Short answer: roughly one in five registered guests does not turn up, and the number climbs when the invitation is free. PheedLoop's Event Data Lab Report #05, published on 29 April 2026, analysed check-in data from more than 1,070 live events and found a median no-show rate of 20 percent, with a mean of 29 percent. Free events sat at a median of 28 percent, paid events at 17 percent. Corporate invitations are almost always on the free side, so planning the room, the catering and the transfers with a 25 to 30 percent shortfall built in is closer to reality than planning for a full list. That rate is not fixed. How the invitation is framed, how reminders are scheduled and how registration is managed all move it. Below: the real distribution of no-shows, why free invitations leak, the late registration myth, a forecasting method, how to negotiate guarantee numbers with a venue, and what to do in the final two weeks.
What does the no-show distribution actually look like?
The average on its own misleads. In the same report the median sits at 20 percent while the mean reaches 29 percent, which tells you a cluster of events loses close to half the list. The spread: 21 percent of events kept no-shows under 10 percent, and 49 percent stayed under 20 percent. Size matters too. Events expecting 10 to 49 people had a median no-show rate of 32 percent, 50 to 149 people 23 percent, 150 to 499 people 19 percent, and anything above 500 landed in the 18 to 19 percent band. A small gathering feels like the safe one. The data says the opposite, and the reason is visible on site: small invitations usually go out informally, and skipping them costs nobody anything. A dealer meeting where the company pays for flights, transfers and hotel rooms loses far fewer people, because there the cost of not showing up is in plain sight.
Why do free invitations produce more empty chairs?
This is the sharpest split in the data. Free events show a median no-show rate of 28 percent against 17 percent for paid ones, and that ten point gap repeats at every size. The tails are worse: 22 percent of free events lose more than half their expected attendance, against 14 percent of paid events. The instinctive explanation is registration friction, the idea that an easy form attracts casual sign-ups. Report #06 in the same series, published 26 May 2026 across more than 860 events, tested that and did not find it. Among paid events, no-show rates stay flat between 16 and 19 percent no matter how easy or hard the form is. Commitment comes from the payment, not the friction. You usually cannot charge your own guests, but you can borrow the mechanism. Skift Meetings reported a clean example on 2 September 2026: 37signals founder Jason Fried collected 100 dollars at registration for a Chicago breakfast and handed the same 100 dollars back at the door. The earlier free version drew 84 registrations and 38 attendees. The deposit version drew 55 registrations and 50 attendees. Fewer sign-ups, more people in the room.
Are late registrants the ones who skip?
The common belief is that anyone registering in the final week was never serious. Pooled data appears to confirm it. PheedLoop's Report #09, published 10 September 2026, covered 216,282 registrations across 467 live events: people who registered during event week had a 42.4 percent no-show rate, while those registering six or more months out sat at 18.4 percent. That 24 point gap disappears once you compare early and late registrants inside the same event. The median within-event difference is 0.6 points, and at 44 percent of events the late registrants actually showed up more reliably. Late guests are not flakier; events that collect late registrations are different events. Free events register a median of 25 days later than paid ones, and that is where the loss really comes from. The operational takeaway is simple: closing registration early does not lift attendance, it just shrinks your room. Keep the list open, accept final-week sign-ups, and build badge printing around that.
How do you forecast attendance?
The calculation has three factors: how many people you invite, how many of them register, and how many registrants walk through the door. Take the second and third from your own history rather than an industry average. As a reference point, Bizzabo's 2026 State of Events Benchmark Report, published 5 March 2026, found that mature multi-event organisations on its platform average 412 registrations and 269 attendees per event, a 52 percent registration-to-attendance rate. That figure blends formats, so read it as a warning rather than a target. Building your own ratio takes one spreadsheet: invitations sent, registrations received and door check-ins for your last three events, side by side. Do not treat the list as one block either. Dealers, employees, customers, press and students answer the same invitation at very different rates. Four more variables move the forecast: distance and traffic, day of week and start time, weather, and who the invitation comes from. An invitation signed by the managing director does not perform like one sent from an agency mailbox.
How should you negotiate guarantee numbers with the venue?
Three lines in the venue contract turn your forecast straight into an invoice: the guaranteed number, the final-numbers deadline, and how many extra covers the venue will set above the guarantee. The first is the billing floor, the number you pay for even if nobody arrives. The second is your last chance to revise downwards, usually a few working days before the event, after which reductions are not accepted. The third is missing from most contracts and survives only as a custom, which is worth nothing in a billing dispute. The workable setup: guarantee below your forecast, keep an upward revision clause open until the event day, and write the number of extra covers into the contract as a figure. For a residential dealer meeting the same logic applies to the room block. Put the release date in your calendar, because unreleased rooms can cost more than the missing guests ever would.
Where does the operation break when the forecast is wrong?
Under-attendance hurts the budget, over-attendance hurts the brand. On the high side, seating breaks first: with no spare chairs in the room, the guest left standing remembers the unpreparedness rather than the stage. Badges break second. A registration desk working from pre-printed badges pushes anyone outside the list out of the queue, so keep an on-site printer at the desk. Catering is third. Producing extra covers takes tens of minutes rather than a few, so agree the surge capacity with the kitchen in advance. Transport is fourth, since shuttle capacity is fixed and adding a coach on the day rarely works. On the low side, the one fix that saves the room is the seating plan. If 300 people arrive in a 500-seat room, close the back blocks, pull the chairs forward and move to a tighter layout around the stage. To be able to make that call on the morning, agree a re-set window with the venue at contract stage.
Not everyone in the building attends the sessions
Conference and congress formats carry a second layer of loss. PheedLoop's Report #08, published 3 September 2026, analysed 967,701 session check-ins across 411 events. Attendees who entered at least one session attended a median of 5.5 sessions, with the interquartile range running from 3.3 to 8.0. The striking figure is the other one: about one in four people who check into the event never checks into a single session, which puts median session reach at 75 percent of the people in the building. Three consequences follow. Size breakout rooms against realistic session attendance rather than total registrations. Report door numbers and session numbers separately to sponsors and to your own measurement, because they are not the same metric. And treat the foyer, the exhibition area and the networking block as the place where a large share of attendee time is actually spent, not as a gap in the programme.
Seven moves that lift attendance, and the final two weeks
One: personalise the invitation and send it under a named executive. A bulk mail from a corporate address is the weakest channel you have. Two: attach a calendar file to the confirmation, because an invitation that takes no space in the calendar loses to whatever meeting does. Three: remind three times, one week out, two days out and on the morning. Four: put transport, parking and valet details inside the reminder. In a city like Istanbul those three lines often decide the answer. Five: open a waiting list when capacity fills and backfill cancellations the same day. Six: confirm VIPs and speakers by phone, since a thirty-call round prevents a visible gap on stage. Seven: do not close registration early. Keep consent and privacy notices in the invitation flow from the start for any SMS or email reminders. In the final fortnight, work in this order: refresh the forecast segment by segment, submit the guaranteed number to the venue, build badge and check-in staffing around the final count, and put the spare-seat and extra-cover scenarios in writing. Tales Event is based in Istanbul and delivers dealer meetings, launches, openings, gala nights and congresses across Türkiye, with stage, sound, lighting, LED screens, decor and creative design handled by one team. Send us the size and profile of your invitation list and we will build the attendance forecast alongside the room, catering and setup plan.
Let's talk about your event
Tell us your idea and we will prepare the concept and quote.