The webinar funnel: every stage, end to end.

A webinar funnel is eight jobs, and the session is one of them.

The most common failure in this whole subject is treating the session as the funnel. A genuinely good hour can sit inside a machine that produces no revenue, because seven other stages decide who arrives, what they do next, and whether any of it is traceable afterward.

Every stage below exists whether or not you built it deliberately. A funnel with no reminder sequence still has a confirm-and-remind stage — it is simply performing badly and silently. Naming the eight makes the silent ones visible, which is most of the value of a map.

Two properties of the order matter. First, each stage hands the next one an audience it has already filtered, so downstream rates are computed on a population upstream decisions selected — cheap traffic that converts well on the page will wreck the show-up rate, and the reminder sequence gets blamed. Second, the stages fail independently. Fixing the page does nothing for a checkout problem, and no amount of follow-up rescues a session nobody stayed for.

This page is the map. Every stage below is somebody else’s guide in full; the job here is to say what the stage does, which number judges it, and where to go for the detail. The same shape holds for other funnels — a lead magnet funnel swaps the session for a download and keeps the rest, and customer journey mapping is the same exercise across every funnel you run at once.

Where the room comes from, tagged before the first link goes out.

The job of this stage is not volume. It is to put the right people in front of the registration page and to know, afterward, which decision produced them.

Sources are not interchangeable. A partner send, your own list, a personal invitation and a cold ad produce audiences with different intent, and that difference survives all the way down the funnel — it shows up as show-up rate, as stay rate, and eventually as refunds. Treating traffic as a single undifferentiated number hides the one variable most likely to explain the rest of your results.

Which is why the operational instruction for this stage is boring and non-negotiable: one tracked link per channel and per partner, built before anything is sent. Untagged traffic cannot be tagged retroactively. Every run without source tracking is a run you can never compare anything to.

Order the channels by what they return for a funnel that has not proven itself yet: partners, your list, direct invitations, owned surfaces, and paid last — paid being the only channel that charges you for finding out that the funnel does not convert.

One job: make the trade obvious.

This page converts attention into a contactable person. It is the only asset every source depends on, which makes it the highest-leverage thing to get right and the worst thing to leave until the week of.

The page has to answer four questions without scrolling: what will I be able to do afterward, when exactly is it in my timezone, who is teaching and why is that credible, and is there a replay. Being coy about the replay costs more registrations than it saves attendance. Burying the date costs both.

Ask for as little as you can defend. Name and email is usually enough, and every extra field is a tax you pay in registrations for data you might never use. If a field exists because a sales process needs it, keep it and accept the cost knowingly — that is a different decision from keeping it because the form template had it.

Judge this stage per source as well as overall. A page that converts at one rate for your list and a very different rate for cold traffic is not one page with one number; it is a message-match problem wearing a conversion rate.

The stage that decides how much of the room turns up.

Between signing up and starting, a registrant’s intent decays. This stage exists to slow that decay and to remove every practical reason not to join.

Two of the sends do most of the work. The confirmation is opened more than anything else you will send, so it carries the join link, the calendar file and one sentence restating what they will be able to do afterward — not logistics alone. The short reminder shortly before the start is the highest-yield message in the sequence and the one most often left as a draft rather than a schedule.

Between them, a useful send a few days out proves the hour is worth spending, and a friction-removal send the day before answers the two questions that cause no-shows: when exactly, and will there be a replay.

One structural note that belongs to this stage rather than to promotion: anyone who registers should stop receiving invitations immediately and start receiving this sequence. Two clocks, two audiences. On a list of any size that has to happen automatically.

Teach something real, then make the offer without a swerve.

The session’s job inside the funnel is narrow: move a room from “interesting” to “I need to do this and I would rather not do it alone.” Everything else it does is a bonus.

That job is why the offer is chosen before the topic. Work backwards — name what you are selling, ask what someone would have to believe to buy it, and teach that belief well enough that a reasonable person reaches the conclusion honestly. The teaching is real and the conclusion is earned; that is the whole mechanism, and it is also why sessions built topic-first produce a transition that feels like a swerve.

Say up front that there is an offer at the end. It costs nothing and removes the suspicion that otherwise shapes the next forty minutes. Then stay until the questions genuinely stop — the last ten minutes hold the people closest to buying.

Engagement during the session is data, not decoration. Polls, chat and the question queue are the signals the next stage branches on, so whether they are useful on Tuesday depends entirely on whether they land somewhere your email system can read.

The stage most funnels skip, and the one that changes the arithmetic.

The moment the session ends, one audience becomes five, and treating them as one is the single most expensive simplification in this whole machine.

Five segments cover nearly every session: the people who bought, the people who stayed through the offer and did not buy, the people who left before the offer, the people who left early, and the no-shows. They want completely different next messages. A buyer who receives a discount email an hour after paying learns something about you that no follow-up recovers. A no-show who gets the attendee email is being congratulated for something they did not see.

Start with two segments rather than five — attended and did not attend. A coarse split that actually runs beats a fine one that exists on a whiteboard, and you can add branches once the coarse one is working.

The wiring constraint is the whole story here. The split can only branch on behavior the system sending the emails can read. Where attendance lives in one product and the sequence lives in another, the split becomes a manual export at nine in the evening — and nobody sorts a registrant list by hand twice.

Two numbers, because there are two ways to lose the sale.

Between “that is interesting” and money arriving there are two distinct events: the click and the completed purchase. They fail for unrelated reasons and have to be counted separately.

A low click rate against a high stay rate is an offer problem: people heard it and declined. A healthy click rate against a low completion rate is a checkout problem: price surprise, required fields, a payment option you do not offer, a flow that breaks on a phone. Reading those two as one number sends you to rebuild the wrong thing.

The mechanics of the offer moment are simple and mostly about removing ambiguity. Say what it is, what happens immediately after purchase, the price, the deadline, and put the link in the chat while saying it out loud. Then hold the deadline you announced — extending it teaches your audience that your deadlines are decorative, and you only get to teach that once.

Test the whole path with a real transaction before anyone else sees it, including on a phone. This is also the natural home for an order bump or a post-purchase offer — but add those only once the base path completes reliably, because they multiply whatever the checkout is already doing.

If the offer is high-ticket enough that a conversation belongs in the middle, the checkout becomes a booked call and the funnel grows a stage: the webinar-to-call funnel covers who earns the invitation and what the booking page has to do.

The window that decides what the session was worth.

The session created the intent. This stage converts it or wastes it, and the difference is almost entirely whether the split in the previous stage actually happened.

Send within a couple of hours, while the session is still the most interesting thing that happened to them today. Give each segment the message its behavior earned: the people who stayed through the offer get the specific objection the questions surfaced; the people who left before it get the offer section rather than the whole replay; the no-shows get the replay with a real deadline and then a separate, longer path.

Then work the deadline properly. Most of the revenue in a well-run funnel arrives after the session ends, which is exactly why the post-session share is worth counting — a funnel with a decent live conversion and nothing afterward is leaving the whole window unworked.

Buyers leave the sales sequence immediately and enter onboarding. Nothing reads worse than a “last chance” email arriving after somebody has paid, and it is the cheapest mistake on this page to prevent.

Nine numbers, defined precisely, compared only to yourself.

Measurement is not a stage that happens at the end. It is a stage that has to be wired at the beginning and read at the end, which is why it is the one most often missing when it is needed.

Definitions travel; numbers do not. Published funnel benchmarks average across industries, price points and traffic sources that have nothing to do with yours, and borrowing one produces a target with false precision attached. What makes these numbers useful is that you compute them the same way every run, so your fourth funnel can be compared to your first.

Revenue per registrant is the one to optimise against. Every other rate here can be improved by making the funnel smaller — invite fewer people, admit only your warmest list, and watch six percentages rise while the business gets worse. Revenue per registrant does not move that way, and it is also the number that tells you what a registration is worth paying for when you eventually buy traffic.

One honesty rule attaches to this stage. Attributed conversion requires the registration, the attendance and the purchase to sit on one record. Where they do not, the number is an estimate — still useful, provided it is labelled as one and not compared against a number that was measured differently.

When one of the nine is wrong, the work is diagnosis rather than redesign — read the stages top down, stop at the first one out of line with your own last run, and change one thing there. Funnel diagnostics is that procedure in full, including how much data you need before a difference means anything.

If you are starting from nothing, build it in this order.

The map has eight stages. The first version does not need eight polished ones — it needs the ones that make the rest measurable, in an order that fails cheaply.

1. 01The offer and a working checkout. Before anything else, because everything upstream is built to point at it. Buy it yourself with a real card and refund the transaction. A funnel that fills a room and then fails at payment is a worse outcome than one that never filled the room.
2. 02The registration page. The first asset a stranger touches, and the only one every source depends on. Build it once properly and it is reusable for every session afterward.
3. 03The confirmation email and one reminder. Two sends, not eight. The confirmation carries the join link and gets opened more than anything else you send; a short reminder shortly before the start is the single highest-yield message in the sequence. The rest of the sequence is an improvement, not a prerequisite.
4. 04The session itself, run live. Not recorded, not automated. You need to hear the questions and find out which part loses the room before you commit any of it to a machine that repeats it.
5. 05A two-way split. Attended and did not attend. Two segments that exist beats five segments you plan to build. Add the finer branches once the coarse one is running.
6. 06Tagged links, from the first send. Untagged traffic cannot be tagged retroactively, so this is the one item that is genuinely cheaper to do now than later — ten minutes at the start against a permanently unreadable first run.
7. 07The rest of the reminder sequence, then automation. In that order. Automation repeats whatever you built, including the parts that were not working, so it comes after a session that has held up across a few live runs and not before.

What that order optimises for is finding out early. Everything expensive — the full sequence, the fine-grained segments, the automation, the paid traffic — comes after you have watched a real audience move through the machine once. A funnel that has sold something can be improved with confidence; a funnel that has not is a set of assumptions with a schedule attached.

Once a session holds up across a few live runs, the same eight stages can run on a clock instead of a calendar. Automated webinar setup covers what changes when nobody is in the room, and the evergreen playbook covers keeping the deadlines honest once they are automatic.