The Marketing Funnel: Stages, Metrics, and Where Deals Actually Leak
A marketing funnel breaks the buyer's path into stages - awareness, consideration, conversion, retention - and each stage runs on a different metric. Funnels rarely fail everywhere at once; they leak at one specific seam, and finding that seam beats optimizing everything by instinct.
01What is a marketing funnel?
A marketing funnel is a model of the path a buyer takes from first noticing a brand to becoming a paying, repeat customer, usually split into awareness, consideration, conversion, and retention. Each stage carries its own metric, and most of the money a business loses hides in the gaps between stages.
The name comes from the shape: wide at the top, where anyone who's ever seen an ad or a search result counts as in it, and narrow at the bottom, where only paying customers remain. It's a simplification - real buyers loop back, compare prices in five tabs, disappear for months, and return - but the model still earns its keep. It forces a business to ask a sharper question than how do we get more customers. It asks which specific stage is losing them, and what that stage's own number actually looks like.
02The four stages, and the metric that matters at each
Most working versions of the funnel collapse into four stages. Each one answers a different question, and each has a metric that only makes sense inside that stage.
Mixing these up is the single most common funnel mistake: judging an awareness campaign by conversion rate, or judging a checkout page by reach. A TikTok clip pulling a 0.3% CTR can still be doing its job as an awareness asset if CPM stayed low and branded search volume rose afterward. A landing page converting at 4% is doing well even if only a few thousand people ever see it.
- Awareness (TOFU) focuses on pure reach, ahead of any intent to buy. Metric: CPM. Typical range: $5-$20 on Meta and Google Display, $8-$25 on TikTok and Instagram Reels; higher in finance and legal, lower in mass-consumer apps and games.
- Consideration (MOFU) happens when the audience knows the brand exists and is deciding whether it's relevant. Metric: click-through rate and opt-in rate. Typical range: 0.5%-2% CTR on cold paid traffic, 10%-25% opt-in rate on a landing page built for a warm click.
- Conversion (BOFU) is the moment money or a real commitment changes hands. Metric: conversion rate and cost per acquisition. Typical range: 1%-3% site-wide e-commerce conversion, 15%-30% trial-to-paid for self-serve SaaS, 15%-25% close rate once a qualified opportunity exists in sales-led B2B.
- Retention measures whether the customer sticks around long enough to be worth what it cost to acquire them. Metric: churn rate and repeat-purchase rate. Typical range: 3%-7% monthly churn for self-serve SaaS, 20%-40% repeat-purchase rate within 90 days for e-commerce.
03How conversion rates typically move between stages
Visitor to lead loses the most people, and that's fine - the top of the funnel is supposed to be wide. Cold paid traffic converts to an email address or a form fill at roughly 2%-5%; retarget that same audience with a warmer offer and the rate climbs to 10%-25%.
Lead to qualified lead: MQL to SQL, in B2B shorthand - runs 10%-30%, depending on how much friction sits inside the qualification step itself. A discovery call filters harder than a quiz ever will.
Qualified lead to paying customer closes at 15%-30% in sales-led B2B, and self-serve trial-to-paid usually lands in that same 15%-25% band. Freemium's free-to-paid conversion sits meaningfully lower, often 2%-5%, since a free tier attracts plenty of people who never intended to pay anything.
Customer to repeat customer is where compounding growth actually happens: 20%-50% of customers make a second purchase or renew within 90 days in a healthy business. That single number matters more than any of the ones above it, because a repeat customer costs close to nothing to reactivate.
04Worked example: tracing 10,000 visitors to customers
Take 10,000 visitors arriving from a mix of paid search and social - a believable month for a mid-size e-commerce or SaaS campaign. The figures below are round and hypothetical, built purely to show the mechanics.
10,000 visitors reach the landing page. At a 3% visitor-to-lead rate, that produces 300 leads - a signup, a cart start, a form fill. Of those 300, 25% qualify themselves further by opening a follow-up email, booking a call, or returning to finish checkout, leaving 75 qualified leads. Those 75 close at a 20% rate, producing 15 paying customers. Within 90 days, a third of them buy again or renew, adding 5 repeat purchases.
At a $150 average order value, those 15 customers generate $2,250 in first-purchase revenue before a single renewal is counted. If the media spend behind those 10,000 visitors ran $3,000, blended CAC lands at $200 - a number that only means something once it's checked against the average order value and the repeat-purchase rate sitting above it.
The funnel narrows from 10,000 to 15 - an overall conversion rate under 0.2%, which looks alarming until the ratio gets checked stage by stage instead of end to end. The visitor-to-lead step, at 3%, sits at the tight end of its typical 2%-5% band here. Lift it to 5% and everything downstream scales with it: 500 leads, 125 qualified, 25 customers - a 67% jump in paying customers from the same ad spend, because the fix landed at the weakest link instead of adding more traffic on top of an already-leaking top of funnel.
05Where marketing funnels leak, and how to spot it
The awareness-to-consideration seam usually leaks because the audience is imprecise, not because the creative is weak - a well-targeted list with a mediocre ad still tends to beat a broad list running a great one. Chasing a creative fix for a targeting problem burns a testing budget without ever touching the real issue.
The consideration-to-conversion seam leaks at the landing page: a headline that doesn't match what the ad promised, a form asking for more than the offer justifies, missing proof that the product actually works, a checkout with too many steps. This is exactly the seam a focused landing-page conversion audit exists to fix, and it's usually the cheapest leak to plug because the traffic is already paid for.
The conversion-to-retention seam leaks in the first week after purchase or signup - a new customer who never gets a clear next step, never opens the product a second time, never hears from the brand again until a renewal charge shows up on their card and gets disputed.
To spot the leak, don't chase whichever stage's absolute number looks smallest - a stage sending through only 50 people can be healthier than one sending through 500, if the percentage it keeps sits close to what's typical for that stage. Compare each stage's own conversion rate against the ranges above, then look for the one furthest below its normal band. That's the leak, and it's almost always cheaper to plug than to compensate for by buying more traffic at the top.
06Marketing funnel vs pirate metrics (AARRR)
The funnel isn't the only lens for this journey. Dave McClure's AARRR framework - Acquisition, Activation, Retention, Referral, Revenue, nicknamed pirate metrics for the acronym - slices the same buyer path differently, and product-led SaaS teams already tracking in-app events tend to reach for it instead of the classic funnel.
Neither model is more correct than the other; they're built for different reporting habits. Paid-media and ad-platform reporting speaks in funnel language because that's how Meta, Google, and TikTok structure their own dashboards. Product teams speak in AARRR because activation is a specific, trackable event inside the product - the moment a user experiences the thing they signed up for - and referral earns its own line since word-of-mouth compounds on a separate curve from paid acquisition.
07Where you meet the funnel in practice
GA4's funnel exploration report lets a team define each stage as an event and see the exact drop-off percentage between them. Meta Ads Manager and Google Ads both surface a similar funnel view once conversion tracking is set up correctly. A CRM's deal pipeline - new, qualified, proposal, closed-won - is a sales funnel wearing different labels for the same four-stage logic.
A tracking stack like Keitaro or Binom, sitting between the ad click and the postback that confirms a sale, is really a funnel-instrumentation layer. It's what confirms, server-to-server, whether the click that cost eighty cents in one geo and two dollars in another actually turned into revenue two or three stages later - the only honest way to know which channel's cheap clicks were worth buying and which were just cheap.
Auditing which stage is genuinely leaking, rather than assuming the media buy is at fault, is most of what a funnel review with a consultant like Ioann Putevoy looks like in practice: pull the numbers stage by stage, compare them to what's normal for that stage and vertical, and fix the seam that's actually underperforming.
08Related terms worth knowing
A handful of terms show up constantly alongside the funnel, and each one answers a slightly different question than the funnel itself does.
- CAC (customer acquisition cost) - the fully-loaded cost of turning a stranger into a paying customer, sitting at the bottom of the conversion stage.
- LTV (lifetime value) is the total revenue a customer generates over their whole relationship with a business, which is what CAC gets measured against.
- MQL / SQL stands for marketing-qualified lead and sales-qualified lead, the B2B version of the consideration-to-conversion handoff.
- Cohort analysis groups customers by the period they arrived in, so retention gets measured against people who started at the same point.
- Attribution window is the time period after a click or view during which a resulting sale still gets credited to that ad.
- Churn rate is the share of customers who leave or cancel in a given period, the direct opposite of retention.
- s2s postback is a server-to-server signal confirming a sale actually happened, used by tracking platforms to close the loop between click and revenue.
| Funnel stage | AARRR equivalent | Where they differ |
|---|---|---|
| Awareness | Acquisition | Same idea in both models: how someone first finds the brand. |
| Consideration | Activation | AARRR narrows this to one specific, trackable in-product event rather than general interest. |
| Conversion | Revenue | AARRR treats revenue as its own tracked stage; the funnel folds it into a single conversion-rate metric. |
| Retention | Retention | Matches directly in both models - the stage that determines whether CAC ever pays back. |
| No direct equivalent | Referral | AARRR gives word-of-mouth its own stage; the classic funnel usually treats it as a side-effect of retention. |
09FAQ
What's the difference between a marketing funnel and a sales funnel?
A sales funnel is the back half of the marketing funnel - it usually starts once a lead reaches a salesperson and ends at closed-won or closed-lost. The marketing funnel spans the whole journey, from the first ad impression through repeat purchase, across every channel a buyer touches first.
How many stages should a marketing funnel have?
Four works for most businesses: awareness, consideration, conversion, retention. More granular versions split consideration into MQL and SQL, or adopt AARRR's five stages - useful only when each added stage maps to an action someone on the team actually tracks and acts on.
What's a good conversion rate between funnel stages?
It depends heavily on the stage and the business model, so treat the ranges in this article as a starting point for comparison. The more useful benchmark is a funnel's own history: compare this month's stage-to-stage rate against last quarter's before reaching for an industry number.
Does the funnel model still work when buyers don't move in a straight line?
Buyers loop, skip stages, and vanish for months before returning, so a straight line rarely describes what actually happens. The funnel survives as a model anyway because every transition between stages still needs its own number attached to it, and that's what makes a leak visible in the first place.
What tools actually track a marketing funnel?
GA4's funnel exploration report, the built-in funnel views inside Meta Ads Manager and Google Ads, a CRM's deal-stage pipeline, and - for paid traffic specifically - a tracking stack like Keitaro or Binom feeding stage data back through s2s postbacks.
- A marketing funnel splits the buyer's path into stages - awareness, consideration, conversion, retention - and each stage runs on its own dedicated metric.
- Conversion rates between stages fall in predictable ranges (2%-5% visitor-to-lead, 15%-30% qualified-to-customer, 20%-50% customer-to-repeat); treat them as a baseline for comparison.
- Most funnels leak at a single seam rather than everywhere at once - find the stage sitting furthest below its normal range before spending more on traffic at the top.
- AARRR (pirate metrics) slices the same journey differently and suits product-led teams tracking in-app activation; neither model is more correct, they answer different reporting habits.
- A tracking stack like Keitaro or Binom, and s2s postbacks, are what actually confirm which stage a paid click made it to - without that, funnel leaks stay invisible.
Want this on your product?
I consult on acquisition, funnels and retention - including hard verticals.
