Sales Says the Leads Are Bad. Marketing Says Volume Is Fine. Here's How to Settle It.

You know the meeting.

Sales opens with a story about three terrible calls in a row. Marketing pulls up a dashboard showing lead volume up 34% and cost per lead down. Someone suggests better lead scoring. Someone else suggests faster follow-up. Nothing gets decided, and the same meeting happens again next month.

The argument never resolves because both sides are looking at real data that supports their position. Marketing's numbers are accurate. Sales' experience is accurate. They are measuring different things, and the number that would settle it usually does not exist in either system.

Here is what that number is and how to get it.

Why Both Teams Are Right

Marketing measures what happens up to the form submission. Volume, cost per lead, conversion rate on the landing page. All of it is real and all of it can improve while the business gets worse.

Sales measures what happens after. Did the person answer, did they have budget, were they anywhere near ready to buy. Also real.

Between those two views sits the metric almost nobody tracks properly: MQL to SQL conversion, broken down by source.

Aggregate benchmarks tell you the scale of the gap. The cross-industry median MQL-to-SQL conversion rate sits around 13%. B2B SaaS runs 18–22%. Top performers with tight qualification reach 30–40%.

Put another way: at average rates, 87 out of every 100 marketing-qualified leads never become sales-qualified. Roughly 79% of marketing-generated leads never convert to sales at all.

So when sales says most leads are bad, they are describing the industry norm, not an anomaly. And when marketing says volume is fine, they are also correct. Neither statement diagnoses anything.

The Number That Ends the Argument

MQL-to-SQL conversion varies enormously by channel. Not slightly, by a factor of three or more.

Same company. Same sales team. Same definition of qualified. Wildly different outcomes depending on where the lead came from.

This is what makes the argument unresolvable when you only look at totals. If your channel mix shifted toward paid social over the last two quarters, your aggregate lead quality declined even though nothing about your sales team or your qualification criteria changed.

Marketing did not send bad leads. Marketing sent different leads. Volume went up because paid social is cheaper per form fill. Quality went down for exactly the same reason.

You cannot see this in a lead volume report. You can only see it when MQL-to-SQL is broken down by source.

The Three Real Explanations

Once you have that breakdown, the cause is usually one of three things.

1. Channel mix shifted

The most common. Cost per lead improved because spend moved to cheaper channels with lower declared intent. The dashboard rewarded the shift. The pipeline did not.

How you know: MQL-to-SQL held steady within each channel, but the blended rate dropped because the mix changed.

2. Your definition of "qualified" is too generous

Marketing defines MQL as anyone who filled a form. Sales defines it as someone ready to buy this quarter. Both use the same word.

An MQL-to-SQL rate below 15% is usually a definition problem rather than a sales execution problem. If marketing is compensated on MQL volume, the definition will drift toward generosity over time; not through bad faith, but because the incentive points that way.

How you know: MQL-to-SQL dropped across every channel simultaneously, with no change in mix or spend.

3. Follow-up speed collapsed

This one is invisible to marketing and usually invisible to sales leadership too.

Landmark Harvard Business Review research analysed 1.25 million leads across 29 B2C and 13 B2B companies. Firms that attempted contact within one hour were nearly seven times more likely to have a meaningful conversation with a decision maker than firms that waited just one hour longer and more than sixty times more likely than those waiting 24 hours or more.

The same research audited 2,241 companies and found that 23% never responded to a web-generated test lead at all.

Contemporary data shows the same pattern: follow-up within the first hour produces roughly 53% MQL-to-SQL conversion, against 17% for follow-ups after 24 hours. A three-times difference from speed alone.

How you know: lead quality complaints correlate with volume spikes. When more leads arrive, response time slips, and conversion drops for reasons that have nothing to do with lead quality.

The Diagnostic

Four steps. You need CRM data and analytics data in the same view, which is usually the hard part.

Step 1 : Break MQL-to-SQL down by source. Not by campaign, by source and medium. If your CRM does not capture the original traffic source on the lead record, that is the first thing to fix. Without it, none of this analysis is possible.

Step 2 : Compare against the channel benchmarks above. A paid social channel converting at 12% is performing normally. A paid search channel converting at 12% is broken.

Step 3 : Plot response time against conversion. Bucket leads by time-to-first-contact: under an hour, one to four hours, four to twenty-four, over twenty-four. If the drop-off is steep, you have found your problem and it is not lead quality.

Step 4 : Check whether the definition moved. Pull MQL criteria from twelve months ago and compare it to today. Definitions drift quietly, usually toward whatever makes the volume target achievable.

The Deeper Problem: You Are Optimising for the Wrong Event

Here is the part that turns a reporting problem into a spending problem.

Your ad platforms learn from the conversion events you send them. If you send a conversion event on form submission, Meta and Google go looking for more people who submit forms.

They are extremely good at this. And people who fill in forms are not the same population as people who buy.

So the algorithm optimises toward form-fillers. Cost per lead drops. Volume climbs. Marketing's dashboard improves. Sales' pipeline does not, because the ad platform was never optimising for pipeline, it was optimising for the event you told it mattered.

This is the same failure pattern we see in ecommerce, where brands send a purchase event for orders that later come back as returns, and unwittingly train the algorithm to find customers who place orders and refuse delivery. We covered that version in our post on RTO as a measurement problem.

Different funnel, identical mechanism: the algorithm optimises for whatever event you call success.

How to Fix It

Send qualified lead events back to the ad platforms. Not form fills; qualified leads, fired when sales accepts the lead. Meta's Conversions API and Google's offline conversion import both support this. It requires passing an identifier at form submission and matching it back when qualification happens, which is a real implementation project rather than a settings change.

The payoff is that your ad platforms start optimising toward people who become pipeline rather than people who complete forms.

Pass the traffic source into your CRM. Capture source, medium, and campaign as hidden fields on every form. Without this, you cannot run the channel breakdown, and every conversation about lead quality stays anecdotal.

Reconcile your CRM and analytics monthly. Lead counts in GA4 and lead counts in your CRM will differ, for the same structural reasons that GA4 and ecommerce platforms differ; event loss, attribution windows, duplicate submissions. Document the expected gap so you can spot when it changes. Our guide to reconciling revenue across platforms covers the same method applied to ecommerce.

Change what marketing is measured on. If marketing is accountable for MQL volume, MQL volume is what improves. Measure marketing on SQLs and pipeline contribution instead. As one benchmark analysis puts it: 500 MQLs producing 65 SQLs is more valuable than 2,000 MQLs producing 60.

The Uncomfortable Conclusion

Most lead quality arguments are actually measurement arguments wearing a costume.

Sales is not wrong about the calls. Marketing is not wrong about the numbers. The disagreement persists because the organisation is measuring two ends of a funnel and nothing in the middle and because the ad platforms are being trained on an event that only one of the two teams cares about.

Fix the measurement and the argument mostly resolves itself. You will find out whether the mix shifted, whether the definition drifted, or whether nobody is calling the leads back. All three are solvable. None of them are solvable while the two teams are comparing different dashboards.

If your CRM does not know where your leads came from, that is where to start. Everything else in this post depends on it.

Running paid lead gen and stuck in this argument? Talk to FunnelFreaks, we connect analytics to CRM data so lead quality becomes a number rather than an opinion.