Paid versus free registration and what each one does to turnout

The free expo badge went live two editions ago. Registrations went from 14,200 to 18,600, which is growth of 31.0 per cent and the largest single jump in the show's history. It is still quoted in the sales deck, and it settles the paid versus free registration argument every time that argument comes up.

Verified attendance over the same two editions went from 10,700 to 11,150. That is 4.2 per cent. The 4,400 additional registrations produced 450 additional people on the floor, a turnout rate on the increment of 10.2 per cent.

Both numbers are true and only one of them is in the deck. What follows is how to get the second in front of the people making the pricing decision while the decision is still open.

What does price actually move?

Price moves three things and they are usually collapsed into a single argument about attendance.

Turnout is the share of registrations that become a verified body on site. Qualification is the share of the file that matches the audience your exhibitors bought. Cancellation is the share that actively withdraws before the show. Price pushes turnout up, pushes qualification up, and makes cancellation visible where it was previously invisible.

Keeping them separate matters because a free tier can improve one while destroying another, and whichever one is in the report wins the meeting. Report all three by registration type every week and the argument becomes about evidence. Doing that across every type at once, and what it does to the headline number, is the type mix report and A31's subject.

Turnout needs a definition before it needs a number, and the one I would use is badge collection: a registrant counts once a badge has physically been picked up on site. Collection is a single act with a timestamp, it happens once per person, and it does not depend on anyone remembering to switch a scanner on at a hall door. Use that, or use a first entry scan if you have one, and use the same rule for both populations.

Hold geography constant or measure nothing

The raw comparison between paid delegates and free expo visitors is worthless, because the two populations differ in ways that have nothing to do with price.

Paid delegates travel further. They have flights and hotel nights already sunk, an employer who approved a budget line, and a diary blocked out for three days. Free expo visitors skew local, register later, and are frequently deciding on the morning. If you compare turnout across those two groups you are measuring travel commitment, seniority and lead time as much as you are measuring price.

The minimum control that costs nothing is geography. Restrict both populations to registrations from the same catchment, defined by postcode or by drive time from the venue, and compare inside it. The paid delegates who live forty minutes away have no sunk travel cost either, which removes the largest confound in one step.

Lead time is worth controlling second. Registrations created in the final fortnight behave differently from registrations created in October regardless of what they cost, so either band both populations by days out and compare band to band, or report the median lead time for each group next to its turnout rate so a reader can see how different they are.

The same edition, one metro area

Take one edition and restrict to registrations with a postcode inside ninety minutes of the venue.

Inside that catchment the file holds 6,400 free expo badges and 1,900 paid delegate passes.

Of the 6,400 free badges, 3,968 collected a badge on site. That is 62.0 per cent.

Of the 1,900 paid passes, 1,691 collected a badge. That is 89.0 per cent.

Twenty seven points, on a population matched for the single biggest confound, in the same week, under the same weather and the same transport conditions. That is about as clean as an organiser gets without running an experiment.

Two things to hold about that 27 points before it goes anywhere near a slide. It is one edition, so run it on the two before as well and see whether the gap is stable. And the denominator you pick for a no show rate changes the number materially, which is its own argument and A21's, so state the denominator next to the figure or expect to be corrected by somebody using a different one.

What is the gap worth in bodies?

The gap is only interesting if you convert it into the decision it informs, which is whether to charge for the expo badge.

Suppose you put 95 units on the free tier. Turnout rises to the paid level of 89 per cent, which is generous but let it stand. The question is how many registrations you lose.

Work the break-even in bodies. You currently produce 3,968 people from that catchment. At 89 per cent turnout you need 3,968 divided by 0.89, which is 4,458 registrations, to produce the same number of bodies. Going from 6,400 to 4,458 is a fall of 1,942, or 30.3 per cent.

So the rule is: if charging 95 units cuts local expo registrations by less than 30.3 per cent, the floor gets busier and you collect 4,458 times 95, which is 423,510 in revenue you did not have. If it cuts them by more than 30.3 per cent, the floor gets quieter and the revenue is compensation for a worse show.

That single percentage is what the whole decision turns on, and it is the one number your own file cannot give you, because you have never charged. What your file can give you is the elasticity you have already observed elsewhere: what happened to volume when you last moved the delegate price, which is what a price ladder built deliberately exists to expose and A13's subject, and what the exhibitor invitation codes do at their various rates. Both are weak evidence and both are better than the argument the meeting is currently having.

My own preference is a nominal charge rather than a full one. Twenty five or thirty units is far below the 30.3 per cent volume loss that would hurt, and my reading is that most of the commitment effect comes from the act of paying rather than from the amount. A free tier and a thirty unit tier are different products in the registrant's head. A thirty unit tier and a ninety five unit tier are the same product at two prices.

Cancellation is not comparable and gets compared anyway

The third measure breaks in a way that catches people out every year.

A paid registrant who cannot come has a reason to tell you, because there is a refund, a transfer or a substitution available. A free registrant has no reason to tell you anything, so they never cancel. They simply do not appear.

This means your cancellation rate is structurally near zero on the free side and materially above zero on the paid side, and the naive reading is that the paid population is less reliable. The paid cancellations are the visible part of an attrition the free tier is also experiencing and never reporting, so the comparison inverts the truth.

Report cancellations only within the paid population, and never as a comparison across types. If you want one number that covers both, it is turnout, because a no show and a cancellation land in the same place on the floor.

Substitution deserves a line of its own, because it is the mechanism that keeps paid turnout high. CONEXPO-CON/AGG's 2026 registration policy allows name substitutions at no additional cost through 17 February 2026, on badges that are otherwise non-refundable. A company that has paid for eight badges and can move the names around will send eight people. The same company holding eight free badges will send whoever happens to be free that week.

Shows that charge everybody

It is worth knowing that the all-paid model is alive at large scale rather than a historical curiosity.

CONEXPO-CON/AGG 2026 sets its attendee and non-industry guest badge at 289 US dollars and states that every registrant eighteen and older must pay the admission fee. There is no free expo tier. The only complimentary badge in the policy is for children seventeen and under, who must be accompanied by an adult at all times on the show floor. A show with three million square feet of exhibit space charges every adult who walks in.

The NACS Show 2026 rate card does the same thing at a smaller price and with more gradations. There is no free buyer badge on it at all: the cheapest way for a buyer to get in is a one day registration at 255 US dollars at the member early bird rate, or 505 for a non-member, rising to 555 and 805 on site. Everything free on that show's floor is a booth personnel badge or a comp, and both sit under a written entitlement rather than a public price.

That is worth knowing mainly because it makes the free expo badge visible as a choice your organisation took at some point, and choices can be reopened. Which free badges your own show issues, to whom, and under whose signature is a policy question and A17's.

Where this stops

Matching on geography removes the largest confound and leaves the one that cannot be removed, which is self-selection. The people who chose to pay were different people before they paid. A local buyer who spends 595 on a conference pass has already decided the show matters to them, and their 89 per cent turnout is partly a measure of that decision rather than of the money.

That limit is fundamental. No observational cut of your own file separates the effect of the price from the effect of the person who was willing to pay it, and the only clean answer is a randomised price test that almost no organiser will run for good reasons. Treat the 27 point gap as an upper bound on what charging would buy you.

The second limit is that turnout is not the outcome your exhibitors care about. They care about qualified buyers with budget in their aisle. A paid tier that raises turnout by moving your audience towards larger employers with training budgets may deliver fewer bodies and a better floor, or fewer bodies and a narrower one, and turnout cannot tell those apart.

This week, take last edition's file into whatever attendee analytics tooling you have, filter to registrations inside a fixed drive time of the venue, and compute badge collection rates for your paid and free types separately inside that filter. If the gap is under ten points, price is not doing what this post assumes and your free tier is cheaper than it looks.