The Inbound Marketing Sales Funnel: Four Stages, and One With No Real Mechanism

The Inbound Marketing Sales Funnel: Four Stages, and One With No Real Mechanism

Draw the inbound funnel on a whiteboard and it comes out the same way every time. Attract at the top, convert below it, close below that, delight at the bottom. Everyone in the room agrees with it, because it is a fair description of what a company would like to happen.

Then somebody asks what actually moves a person from the second stage to the third, and the room goes quiet for a moment before someone offers a word that means keep emailing them. That pause is the interesting part of the model. Three of the four stages name a mechanism you could point at. The second-to-third transition names a hope.

What the model is genuinely for

The funnel is a shared vocabulary for a handoff. Its value is that two teams with different incentives can point at the same object and agree which one of them owns it right now. That is not a small thing, and it is worth keeping.

What it is not is a description of how buyers behave. Buyers loop. They research, go quiet for five months, come back with a different job title, involve four colleagues who have done none of the reading, and arrive at a shortlist assembled from a conversation you were not part of. The people who popularised the funnel largely replaced it with a flywheel for exactly this reason, which is an unusually honest thing for a vendor to do about its own diagram.

So use it as a filing system for handoffs, and stop expecting it to predict anything.

AttractMechanism: search, content, referral

Real, and measurable in sessions

ConvertMechanism: an offer worth an email address

Real, and set by what you offer

CloseMechanism: a human conversation

Real, and owned by sales

DelightMechanism: the product working

Real, and owned by nobody in marketing

The three seams where funnels leak

Every stage boundary is a handoff between owners, and a handoff is where things fall.

Attract to convert leaks on the offer. The traffic is fine and the thing being asked for is not worth an email address, or it is worth one to a population who will never buy. This is a content-and-offer problem and it is the most commonly diagnosed one, largely because it is the easiest to see.

Convert to close leaks on time and on evidence. Somebody identified themselves, and by the time a human engages them the moment has passed, or the human has nothing to go on beyond a name and a downloaded file. The mechanics of closing this seam are in inbound lead generation.

Close to delight leaks on expectation. What was sold and what gets delivered differ slightly, and the gap shows up two quarters later as churn that marketing never hears about and therefore never corrects for.

The stage labels themselves cause a fourth problem, which is that two teams use the same word for different objects. An MQL that marketing counts and an MQL that sales accepts are separate numbers, and the argument about which is real is really an argument about evidence. The version of the definitions that survives that argument is in MQL vs SQL.

The middle of the funnel, where the standard answer is a schedule

Here is the stage with no mechanism, stated plainly. A person converted. They are not ready to talk. Something is supposed to happen over the following weeks that makes them ready.

The conventional answer is a scheduled run of educational emails, spaced out over some number of weeks, that keeps arriving until the person books something or stops opening. Almost every treatment of this funnel recommends it, and the recommendation usually arrives with a diagram of numbered emails and day counts.

Where we differ from standard practice

We never run those, on our own outbound or on our clients'. One message per campaign, built on one premise, sent once, with no scheduled reminder afterwards.

Squeamishness about volume is not the reason. The reason is that a schedule substitutes for having something to say. When a person is silent, the honest interpretation is that the thing you offered was not what they needed that day. A timer does not change that fact and cannot discover what would change it. What it does is put your name repeatedly beside a feeling of low-grade obligation, which is a strange thing to spend brand equity on. What happened to our meeting rate when we stopped is in we stopped using follow-up emails.

What replaces it

The middle of the funnel does have real mechanisms. They are just harder than a schedule, which is why the schedule is popular.

Give them a reason to return rather than a reason to be reminded. A benchmark that updates, a tool they will need again, a number that changes quarterly. The person comes back because the thing is useful, and their return is a genuine signal rather than a manufactured one.

Let the product be the middle of the funnel. A sandbox, a limited free tier, a sample of the actual work. A person who tried the thing has moved themselves through the stage, and you learn more from ten minutes of their usage than from six months of opens.

Watch for a new premise, then write once. If something changes at their company that makes your case newly relevant, that is a reason to write, and it is a different message from the one you sent before. We treat that as a new campaign rather than as the next item on a schedule, because it is a genuinely different argument rather than a repetition of the last one.

Accept that some people leave the funnel. Most converted people are not going to buy this year, and pretending otherwise is what generates the schedule in the first place. A programme that lets them go, and stays findable when they come back, outperforms one that keeps talking.

The scheduled approachWhat most guides recommend

• A timer decides when you speak
• Content is generic, written before you knew who they were
• Silence is treated as a reason to send more
• Measured in opens and clicks
• Requires no new information to keep running

How we work it insteadOne message, one premise

• A new fact about their business decides when you speak
• Content is written after you know why you are writing
• Silence is treated as an answer
• Measured in conversations
• Requires a genuine reason, which is the point

The last stage is the only one that compounds, and marketing leaves the room

Delight sits at the bottom of the diagram, usually gets the least attention in the plan, and is the only stage whose output feeds back into the top of the funnel.

A customer who succeeds tells colleagues, answers questions in the communities where your buyers already are, agrees to be referenced, and shows up in the searches that matter most because they are searching for you by name. None of that appears in a campaign report, and all of it does more for attraction than another content pillar would.

What usually happens instead is that marketing hands off at close and the stage becomes customer success's problem, measured on retention. Retention and advocacy are different outcomes. A quietly renewing customer is retained and generates nothing at the top of the funnel. A customer who will publicly say why they chose you is doing attraction work that no amount of budget replicates.

The practical version is small: know which of your customers would speak on the record, ask them before you need them, and treat that list as a demand generation asset rather than as a sales collateral request. It is the cheapest attraction available and it is the one that requires somebody to have kept a relationship warm, which is why it usually does not exist.

Measuring it without fooling yourself

Three numbers, and one habit.

Stage-to-stage conversion rate, tracked separately. A single top-to-bottom rate hides which seam is leaking. Two adjacent rates tell you where to spend the quarter.

Time in stage. A stage where people sit for months is either mislabelled or has no mechanism, and both are worth knowing. This is usually the number that exposes the middle-of-funnel problem long before anyone admits to it.

Where the eventual customers first arrived, recorded separately from where they converted. These are almost always different pages, and standard attribution hands all the credit to the second one. A page doing the hardest work in the funnel can look like a failure for years.

The habit: count people who left. A funnel that only reports the survivors of each stage will always look reasonable, because a rate computed over the people who made it through is a rate over survivors. Knowing that four hundred people entered the middle stage and eleven emerged is a more useful sentence than any conversion percentage.

Where this funnel does not reach

Every stage of this model assumes the person arrived. An account that has never searched for anything in your category never enters at the top, and no amount of work on the four stages will find them. Deciding which accounts need a different play entirely is the subject of demand creation, capture and conversion, and the honest comparison of the two instruments is in inbound marketing vs outbound.

The short version

The funnel is a good filing system for handoffs and a poor model of buyer behaviour. Three of its four stages name a real mechanism. The one that does not is the middle, and the conventional filler for it is a schedule that runs whether or not you have anything to say.

Replace the timer with reasons: something worth coming back to, a product surface they can try, and a single message written when a real change gives you something specific to write about. Then measure the seams separately, count the people who left, and let the ones who are not buying this year go quietly.

If you want to see what one researched message to twenty named accounts looks like, aimed at people who will never enter this funnel on their own, we will build it: free campaign.