A founder showed me a marketing slide earlier this year. One number on it, set in huge type: 412 registrations. Their best webinar ever. I asked what it produced in pipeline and the room went quiet. Two days later I got the answer: three opportunities, and one of those was already open before the webinar and would have closed anyway.
That gap is the normal state of things. I have looked at maybe thirty webinar programs across client CRMs and the pattern almost never changes. Marketing reports registrations. Sales reports that webinar leads are junk. Finance sees a line item for the platform and the promotion spend and cannot tie it to anything. Everybody is a bit right, which is why the argument never ends.
The webinar is rarely the problem. The problem is the 40 metres of plumbing between the webinar platform and the CRM, and the fact that nobody owns it.
Registrations is the number that matters least
Here is the chain that actually happens: someone registers, some fraction shows up, some smaller fraction stays long enough to absorb anything, a smaller fraction again does something that signals intent, and then sales either follows up in a way that respects that signal or blasts everyone with the same recording email.
Registrations sit at the top of that chain and tell you almost nothing except that your promotion worked.
The attendance benchmarks are worth knowing because they are wider apart than most teams assume. ON24's 2026 Digital Engagement Benchmarks report a 60% registration-to-attendance rate across their platform, an average of 239 attendees per webinar, and live viewers staying about 51 minutes. Independent cross-platform B2B benchmarks are less flattering, with a median live attend rate closer to 41.6%. Both can be true. ON24 skews toward enterprise marketing teams with warm house lists. If you are buying cold registrations off paid social, the low end is your reality.
The number from that report I care about most is different: on-demand viewing accounts for roughly half of all attendance, and those viewers watch about 33 minutes rather than 51. Half your audience shows up after the event is over, watches two thirds as much, and in most CRMs I open, none of them exist as a tracked action at all.
A webinar is a qualification event, not a lead source
This is the reframe that changes what you build.
Most teams treat webinars as top of funnel. Fill the form, capture the email, count the lead. Then they are surprised when the list converts badly and sales stops working it.
Webinars are the opposite. They are one of the few things you run where a stranger voluntarily gives you 30 to 50 minutes of attention and tells you, through what they do in those minutes, how serious they are. The Content Marketing Institute's B2B research keeps putting webinars near the top of mid-funnel formats, with 78% of marketers rating them effective. That rating only makes sense in the middle, where somebody already has a problem and is shopping for an approach.
The conversion data points the same way. Across B2B, Salesforce's State of Sales work put MQL to SQL conversion around 13%. Webinar-sourced contacts are generally reported somewhere between 17% and 30% depending on whose dataset you read. The spread is wide and I would not build a forecast on it, but the direction is consistent across every client CRM I have queried myself: webinar contacts convert better than gated PDF contacts, and they convert much better when the follow-up knows what they did.
So the job is not to generate more registrations. The job is to capture behavior and act on it.
The four fields your CRM needs, and the three it usually lacks
When I audit a webinar program I run one query. I look for four properties on the contact record:
Almost every company has the first one. Maybe a third have the second as a crude boolean. Fewer than one in five have minutes watched sitting on the contact record in a form sales can filter on. Actions taken are usually trapped in a CSV export that somebody downloads once and never looks at again.
Without fields three and four, no amount of clever follow-up copy will help, because you have no way to tell the person who asked a pricing question at minute 38 apart from the person who opened the room and walked away.
Where the plumbing breaks
Four failure points, in the order I usually find them.
The registration form asks for the wrong things. Company size and use case dropdowns look useful and mostly generate garbage, because people lie to get into a webinar. Work email plus company name plus job title is enough. Everything else you should enrich, not ask for. Run the list through an enrichment step before it hits the CRM so you get firmographics you can trust rather than self-reported noise. This is the same discipline that makes CRM data enrichment worth the spend anywhere else.
Registrations create contacts with no account. People register with personal addresses more than you expect, particularly for anything with a compliance or career angle. Those records land as orphans and never get associated with the target account, so the ABM report shows zero engagement from an account where three people attended. Fixing lead-to-account matching once solves this for every channel, not just webinars.
Engagement data never syncs back. The native Zoom and HubSpot connectors will usually write "attended" and stop. Minutes watched and in-session actions sit in the platform's own reporting. If your platform exposes them via API, a small n8n workflow that pulls the post-event report and writes four properties onto the contact takes an afternoon and pays for itself the first quarter.
On-demand views are invisible. The replay page is often ungated or sits on a separate tool, so the 50% of your audience who watch afterwards produce no record at all. Put the replay behind the same tracking as the live session, even if you keep it ungated for known contacts.
You are not measuring a webinar. You are measuring minutes of attention per account.
Once engagement lands on the contact and rolls up to the account, the webinar stops being a marketing campaign and starts being a signal source your sales team can work like any other.
Segment into three tiers, not one list
The most expensive mistake in this whole area costs nothing to fix: sending one follow-up email to everyone who registered.
Tier 1 is usually defined as attended live for more than half the session, or took an in-session action, or matches your ICP and watched on demand for more than 20 minutes. In a 400-person webinar that is typically 30 to 60 people. Those go to a rep with the actual context in the task: which session, how long they stayed, what they asked.
Tier 2 gets a sequence about the specific topic, not a generic newsletter. If the webinar was about renewal forecasting, the follow-up talks about renewal forecasting for the next three touches. Treat it as lead nurturing with a known subject, which is far easier than nurturing a cold list.
Tier 3 gets the replay and nothing else until they do something. No-shows are not worthless, they are just unproven. Sending them to a rep is how you teach sales that webinar leads are junk.
On timing, there is vendor research claiming that following up within an hour converts at 53% to SQL versus 17% after 24 hours. I do not believe those exact figures and neither should you, because the people who follow up fast are also the people who do everything else well. The direction is right though, and it matches what I see when I compare speed-to-lead cohorts inside client CRMs. Same-day for Tier 1, next-morning at the latest. The broader case for that sits in our write-up on speed to lead and routing rules.
The build, end to end
The stack we use most often is HubSpot as the system of record, whatever webinar platform the client already pays for, Clay for enrichment on the registration list, and n8n for the sync and scoring logic that the native integrations will not do. None of that is exotic. The reason it does not exist at most companies is that it belongs to nobody: marketing owns the webinar, sales owns the follow-up, and the wiring in between is an orphan. That is the gap our CRM and RevOps work exists to close, and the automation side is usually a week of build, not a quarter.
How to report webinar pipeline without lying
This is where most programs lose their budget, so be careful here.
First-touch attribution will make webinars look like a miracle, because a lot of people register before they ever talk to sales. Last-touch will make them look worthless, because almost nobody signs a contract straight off a webinar. Both are wrong, and if you pick whichever flatters the program you will get caught the first time a CFO looks closely.
Use an influence rule instead, and write it down before the quarter starts. The one I use: a deal is webinar-influenced if any contact on the deal attended a session within 90 days before the deal was created, or attended while the deal was open. Report two numbers side by side. Influenced pipeline, and influenced pipeline where the webinar was the first recorded touch. The second number is your honest sourcing claim. The first is your honest influence claim. Do not average them into one figure.
Healthy B2B SaaS companies tend to see marketing source somewhere between 28% and 40% of pipeline in total. Webinars are one channel inside that. If your webinar program is claiming to source half your pipeline, your attribution is broken, not your program. There is more on picking a model you can defend in our piece on why multi-touch attribution models lie.
Things I would cut
A few opinions that tend to start arguments.
Run fewer webinars. Monthly is usually the right cadence for a company under 100 people, and most teams doing weekly are producing content nobody watches to keep a calendar slot full. The average attendee count falls off a cliff when you over-produce.
Stop gating the replay for known contacts. You already have their email. Making them fill a form again to watch something they registered for costs you the on-demand half of your audience, which is the half you were not measuring anyway.
Kill the internal product webinar. If your speakers are all employees and the topic is your roadmap, that is a customer update, not demand generation. Send it to customers and stop promoting it as a pipeline play.
Run partner webinars instead. Two audiences, half the promotion cost, and the list you get is usually better qualified than anything you buy. It is one of the cheapest additions to a go-to-market plan that most teams already have written down. The mechanics of the follow-up are identical to what you do with trade show leads, and the same rule applies: agree who works which names before the event, not after.
The scoreboard
Four numbers on one dashboard, reviewed monthly. Not twelve.
- Attendance rate, live and on demand, split out. This tells you whether promotion or the topic is the problem.
- Tier 1 count per webinar. The absolute number of people worth a sales touch. This is the real output.
- Tier 1 to meeting rate. If it is under 15%, your tier definition is too loose or the follow-up is generic.
- Influenced pipeline per webinar, 90-day window. The number you take to the board.
Registrations can stay on the slide if it makes marketing happy. Just put it in small type at the bottom where it belongs.
Webinar data stuck in the platform?
Book a free 30-minute audit. We will look at what your CRM actually records after an event and show you the three fixes we would make first.
Book an audit →FAQ
What is a good webinar attendance rate for B2B?
Between 40% and 60% of registrants, depending on how warm your list is. ON24 reports 60% across their platform, while independent cross-platform benchmarks put the median near 41.6%. If you promote to your own database and customers, aim for the higher end. If you buy registrations through paid social, 40% is a realistic ceiling and there is nothing wrong with that as long as your cost per Tier 1 contact still works.
How many webinar registrants should go to sales?
Far fewer than most teams send. In a 400-registrant webinar, 30 to 60 people usually meet a sensible Tier 1 bar: attended more than half the session, or took an in-session action, or matched ICP and watched a meaningful chunk on demand. Sending the full list destroys rep trust in the channel within about two events, and you do not get that trust back easily.
Should we score webinar attendance in our lead scoring model?
Yes, but as behavior, not as fit. Minutes watched and in-session actions belong in the engagement half of the score, and they should decay after 60 to 90 days. A webinar someone watched last March says very little about them today. We cover the split in detail in the guide to fit and engagement scoring.
How do we prove webinars generate pipeline?
Define an influence window before the quarter starts and stick to it. Ours is 90 days before deal creation, plus any attendance while the deal is open. Report influenced pipeline and first-touch-sourced pipeline as two separate numbers rather than blending them. The discipline matters more than the exact window, because a rule you change after the fact is not a measurement.
Do no-show registrants have any value?
They do, but not as sales-ready leads. Half of all webinar attendance happens on demand, so a good share of no-shows will watch the replay if you send it. Track that view the same way you track live attendance, and let it move them into Tier 2 automatically. What kills the channel is routing no-shows to a rep on the day, which trains sales to ignore everything with a webinar tag on it.
Fix the plumbing, not the presentation
The uncomfortable thing about webinar programs is that the content is usually fine. The speakers are decent, the topic is relevant, and people show up. What is missing is four fields on a contact record, one scoring rule, and an agreement about who works which names.
If you are running webinars and cannot tell me the influenced pipeline from the last three, that is a two-week fix, not a strategy problem. Get in touch and we will show you what your own data already says.