Marketing Funnel Audit: Finding Where Your Leads Actually Drop Off
A client came to us with a spreadsheet that made no sense on its own: traffic was up 40% year over year, the sales team was closing roughly the same number of deals as always, and marketing spend had crept up without a clear return. Nobody could say why. The instinct in that room was to blame the website, "the site isn't converting," but that's a conclusion, not a diagnosis. Traffic and sales are just the two endpoints of a much longer chain, and somewhere in the middle of that chain, leads were disappearing without anyone noticing exactly where.
That's what a marketing funnel audit is for. It's a structured review of every stage a prospect moves through between the first time they hear about you and the moment they become a customer, with a measured drop-off rate at each stage, not just a before-and-after comparison of top-line traffic and bottom-line revenue. Most businesses only ever look at the two ends. That's like trying to find a leak in a long pipe by checking the tap and the meter and ignoring the forty feet of pipe in between. The leak could be anywhere, and if you don't measure stage by stage, you'll never know which fix actually matters.
What a Funnel Audit Actually Involves
A proper marketing funnel audit does three things, in order:
- Defines the stages a real prospect passes through for your specific business, not a textbook version of a funnel.
- Measures the volume and drop-off rate at each stage transition, using actual data rather than assumptions.
- Diagnoses the cause of each meaningful drop, so you know whether you're looking at a traffic problem, a messaging problem, a friction problem, or something that isn't actually a problem at all.
The output isn't a report that says "your funnel needs work": every funnel needs work, that's not useful information. The output is a ranked list: here is the stage where you're losing the most potential revenue, here's why it's happening, and here's what to fix first.
Defining Your Funnel Stages (Not the Generic Template)
Every marketing blog shows the same five-stage funnel (awareness, engagement, lead capture, qualification, conversion), and it's a reasonable starting skeleton. But applying it without adapting it to how your business actually sells is one of the most common ways a funnel audit goes wrong before it even starts.
Take that generic skeleton and ask, honestly, what each stage looks like for you:
- Awareness / traffic: how people first encounter you: organic search, paid ads, referrals, social, direct. This is visits, not visitors interested in buying yet.
- Engagement: someone actually consuming content, browsing more than one page, watching a demo video, reading a case study. This is the stage that separates a bounce from a real prospect.
- Lead capture: the point where an anonymous visitor becomes an identifiable one: a form fill, a booked call, a downloaded resource, a newsletter signup, a chat conversation.
- Qualification: for businesses with a sales process, this is where a lead is assessed against budget, fit, timing, or authority. For self-serve products, this might instead be a free-trial signup or a pricing-page visit.
- Conversion / sale: the actual purchase, signed contract, or closed deal.
If you sell a $50 product through an ecommerce store, "qualification" might not exist as a distinct stage at all: add-to-cart and checkout initiation do that job instead. If you sell a $30,000 B2B contract, the funnel might have two or three qualification sub-stages (marketing-qualified lead, sales-qualified lead, proposal sent) that each deserve their own drop-off number, because a 60% drop between MQL and SQL means something completely different from a 60% drop between proposal and close. The point isn't to copy a diagram. It's to map the actual sequence of things a real prospect does before they buy from you, in the order they do it, using language your own team already uses internally. If your sales team calls something a "qualified opportunity," use that term in the audit instead of inventing new terminology nobody recognizes.
Measuring Drop-Off at Each Stage
Once the stages are defined, the audit needs real numbers at each transition, not estimates. In practice this means pulling from a few sources:
- Analytics platform (GA4 or equivalent) for traffic volume, engagement rate, and on-site funnel visualization. GA4's Explore reports can build a funnel exploration with your defined steps and show exact drop-off percentages between them.
- Form and CRM data for lead capture and qualification numbers: how many form submissions became leads a salesperson actually followed up on, and how many of those became qualified.
- CRM pipeline reports for qualification-to-close rates, sales cycle length, and where deals stall.
- Ad platform reports (Google Ads, Meta Ads) segmented alongside analytics, so you can see whether drop-off differs by channel: paid traffic and organic traffic often behave very differently at the same funnel stage.
Build one table with volume at each stage and the percentage drop to the next. Do this by channel and by time period where you can, because an average across all traffic sources hides more than it reveals: a channel bringing in cheap but low-intent traffic will drag down your average engagement rate even if your best channel is performing fine.
A Genuine Problem vs. a Stage That's Supposed to Filter People Out
This is the part that trips up a lot of first-time funnel audits: not every drop-off is bad. Some stages exist specifically to filter people out, and a high drop-off rate there is the system working correctly.
A qualification stage that filters out 70% of leads because they don't have budget or don't fit your ideal customer profile isn't broken. It's doing its job, assuming those leads genuinely weren't a fit. The question to ask at every stage isn't "is this drop-off number high?" It's "are the people dropping off here people we wanted to keep, or people we were right to lose?" A landing page that loses visitors who searched for something unrelated and landed there by accident isn't leaking real leads. It's correctly not converting people who were never prospects.
The genuine problems look different: a drop-off that's higher than industry norms for that specific stage type, a drop-off that's inconsistent across otherwise-similar traffic sources (meaning something in the experience, not the audience, is causing it), or a drop-off that increased after a specific change: a redesign, a new form, a pricing update. Those are signals worth chasing. A qualification stage filtering out unqualified leads at a stable, expected rate is not.
Diagnosing the Cause Once You've Found the Leak
Finding the stage with the problem is only half the audit. The same drop-off number can have completely different causes, and the fix only works if it matches the actual cause. Broadly, a leak at any given stage comes down to one of four things:
1. Traffic Quality
The people arriving at that stage were never a good fit to begin with. This shows up as a channel or campaign with a much steeper drop-off than your other sources feeding the same stage. A paid campaign built around a broad keyword or a broad audience can flood the top of the funnel with volume while quietly wrecking every conversion metric below it, because the traffic itself doesn't match what you're offering.
2. Messaging / Relevance
The traffic is fine, but what they land on doesn't match what they expected or what would move them forward. An ad promising one thing and a landing page delivering something else. A pricing page that doesn't answer the question the visitor actually has at that point in their decision. This tends to show up as a drop that's consistent across traffic sources: it's not about who's arriving, it's about what they find.
3. Friction / UX
The offer and the audience are both right, but the mechanics of moving forward are harder than they need to be. A form asking for too much information too early. A checkout flow with unexpected steps. A "book a call" button that's genuinely hard to find on mobile. This is often diagnosable through session recordings, heatmaps, or simply testing the flow yourself on a phone with a slow connection.
4. Cost / Fit Mismatch
Sometimes the drop-off is accurate information, not a fixable problem. The price is genuinely higher than what that segment is willing to pay, or the product genuinely doesn't solve their specific problem. No amount of UX polish fixes a real value mismatch: the fix here is either a pricing or packaging change, a different target segment, or accepting that this particular drop-off is correctly filtering people out (see above).
Separating these four causes usually takes more than the funnel numbers alone: a handful of exit-intent surveys, a few recorded sales calls, or an hour with session recordings at the specific stage usually tells you more than another week of aggregate data will. For a closer look at diagnosing exactly this kind of leak once traffic reaches your website specifically, our piece on why a site gets traffic but not customers walks through the on-site version of this same diagnostic process in more detail. This article is the wider view across the whole marketing funnel, from first touch through to sale, of which the website is only one part.
Prioritizing Which Leak to Fix First
Once you have drop-off percentages at every stage, the instinct is usually to fix whichever stage has the worst-looking number. That's often the wrong place to start. A 90% drop-off at a stage that only 50 people reach per month matters a lot less than a 20% drop-off at a stage that 5,000 people reach per month.
The number that actually matters is volume lost, not drop-off percentage in isolation: roughly, the number of people entering that stage multiplied by the percentage that drops off, translated into how much revenue that represents downstream. A stage that feels the most broken because the percentage looks dramatic is often not where the money is. A stage with a merely mediocre percentage but enormous volume is frequently where a small improvement produces the biggest absolute gain.
There's a second factor worth weighing alongside volume: how fixable the cause is. A traffic-quality problem on a single underperforming campaign can sometimes be fixed in a day by pausing or retargeting that campaign. A friction problem on a form might be a one-week fix. A genuine cost/fit mismatch might require a pricing strategy change that takes months to properly test. When two leaks are close in lost-volume terms, start with the one you can act on fastest and measure the impact of before committing more time to the harder fix.
A Worked Example
Here's how this plays out on a hypothetical funnel for a mid-sized B2B services company running both organic and paid traffic.
| Stage | Volume (monthly) | Drop-off to next stage |
|---|---|---|
| Website visits | 10,000 | 82% |
| Engaged visits (2+ pages, 60s+) | 1,800 | 72% |
| Leads captured (form/call booked) | 500 | 60% |
| Qualified leads | 200 | 75% |
| Closed deals | 50 | N/A |
At first glance, the 82% drop between visits and engaged visits and the 75% drop between qualified leads and closed deals both look alarming. But run the volume math: the visits-to-engagement drop loses roughly 8,200 people a month, while the qualified-to-close drop loses 150 leads a month. In absolute terms, the top of the funnel is where the biggest number of prospects is disappearing, but before assuming that's the priority, the audit needs to separate cause from filtering.
Digging in: the 82% drop at the top turns out to be split unevenly by channel. Organic search traffic engages at a healthy 30% rate, close to what's expected for that stage. A specific paid campaign, however, brings in high volume at low cost but engages at under 5%: a traffic-quality problem, not a site problem. That campaign is inflating the overall drop-off number and making the whole top-of-funnel look worse than it actually is for the traffic that matters.
Meanwhile the 75% drop between qualified lead and closed deal, while it involves fewer total people, represents far more revenue per lead at this company's price point: those 150 lost qualified leads per month, at an average deal size of several thousand dollars, are worth more in absolute terms than fixing the paid campaign's targeting. Sales call recordings at that stage reveal a recurring pattern: qualified prospects are stalling after the proposal stage because the proposal document doesn't address a common objection about implementation timeline. That's a messaging fix, not a traffic fix, and it's cheap to test: rewrite the proposal template, track the next month's close rate, done.
The takeaway isn't "always fix the bottom of the funnel": it's that the worked example only reveals the real priority once volume, revenue-per-lead, and cause are all considered together, rather than reacting to whichever percentage looks the scariest on first read.
Turning the Audit Into Action
A funnel audit is only useful if it ends in a short, ranked list of specific fixes tied to specific stages: not a general sense that "the funnel needs work." Run it quarterly if your funnel and traffic mix are fairly stable, or monthly during periods of active change (new campaigns, a redesign, a pricing shift), since the biggest leak can move between stages as soon as you fix the current one. The businesses that get the most value from this exercise treat it as a recurring discipline rather than a one-time report that sits in a folder.
If you've got the traffic and sales numbers but no clear view of what's happening in between, that's exactly the kind of diagnostic work a structured conversion audit is built for. Our CRO service covers this stage-by-stage analysis in depth, or if you'd rather talk through your specific funnel first, get in touch and we can walk through it together.