Should You Offer the Webinar Replay? Nobody Has Tested It

The replay is the recording of a live webinar, sent afterwards - usually to everyone who registered, whether they attended or not. The replay question is whether promising that recording is what stops people attending live, and therefore whether you should keep sending it, restrict it, or stop. It is the decision that follows directly from the most-cited reason registrants give for skipping a live session, and it is the decision almost nobody answers.

This page is the sequel to a ranking. In why webinar registrants do not show up we put the replay at the top of the causes, because that is where registrants themselves put it. Establishing the cause and then never addressing the policy it implies is the gap this page exists to close. We are not going to restate the ranking; we are going to take the decision seriously.

Read the source label on every one of those. It is vendor platform data, from 2021, from a US and global English-language syndicated network, and the report itself does not disclose how many webinars are behind it - only "thousands of data points" from its platform and its own user survey. The same report prints three different numbers for the live figure: 36% on its benchmark page, 34% on the audience-conversion chart, and 33% in the prose. We publish the chart figure, disclose the other two, and would rather you knew the benchmark disagrees with itself than pretend it does not.

Has anyone actually tested whether the replay reduces live attendance?

No. We looked specifically for this and found nothing: no academic study, no vendor report that varies replay policy and measures the result, no operator publishing a split test with a disclosed sample. What exists is a large volume of consistent, unmeasured testimony pointing one way, and a large volume of confident advice built on top of it.

That sentence is the whole problem in fourteen words, and it is worth being precise about what kind of thing it is. It is one person, saying what they do, in a public discussion, agreed with by many others in the same thread. It is the highest-voted explanation in a corpus that contained a great many explanations. That makes it strong evidence of what people say and no evidence at all of what happens to a show up rate when a business changes its policy.

The asymmetry is the interesting part. The replay is the strongest reported cause of no-shows and the weakest evidenced one, while the cause with the best evidence behind it - reminder timing - is near the bottom of any honest ranking. The subject has its research effort pointed at the wrong end of its own problem, and it has been that way long enough that a category of vendors now sells against the least important cause.

What do you gain by not sending the replay?

The argument for removing the replay is a scarcity argument, and it is a coherent one. If attending live is the only way to get the material, the choice the registrant faces changes shape: it stops being "now or later at no cost" and becomes "now or not at all". That is a real change in the decision, and it is the reason the recommendation keeps resurfacing.

What is claimed to follow, in the order operators claim it:

• Live attendance rises, because the alternative has been priced out of existence. This is the headline claim and it is the untested one.
• The room is warmer. People who chose to be there rather than defaulting into a recording are, on average, more engaged - which matters if you sell from the session.
• The event becomes an event again. A session that cannot be caught up on has a deadline, and deadlines are the only thing a calendar respects.
• Your reminders start meaning something. A reminder for something that will be available anyway is a notification. A reminder for something that expires is information.

What do you lose by not sending the replay?

Here the evidence is better, because at least one large dataset speaks to the size of the on-demand audience. In the BrightTALK Benchmarks Report 2021, registration-to-view conversion was 34% live, 15% on-demand, 46% in total. Roughly one in three of the registrants who watched anything at all watched only the recording. Remove the replay and, in that population, you are removing about a third of your total viewing to try to grow the other two thirds.

Label it properly before you use it: vendor platform data, 2021, US and global English-language, a syndicated content network rather than a house list, no webinar count disclosed, and three internally inconsistent figures for the live number. It is not your business. It is the closest published measurement of the thing you would be giving up.

And the people behind the percentage are not interchangeable with the people you are trying to convert. Removing the replay costs you, specifically:

• Everyone who genuinely could not make the time. A client emergency, a school run, a timezone, a shift. These are not people avoiding you, and punishing them is a strange way to build a list.
• The international part of your audience. One live slot cannot be convenient in three timezones, and no reminder solves a clock.
• The internal share. B2B recordings get forwarded to a colleague, a manager, a procurement contact. That forward is often where the deal actually starts, and it cannot happen without a link.
• The searchable asset. A recording keeps earning after the session. Withdrawing it converts a durable asset into a one-day event to buy an unmeasured attendance gain.
• Goodwill you will spend later. "I registered and got nothing" is a bad first impression, and it is the impression you create for the people your live slot did not fit.

What sits between "always send it" and "never send it"?

Almost the entire argument is conducted between two extremes that few businesses actually need. The useful moves are in the middle, and they are all cheap to implement. None of them has been tested either - but a policy you chose beats a platform default you never noticed.

Three implementation notes that decide whether any of these work:

• Say the policy on the registration page. Whatever you choose, the registrant should learn it before they sign up, not after they miss the session. A replay withdrawn by surprise reads as a bait and switch; a replay that was never promised reads as a rule.
• Enforce the expiry. An expiring replay that gets re-sent to whoever complains is not an expiring replay - it is a full replay with an extra email, and your list learns that within two events.
• Cut the offer, not the teaching. If the session sells something, the version that removes the pitch and the live Q&A gives away the least while withdrawing the most. It is the least resented of the restrictions because nobody feels short-changed on the substance.

Is there a way to beat the replay without removing it?

One idea in the entire corpus claims to have done this, and it is a format change rather than a marketing one.

The logic is that the replay only competes with a lecture. If the live session is a lecture, the recording is a strictly better version of it - same content, at 2x speed, when convenient. If the session is something you do, the recording is a video of other people doing it, and that is not the same product at all. The replay stops being a substitute without anyone having to withhold it.

What makes a session genuinely un-replayable, in ascending order of effort:

• They bring something. A page, a number, a draft, a screenshot - submitted at registration and reviewed in the room.
• Their case gets answered. Questions taken on specifics, not on the topic. The recording answers somebody else's question; live answers theirs.
• They finish something. Time inside the session to build, write, or configure the thing, with help available while they do it.
• Something is decided in the room. A choice made, a plan committed to, a template filled in - an output that leaves with them.
• A resource is limited. Hot seats, review slots, a headcount cap on a breakout. Real scarcity, honestly bounded.

It also has a property the other options do not: it is the only move that raises live attendance without taking anything away from anyone. The ranked list of everything else that plausibly moves a show up rate is in how to increase webinar attendance, including the lever that is larger than any of this and that most vendors leave out because they do not sell it.

What if you run evergreen or automated webinars?

Then most of this page does not apply to you, and it is worth saying so plainly rather than letting you read to the end and work it out. Evergreen is a different product, not a setting on the same one. The session is already a recording; the registrant already knows it; there is no live moment to protect and no replay to withhold. The replay question, as a question, mostly dissolves.

Those four numbers come from one practitioner who discloses a basis - 114 webinars built and optimised over ten years - but no methodology, no measurement period per figure, and no per-figure sample. We include them because they are the only course-creator-specific figures in the literature, and we would rather show you a labelled weak number than a confident invented one. Do not put them next to a platform benchmark and treat the difference as a finding.

Two structural things follow for anyone running evergreen. The gap between registering and the session starting is frequently minutes, not days, which means most pre-event sequences have nowhere to happen. And the thing that decides whether an evergreen funnel works is the stay rate and the offer, not attendance mechanics.

How would you test the replay policy on your own list?

Since nobody has tested this, the only real answer available is: you test it, on your own list, and you become the first person you know of with an actual number. That is less satisfying than a benchmark and considerably more useful, because your traffic source and your offer are the two variables the missing research would not have controlled for anyway.

The practical design is to alternate the policy across your next several events, holding everything else still:

• Same traffic source and same mix. If one event runs on paid traffic and the next on your house list, you have measured the traffic, not the replay. This is the mistake that ruins most operator tests.
• Same offer, same session, same length. Change one thing.
• Same day of week and time slot. A Tuesday event and a Friday event are not comparable, and neither is a session that moved by three hours.
• Alternate rather than switch. Replay, no replay, replay, no replay - so a seasonal drift affects both arms roughly equally instead of landing entirely on one.
• Several events, not two. A single pair tells you about two Tuesdays. Run enough of them that a difference has to survive being repeated.
• Announce the policy at registration, both times. Otherwise you are testing surprise, not scarcity.

Alternating across events is not a clean randomised trial, and it is worth knowing why. Time is confounded with the treatment: your list gets more saturated, your ads get more expensive, the season moves. The cleaner design is to split one event's registrant list in half and send the replay to only one half - same day, same session, same everything - but that means treating two groups of customers differently for the same event, which many operators will not do and some are right not to do. If you are willing, it answers the question far faster. If you are not, alternate, and accept a noisier answer honestly obtained.

One more note on what a result would mean. Whatever you find applies to your list, your offer and your audience, and to nothing else. That is not a limitation of the method - it is the actual state of knowledge in this subject, which is why we have been careful all the way down this page not to hand you someone else's number dressed as an answer.

So should you offer the replay?

Our position, stated as a position rather than a finding, because a finding does not exist:

1. Do not kill the replay to fix attendance. You would be trading a measured audience - roughly a third of total viewing in the one dataset that speaks to it - for an attendance gain nobody has ever measured. That is a bad trade to make on testimony alone.
2. Do make it a decision. The default is "send everything to everyone forever", and almost nobody chose it. Choosing an expiring window, or an attendee-only rule, or a cut without the offer costs an afternoon.
3. Change what live is for, first. The working session is the only reported way to beat the replay without taking anything away, and it is the change that would still be worth making if the replay question were settled tomorrow in either direction.
4. Announce whatever you decide at registration. Half the resentment in this subject comes from the policy changing after the sign-up, not from the policy itself.
5. Test it, and count registrations as well as attendance. Several events, alternating, one variable, four numbers.
6. Do not buy a reminder cadence to solve this one. A reminder that says the session starts in an hour is arguing with somebody who has already priced the alternative at zero. That is a different problem with a different fix, and we have written about which is which.

If you want the full picture of what does move a show up rate - and the honest ranking of how strong the evidence is for each item - start with why webinar registrants do not show up for the diagnosis and webinar attendance statistics: every study for what the published numbers really measure. If you want the system rather than the reading, that is AI webinar attendance, and the part that usually matters more than the empty seat - the hundreds of registrants nobody speaks to while the offer window is open - is post-webinar follow-up calls.

Frequently Asked Questions