Your webinar analytics show page views, registrations, attendance, clicks and sales. Reading off the numbers takes a minute. Understanding them takes longer. A figure that went up might be good news, a measuring glitch or pure chance. This article shows how to read your webinar analytics so you end up with a decision you can justify, not just a good or bad feeling.
Before you open the report
Settle three things before you look at a single number.
The question. “How’s it going?” isn’t something numbers can answer. Try “Since the new reminder email, are more registrants turning up?” or “Where are we losing the most people?” A clear question tells you which figures matter and which you can safely ignore.
The period. Decide on it before you see the results. If you look first and pick the dates afterwards, you’ll nearly always find what you hoped to see.
The changes. Keep a simple log of what you changed and when: a new headline on the registration page, different session times, a new slide at the offer, an ad campaign started. Without it, you’ll be guessing later why a line bends in week three.
Read the figures as a chain
The individual figures depend on each other. Each step builds on the one before it: people who never see the registration page can’t register, and people who don’t attend can’t click the offer. So a single figure says little on its own. Its neighbours tell you what’s going on.
Some typical patterns:
- More registrations, same attendance. The extra sign-ups probably come from a source with weaker interest, such as a giveaway or a very general post. Your sign-up rate rises while your show-up rate falls.
- Same attendance, fewer clicks on the offer. Something changed on the way to the offer. Check whether people leave earlier or whether the offer is now presented differently.
- Lots of clicks, few sales. The problem sits after the webinar, on the checkout page. Does the price there match the one in the talk? Does payment work on a phone?
- Everything drops evenly. Fewer people are reaching the registration page. The cause lies with your promotion or a traffic source that has dried up, not with the webinar itself.
For how the stages of a sales webinar fit together, see Sales webinar structure from welcome to offer.
Periods that can be compared
A comparison is only as good as the periods you put side by side. Webinars come with a few traps.
Uneven weeks. A week with a bank holiday, the start of school holidays or a clock change behaves differently from a normal week. Compare whole weeks where you can and mark the exceptions. Holidays and time zones are covered in How to plan an automated webinar schedule.
Different numbers of sessions. If you ran six sessions a week in June and four in May, raw totals aren’t comparable. Use rates or figures per session.
Late purchases. Many people don’t buy during the webinar but hours or days later. If you review last week on Monday morning, the weekend’s sales may still be missing. Give each period a fixed follow-up window before you judge it.
Cohort or calendar. Where possible, assign attendance and purchases to the period in which people registered, not the day they happened. Someone who signs up on 28 June and buys on 2 July belongs to the June group. That way you compare groups that saw the same version of your page and your webinar.
Rolling averages. Daily figures jump about. A rolling average over seven or 28 days smooths out the noise and shows the trend.
Small numbers and chance
With few events, chance plays a big part. An example with made-up figures: in May you have four sales from 80 attendees, in June six from 80. That looks like a 50 per cent rise. In reality the difference is small enough that it could easily have happened without any change at all.
A simple rule of thumb helps. When you count events that happen independently and fairly rarely, the count fluctuates by chance by roughly its square root. For four sales that’s about two either way, for 100 registrations about ten, for 400 about twenty. If a difference sits inside that range, it proves nothing.
Three habits follow from this:
- Look at counts, not just percentages. “7.5 per cent purchase rate” sounds precise. “Six sales” shows more honestly how thin the evidence is.
- Combine data until it holds up. Look at four weeks together rather than at each session on its own.
- Check whether a single event flips the result. If one sale more or less turns “better” into “worse”, you don’t have a result yet.
From symptom to cause
Analytics show you where something happens, rarely why. For the why, you need hunches that you then test on purpose. This overview is a starting point:
| What you see | Possible causes | What to check |
|---|---|---|
| Show-up rate drops suddenly | Emails going to spam, broken joining link, session moved | A test registration of your own, delivery of reminders, the time in the email |
| Show-up rate declines slowly | Too long between sign-up and session, weaker traffic source | Gap to the next session, where registrations come from |
| Viewing time drops at one point | A slide bores or confuses, sound or picture stutters | Watch and listen to that point yourself, on a phone too |
| Fewer clicks on the offer | Offer comes later, button doesn’t appear, webinar got longer | Running order, timing of the offer, length |
| Clicks without sales | Price doesn’t match the checkout page, payment fails | Go through the checkout yourself |
| Sign-up rate falls | New traffic source, page loads slowly, form changed | Where visitors come from, load time, recent edits |
Some of these points have their own articles. Reminders are covered in Webinar reminder emails, the page itself in How to build a webinar landing page that converts, and when people expect a buy button in When to show the buy button in a webinar.
Numbers don’t always explain the why. So read what guests write in the chat, and ask now and then. A single question after the session, such as “What was missing for you?”, often explains more than another table.
Spotting measurement errors
Sometimes the problem isn’t the webinar but the measurement. These errors are common:
- Your own tests. Test registrations by you and your team distort small numbers. Use a recognisable address and filter it out.
- Automated sign-ups. Bots that fill in forms create registrations that are never confirmed. A jump in registrations without a jump in confirmed registrations is a clue.
- Duplicate people. Someone who registers twice with the same address or visits the page several times may count more than once, depending on the tool.
- Changed definitions. If an update changed what counts as attendance, figures before and after aren’t comparable.
- Time zones. If the report uses a different time zone from your sessions, events just before midnight slip into the wrong day.
If a figure looks too good or too bad to be true, check the measurement first and the webinar second.
A one-page review
To turn reading into a decision, a fixed format helps. One page is enough:
- Period and comparison period, with a follow-up window for late purchases.
- Changes made during that time, from your log.
- The chain from page views to sales, with counts and rates.
- The standout point, where most has changed or most people are lost.
- One hunch about why.
- One next step that tests exactly that hunch, and the date you’ll look again.
If you want to test the hunch properly against a second version, a split test is the next step. Until then, the rule of changing only one thing at a time will do.
Frequently asked questions
How often should I look at my analytics?
A quick weekly look is enough to catch failures such as broken links or emails that aren’t arriving. For real decisions, use longer periods, such as a month or a fixed number of sessions.
Which figure matters most?
That depends on your question. For sales webinars it pays to keep the whole chain in view and look for the step where you lose the most in proportion. A single favourite figure easily leads you astray.
What if I only have a few attendees?
Combine longer periods and pay more attention to patterns than to percentages. With very small numbers, feedback from guests and your own observations often tell you more than any rate.
Why do different tools show different numbers?
Because they count differently: visits or people, with or without repeats, in different time zones. Pick one source and stick with it.
Read, suspect, test
Webinar analytics aren’t a report card. They’re a map that shows you where to look. Read the figures as a chain, compare only periods that are comparable, don’t take small differences too seriously and test each hunch with a single change. That way every month adds a little experience you can build on.
