Email Funnels: How to Map the Stages, Find the Leak and Fix the Right One

Nearly every team that sells anything has drawn a funnel on a whiteboard. Far fewer can say what percentage of people move from the second stage to the third, which is the only part of the drawing that tells you what to do next. The diagram is easy; the measurement is the work.

This guide is about the measurement. What each stage is actually for, when a contact should leave it, which number to watch between any two stages, and how to tell a funnel that needs another email from one that needs a different email. If you are building the individual sequences, Lead Nurturing Emails and Email Marketing Automation go deeper on each one; this is the layer above them.

What an email funnel is, and three things it is not

An email funnel is the path a person takes from first giving you their address to becoming a customer, described as a small number of stages with a measurable step between each pair. It is a model you use to decide what to send and to find out what is not working.

Three common confusions are worth clearing up, because each one leads to a different mistake:

• It is not one automation. An automation is a mechanism, and a funnel usually needs several of them plus some broadcast campaigns. Teams who equate the two end up building one enormous sequence that nobody can safely edit.
• It is not a sales ladder. The stages are not degrees of pushiness. A nurture email that explains one hard thing well does more for the funnel than an email that asks for the sale slightly harder than the last.
• It is not a straight line. People skip stages, re-enter them, and buy from the first email. The funnel is a way to organise your sending, not a claim about how any individual behaves.

The four stages, and the single job of each

Four stages is enough for almost everyone. The discipline is not in having more of them; it is in each one having a single job and a condition that moves the person on.

Notice what is missing: a stage for "send the monthly newsletter". Broadcasts run alongside the funnel and are not part of it. Mixing them in is how teams end up unable to tell whether the nurture stage works, because half the clicks it gets came from an unrelated campaign.

Give every stage an exit condition

This is the part most funnels are missing, and it causes more damage than any other single mistake. If the nurture stage has no exit, a person who booked a demo on Tuesday still receives the "have you considered trying us?" email on Thursday. You look unaware, which is worse than being quiet.

Every stage needs three things written down before you build it:

1. The entry condition. What puts someone into this stage. Usually the exit of the stage before it.
2. The exit condition. What takes them out, including the negative case. "Clicked pricing" and "finished all five emails without clicking" are both exits.
3. The maximum stay. The longest anyone sits here before being moved on regardless. Without this, people accumulate in stages you forgot about.

Measure the handoffs, not the totals

End-to-end conversion is a reporting number. It tells a founder how the quarter went and tells you nothing about what to change. Stage-to-stage conversion is a working number: it names the email that is costing you.

In the example above, the instinct is to blame the final step: only 15% of people who reached pricing bought. But that step is converting a warm, self-selected group, and 15% of them is respectable. The real loss is earlier. Of the 4,100 people who clicked something in the welcome series, 3,120 never clicked again. That is the largest group of interested people in the whole picture, and nothing in the funnel is currently aimed at them.

The fix follows from the number. Rewriting the offer email would improve the smallest step in the chain. Rewriting the first two nurture emails, or segmenting them by what each person clicked in the welcome, addresses the step where thousands of people are being lost. Use clicks for every one of these measurements, not opens: privacy proxies make open rate unusable for anything that branches.

What belongs at each stage

The content question gets easier once each stage has a job. You are not asking "what should email four say?", you are asking "what does this stage need to accomplish, and what is the shortest email that does it?"

Subscribe

A confirmation and the thing you promised. Nothing else. This is not the place for a product tour. If you use double opt-in, the confirmation email has exactly one job and should contain exactly one link. Signup form tactics covers what gets people here in the first place.

Orient

Two or three emails over a week. Who you are, what you will send and how often, and one genuinely useful thing that stands alone. End each with a question or a link worth clicking, because the click is the exit condition and you need it to happen.

Nurture

One idea per email, chosen by what they clicked. Someone who read about deliverability gets the follow-up on authentication; someone who read about segmentation gets the one on behavioural rules. This is where a generic funnel bleeds people and a segmented one does not.

Convert

Two emails, not six. The first makes the offer and names the single most common objection. The second, a few days later, is shorter and gives one piece of proof: a customer who had the same problem, a number, a guarantee. Then stop and return them to the nurture stage or the newsletter.

How long should an email funnel be?

As long as the decision takes, which is a property of what you sell rather than of email. The useful question is how long a buyer typically deliberates, and whether anyone else has to approve the purchase.

• Low price, individual decision (a £30 product, a course): four to six emails over two weeks. Longer and you are padding.
• Mid price, considered purchase (most SaaS): eight to twelve emails over four to six weeks, with the nurture stage branching by interest.
• High price, multiple approvers: the funnel stops being a schedule and becomes a library. Send less often, aim each email at a different person in the room, and expect the sale to happen in a meeting you are not in.

When in doubt, build the shortest version that covers the four stages, measure the handoffs for a month, and add emails only where the numbers show people leaving. Funnels grow fine; they almost never shrink once someone has written the emails.

Five ways funnels quietly fail

• Stages with no exit. The single most common fault. People accumulate, receive irrelevant mail, and unsubscribe at the worst possible moment.
• Escalating urgency instead of escalating relevance. Each email asks harder rather than explaining better. This trains people to ignore you before they ever reach the offer.
• Measuring the whole funnel as one number. It moves slowly, hides the cause, and gives you nothing to act on this week.
• No path for the people who said no. A clear no is valuable. Send them back to the newsletter rather than deleting them or, worse, restarting the sequence.
• Running the funnel on a list that should have been cleaned. Unengaged contacts drag deliverability down for everyone, so the funnel converts worse than its content deserves. Re-engagement belongs before the funnel, not after it.

Building it without it becoming unmaintainable

Build one stage at a time, and let it run long enough to produce a number before you build the next. A funnel assembled in a single afternoon is a funnel nobody has evidence for, and the stage you got wrong will be buried under three more that depend on it.

Keep each stage as its own automation rather than one long flow. Separate automations can be paused, edited and measured independently, and a mistake in the nurture stage does not require touching the welcome series that is working.

Next, go a layer deeper on whichever stage your numbers point at: the welcome series for orient, segmentation for nurture, and metrics for deciding which of these numbers you can trust.