared the higher one.
• A rising disqualification rate is a leading indicator, not sales being difficult. When more than roughly three-quarters of MQLs get rejected before a rep works them, the scoring model isn't finding intent. It's finding activity, and activity and intent aren't the same signal.
• More marketing spend against a broken scoring model doesn't dilute the bad leads, it multiplies them. Doubling volume at a flat or worsening disqualification rate doubles the absolute number of wasted sales hours too.
• Fixing this will shrink your reported MQL number before pipeline improves. That's the trade-off, and it's the one most marketing leaders won't raise with their own board unless someone forces the question first.
• Forrester's own research on lead-centric funnels found a typical inquiry-to-close conversion rate under 1%, which means the MQL-as-scoreboard model most teams still run was built to fail at scale, not just at your company.

The business problem: a scoreboard that rewards the wrong number

MQL count is the easiest marketing number to move, and that's exactly why it survives long past the point where it should have been replaced. A content sprint, a lower scoring threshold, a paid campaign aimed at a broader audience: any of these lifts the count within weeks, shows up green on a dashboard, and buys marketing another quarter before anyone asks what happened to the number underneath it, revenue attributable to those leads.

The problem is that MQL, as most teams score it, measures activity, not intent. Downloading a whitepaper, attending a webinar, visiting a pricing page twice: these are behaviors a scoring model can count, and none of them tell you whether the person behind them can buy anything. A disqualification rate above 75-80% usually means the model is scoring behavior it can observe instead of intent it can't, and calling the output "qualified" anyway.

I've written about the layer below this one, where a rising lead count hid a pipeline that wasn't actually growing, and the layer above it, where traffic climbed while identifiable demand stayed flat. MQL disqualification sits between the two. It's the mechanism that explains why volume at the top of the funnel and volume in the middle can both look healthy while the number that pays your salary doesn't move.

Why the usual fix, more marketing, makes it worse

When pipeline stalls and the CRO asks marketing to help, the default answer is more: more spend, more content, more channels, on the assumption that scaling the input scales the output proportionally. That assumption only holds if the scoring model underneath is accurate. If it isn't, scaling the input scales the error.

That's exactly what happened at Ledgerline. Doubling spend didn't recruit twice the qualified buyers, because the pool of people at Ledgerline's addressable accounts actually shopping for reconciliation software in a given quarter was roughly fixed. The additional spend recruited people adjacent to that pool: similar job titles, similar firmographic profile, no live problem or budget behind the click. A scoring model tuned to profile and engagement rather than verified intent waved almost all of them through as "qualified."

Loosening the scoring threshold is the version of this failure that does the most damage, because it launders a volume problem into a metric that looks fixed. MQL count goes up, the chart the board sees turns green, and the signal that actually matters, how many of those leads a rep can turn into a real conversation, gets worse in the same reporting period. Forrester's research on lead-centric funnels found a typical inquiry-to-close conversion rate under 1%, in large part because scoring models built on guesses about profile and engagement have no real mechanism for detecting a funded, timed buying decision.

The framework: track disqualification by reason, not just by rate

A single disqualification rate tells you something is wrong. It doesn't tell you what to fix. I ask every marketing team I work with, including inside ViitorCloud, to tag every disqualified MQL with one of four reason codes before that lead disappears from the CRM.

• Firmographic mismatch. The account is the wrong size, industry, or geography for what you sell, and the scoring model should have caught this before the lead was ever marked qualified.
• No budget authority present. The engaged person has no visibility into spend and no path to someone who does, which a scoring model built on individual activity can't detect on its own.
• Activity without a named problem. The lead is real and in-market by firmographic profile, but nothing in the engagement, a form fill, a download, a demo request, points to an actual, costed problem.
• Dead on timing. The fit and the problem are both real, but there's no forcing function, no budget cycle, no renewal, no compliance deadline, so the deal has nowhere to go this quarter.

The split matters because each reason has a different fix, and none of them is "more volume." Firmographic mismatch is a targeting fix upstream of the funnel. No budget authority is a scoring fix: weight the model toward title and seniority, not just engagement. Activity without a problem is a content and qualification-question fix. Dead on timing is a nurture and re-engagement problem, not a disqualification at all if you're tracking it honestly.

When Bridget ran this audit at Ledgerline, 61% of the new disqualifications from the doubled campaign fell into one bucket: firmographic mismatch. The paid expansion had targeted job titles that existed at both target accounts and completely unrelated companies, and the scoring model had no firmographic gate to filter the mismatch out. That's a targeting fix that costs an afternoon of campaign rework. It is not a reason to double the budget again.

What the research says about why this keeps happening

This isn't a Ledgerline-specific failure. First Page Sage's benchmark analysis of B2B SaaS client data puts the average MQL-to-SQL conversion rate at 13%, meaning an 87% disqualification rate is closer to the industry average than the exception. If your funnel already runs at 73% and you tune it toward "average," you're tuning toward failure, not fixing it.

The market has started to notice. HubSpot's 2026 State of Marketing research found lead quality and MQL accuracy are now the single most-tracked marketing KPI, at 39.4% of marketers, ahead of raw lead generation volume at 29.2%. That's a real shift from a decade of volume-first reporting, and it means a board that still opens with "how many MQLs did we generate" is running a metric marketers themselves have started to deprioritize.

The most direct evidence against the MQL scoreboard comes from the Forrester research cited above. After two decades of watching lead-centric processes, Forrester's own recommendation is to retire MQLs entirely in favor of tracking engaged buying groups, on the grounds that a model scoring individual behavior can't detect a group decision at all. You don't have to go that far to get the immediate benefit. Tracking disqualification by reason code gets you most of the value without rebuilding your entire revenue process in one quarter.

Here's the trade-off I'd rather name than soften. Fixing your scoring model and tightening firmographic and budget-authority gates will shrink your reported MQL count, likely by a third or more, in the same quarter you make the change. Your board will see a smaller top-of-funnel number and ask why marketing is "producing less." The honest answer is that marketing was never producing what the number claimed, and the smaller number is the first one you can actually defend when sales asks where the leads went.

The ViitorCloud perspective: scoring is a machine job, disqualifying is a judgment call

This is the same AI-Native argument I make about engineering work, applied to the top of the funnel. A model can score firmographic fit, weight engagement signals, and flag budget-authority gaps faster and more consistently than any human reviewer working a queue by hand. What it can't do on its own is decide whether a borderline account, wrong size but a strategic logo, or right size but ambiguous timing, deserves a human follow-up anyway. That call is still a judgment a person has to make, and it's the one that decides whether your funnel gets smarter over time or just gets automated at the same accuracy it started with.

As VP of Growth at ViitorCloud, I run every inbound lead from our own site, including the ones that come in off our case studies, through the same four reason codes before anyone on our team spends time on it. An inbound that reads three case studies and asks a scoped question about delivery timelines gets a same-day reply. An inbound with the right title at the wrong company size gets tagged firmographic mismatch and routed to nurture, not to a rep's calendar. The discipline is the same whether it's our funnel or a client's.

We now run this as a working engagement for revenue teams that suspect their MQL number is lying to them, informally called the Funnel Quality Diagnostic: we pull 90 days of MQL and disqualification data, tag every rejection by reason code, and hand back a scored breakdown instead of a recommendation to spend more. If your MQL count keeps climbing while your sales team keeps saying the leads are bad, ViitorCloud's technology consulting team runs the Funnel Quality Diagnostic against your actual CRM data, not a benchmark deck.

A disqualification audit you can run this week

This takes an afternoon against your last quarter of CRM data, not a new tool or a new hire.

• Pull every MQL disqualified in the last 90 days and tag each one with a reason code: firmographic mismatch, no budget authority, activity without a named problem, or dead on timing.
• Calculate the disqualification rate by lead source, not just in aggregate. A channel running at 95% disqualified is telling you something a blended 80% average is hiding.
• Check whether your scoring model has a firmographic gate at all, or whether it scores engagement first and firmographic fit as an afterthought.
• Ask your SDR team, by name, which reason code they'd assign to the last twenty leads they disqualified. Compare their answer to what the CRM data shows. The gap is your model's blind spot.
• Before approving any new budget for volume, model what happens to the absolute disqualified-lead count if the rate doesn't improve. That number is what you're actually asking sales to process.

Frequently asked questions

What counts as a "bad" MQL disqualification rate?

There's no single universal number, but First Page Sage's benchmark data puts average B2B SaaS MQL-to-SQL conversion at around 13%, meaning an 87% disqualification rate is closer to typical than alarming. A rate meaningfully above that, especially one that rises after a volume push, means your scoring model is finding activity, not intent.

Does a high disqualification rate mean marketing should stop generating MQLs?

No. It means the scoring model deciding what counts as "qualified" needs to change before the volume does. Cutting off lead generation without fixing the model just shrinks a broken number. Fixing the model first, then scaling, is the order that actually improves pipeline.

Why does doubling marketing spend make the disqualification rate worse instead of better?

Because the additional spend usually recruits people adjacent to your real buying pool, similar titles and firmographic profile, but without a live problem or budget behind the click. A scoring model that isn't gating on firmographic fit and budget authority waves most of them through as qualified, which raises the rate instead of diluting it.

Who should own fixing the disqualification rate, marketing or sales?

Both, but the reason-code data has to be shared and reviewed jointly, not owned by whichever side is currently getting blamed. Marketing owns the scoring model and targeting. Sales owns the ground truth on which leads were actually workable. Neither side can fix the number alone, because neither sees the whole failure by itself.