Multi-Touch Attribution & Marketing Mix Modeling (MMM)Playbook3 min readUpdated September 2026

Where Is Your Funnel Actually Losing People?

"Our funnel is leaking somewhere" is a common complaint that's almost impossible to act on until you know exactly where. The aggregate conversion rate from visitor to customer tells you there's a problem but says nothing about which specific step is causing it. Here's a Q&A walking through how to actually locate a leak using your attribution and journey data.

Vendors Covered in this Article

Disclosure: We may earn a commission if you buy through some links on this page. It doesn't change what we recommend.

Where do we even start looking?

Break your funnel into distinct steps, visitor to lead, lead to marketing qualified, marketing qualified to sales qualified, sales qualified to closed, and calculate the conversion rate at each individual step rather than only the overall rate. The step with the lowest conversion rate relative to what you'd reasonably expect is where to focus first. It's common to discover the aggregate problem is really one specific step dragging the whole thing down, while the other steps are performing fine.

Write each step's conversion rate down side by side rather than looking at them one at a time in separate reports. The comparison is what reveals the outlier; a single step's number in isolation rarely looks obviously wrong on its own.

How do we know if a step's conversion rate is actually bad?

If your own landing pages are converting below the 6.6 percent all-industry median1, that's one useful comparison point for the top of the funnel specifically, though it's still just a reference, not a target. For steps further down the funnel, industry benchmarks are harder to find and less useful anyway, so your own historical trend for that same step is usually the better comparison.

What's different about a leak versus a slow step?

A leak is prospects dropping out of the funnel entirely at a given step. A slow step is prospects still moving through, just more slowly than expected. These look similar in a simple conversion rate but need different fixes: a leak usually points to a mismatch, wrong audience, unclear next step, or a broken form, while a slow step usually points to a process or follow-up timing issue rather than anything wrong with the step itself.

Confusing the two leads teams to fix the wrong thing. Redesigning a form because a step looks slow, when the real issue is delayed sales follow-up, spends effort on something that was never actually broken.

How do we tell which one we're dealing with?

Look at what happens to the prospects who don't convert at that step within a reasonable window, say sixty days. If most of them never come back and never convert through any other path, that's a leak. If a meaningful share eventually does convert, just later than expected, that's a slow step, and the fix is more about follow-up cadence than about the step itself.

For example, suppose many marketing qualified leads stall before any sales conversation. If most of them never reappear within sixty days, treat it as a leak and check the handoff rules and lead scoring criteria. If a meaningful share converts a few weeks later than expected, treat it as a slow step and tighten follow-up timing instead. Write the diagnosis down next to that step's conversion rate so the next review can confirm whether the change worked. As a decision rule, fix the leak first when high volume flows through the step, and fix the slow step first when it delays revenue on deals already moving.

What should we actually fix once we've found the leak?

  • If the leak is at the visitor-to-lead step, check message match between the traffic source and the landing page first
  • If it's at marketing qualified to sales qualified, check the handoff process and lead scoring criteria before assuming the leads themselves are weak
  • If it's at sales qualified to closed, that's less a marketing problem and more worth a joint review with sales on what's actually happening in those conversations
  • Re-measure the same step a month after any fix, since a leak fixed on paper doesn't always translate to a fixed conversion rate in practice

How often should we run this kind of analysis?

Quarterly is reasonable for a full step-by-step review, with a lighter monthly glance at whichever step was previously identified as weak, to confirm a fix is actually holding. Funnels shift as campaigns, offers, and the market change, so a step that was fine six months ago isn't guaranteed to still be fine now.

Keep a short record of what you checked and what you found each time, the same way you would for any recurring audit. Without that record, it's easy to re-discover the same leak a year later and waste time re-diagnosing something that was already understood and, supposedly, fixed.

Executive Capability Standard

What Good Looks Like

The standard is a step-by-step conversion rate breakdown that isolates the single weakest step, rather than reacting to one blended top-to-bottom conversion number.

Building The Capability (5-Stage Skill Ladder)

1. Learn:Break your funnel into distinct steps and calculate the conversion rate at each one separately to find where the real weak point is.
2. Do Manually:Manually investigate the weakest step each quarter, checking whether it's a genuine leak or a slow step with delayed conversion.
3. Delegate:Have a marketing operations or growth hire own the recurring step-by-step review and track whether fixes actually hold over time.
4. Automate:Build a dashboard that tracks conversion rate at each funnel step automatically, so a new leak surfaces without a manual quarterly pull.
5. Buy:Bring in conversion rate optimization help for a specific step once you've confirmed, with your own step-by-step data, exactly where the problem is.

How to Get Started

Disclosure: We may earn a commission if you buy through some links on this page. It doesn't change what we recommend.

Frequently Asked Questions

How many funnel steps should we track separately?

Enough to isolate a real problem without making the analysis unmanageable, typically four to six steps. Too few steps hides where the real leak is; too many makes it hard to act on any single one without the analysis becoming its own project.

What's the difference between a funnel leak and normal attrition?

Some drop-off at every step is normal and expected, not every visitor becomes a lead and not every lead becomes a customer. A leak specifically means a step's conversion rate is meaningfully worse than your own historical baseline or a reasonable benchmark for that step, not just that some attrition exists.

Should we fix the weakest step first even if it's not the biggest opportunity?

Not necessarily. Weigh the size of the leak against the volume flowing through that step. A modest leak at a high-volume early step can matter more in absolute terms than a severe leak at a low-volume late step, so prioritize by the actual number of prospects affected, not just the percentage.

Sources

Where we quote a benchmark, we show its source. Other figures in this guide are estimates or general guidance, so check them against your own numbers.

  1. Landing Page Conversion Rate (Median, All Industries). Unbounce Conversion Benchmark Report.

Related Guides