Incrementality Testing for Ticket Sales: Did Your Ads Sell Tickets That Would Have Sold Anyway?

Incrementality testing answers the question every promoter asks about a campaign and almost never gets answered: how many of these tickets would have sold without the ads? The objection behind it is legitimate. For a major artist, 80 percent or more of the tickets come from the artist's own brand, the announcement and the fan base. A platform dashboard that reports a return of 15 on ad spend is taking credit for that brand. The honest question is what happened to the remaining tickets, and whether the campaign moved them. This article explains what incrementality means in ticketing, which test designs are feasible for a promoter who is not Live Nation, how to read the results, and the traps that produce confident wrong answers.

What incrementality means

A ticket sale is incremental if it would not have happened without the marketing activity being measured. Everything else is baseline: tickets the artist, the announcement, the venue's own channels and word of mouth would have sold regardless.

Platform attribution cannot distinguish the two. A fan who was always going to buy, saw an ad on the way, and bought within the attribution window is counted as a conversion. Incrementality testing separates the baseline from the lift by comparing two groups that are alike in every way except one: whether they were exposed to the campaign.

In practical terms, the result of a test is one number: incremental tickets, or the equivalent in incremental revenue, and from that a cost per incremental ticket. That number can be compared across channels, across shows and across seasons, because it does not depend on any platform's attribution model.

Why it matters more in live entertainment

In most categories, the baseline is a steady trickle. In live entertainment, the baseline is the whole story for a large share of shows. An arena headliner sells out in the presale; a campaign running alongside claims thousands of purchases it had nothing to do with. Conversely, a mid-size show in a secondary market has a thin baseline, and a well-targeted campaign can be most of the sales. The same reported ROAS can mean "the ads did nothing" in the first case and "the ads made the show" in the second.

Without a way to separate baseline from lift, a promoter cannot tell which of the two they are looking at. Every budget decision that follows is a guess dressed as a number.

Four test designs that work for promoters

The good news is that live entertainment has a structural advantage for testing: shows are naturally split by market and by date. A promoter running a tour has holdout groups built in.

1. Geo holdout. Run the campaign in some markets and not in comparable others. Compare verified ticket sales per capita, or per capacity, between the two groups over the same window. This is the most robust design for tours and for productions with multiple markets. It requires markets that are similar enough to compare and a promoter willing to leave a few of them without support for a defined period.

2. Time-based holdout. For a single show, switch the campaign off for a defined window and on for another, and compare sales velocity against the baseline curve in each. Weaker than geo tests because demand changes over the life of a show, but usable when there is only one market. The comparison must be made against comparable shows' curves at the same point, not against the previous week.

3. Audience holdout (conversion lift). Meta and other platforms offer lift studies where a randomly selected share of the target audience is held out and never shown the ads. Purchases in the exposed group are compared to purchases in the holdout. This is the direct answer to how to measure incremental conversions in Meta ads: it is the platform's own tool for it. Two caveats: it requires the purchase signal to reach the platform, which depends on the ticketing shop's tracking, and it measures lift within one platform, not across the mix.

4. Matched shows. For promoters with many similar shows, pair shows with comparable attributes (genre, capacity, market size, price) and run the campaign on one of each pair. Over a season this produces a clean read on what the campaign adds for that type of show. It needs enough shows to pair, which makes it a season-level design rather than a show-level one.

For a promoter with forty shows a year, the geo holdout on a tour and the matched-shows design over a season are the two that produce results worth acting on. Both need only the ticketing system's sales data and a willingness to hold something back.

How to read the result

A test returns a lift: exposed markets sold X tickets per 1,000 capacity, holdout markets sold Y. The difference, scaled to the campaign's total reach, is the incremental ticket count. Divide spend by it and the cost per incremental ticket appears.

Three things to check before believing the number:

Was the baseline comparable? If the exposed markets were stronger for this artist to begin with, the lift is overstated. Compare the markets' history on similar shows before the test, not after.

Was the window long enough? Ticket buyers do not convert the day they see the ad. A window that ends too early understates lift; a window that runs to the show date lets late organic demand in. Comparable shows' curves say where the campaign's effect should be visible.

Is the lift stable? One test is a data point. The same design repeated across two or three tours or seasons is a result. Cost per incremental ticket that lands in the same range three times is a number a promoter can plan a budget around.

The traps

Reading platform ROAS as lift. The most common mistake, and the most expensive. Platform-reported purchases include the baseline by construction.

Testing on the show that was going to sell out. A sold-out show cannot show lift; there is no headroom. Tests belong on shows with real audience still deciding.

Switching the campaign off and calling the drop "lift". Sales fall for many reasons in a given week. Without a comparable curve, a drop after switching off proves nothing.

Ignoring cross-channel effects. A TikTok campaign creates searches on Google that convert as branded search. A lift study inside TikTok misses them; a geo holdout catches them. Prefer designs that measure the whole mix against the box office.

Testing once and stopping. A single test on a single tour answers one question about one artist in one season. Incrementality is a discipline, not an event.

Where NYBA stands on this

NYBA does not report platform attribution to promoters. Every campaign is measured against real ticket sales from the connected ticketing system, at ticket level, and read against a baseline built from thousands of comparable shows. What the promoter sees is verified tickets and the movement of the sales curve above what shows like this one sell on their own. Every number can be checked at their own box office.

Structured holdouts on tours and matched-show comparisons across a season are how that baseline is stress-tested. They are also, in our experience, the fastest way to end the argument about whether the artist's brand or the campaign sold the show: the answer is almost always both, in a proportion that changes by show, and the proportion is the number worth knowing.

Frequently asked questions

What does incrementality mean in advertising?
The share of an outcome (here, ticket sales) that would not have happened without the advertising. It is measured by comparing an exposed group with a comparable group that was not exposed, rather than by counting conversions that followed an ad.

How do I measure incremental conversions in Meta ads?
Meta's conversion lift studies hold out a random share of the target audience and compare purchases between exposed and holdout groups. For ticketed events, the purchase signal must reach Meta from the ticketing shop, and the result covers Meta only. A geo holdout measured on the ticketing system's sales covers the whole channel mix.

What is a good cost per incremental ticket?
It depends on ticket price and category, and the honest answer is: lower than the margin on the ticket. A number that stays stable across repeated tests is more valuable than a single impressive one.

Can a small promoter run incrementality tests?
Yes. A tour with several comparable markets, or a season with several comparable shows, provides the holdout structure. The only requirements are sales data from the ticketing system and the discipline to hold something back for a defined period.

What is the difference between incrementality and attribution?
Attribution assigns credit for a sale to a touchpoint. Incrementality asks whether the sale would have happened at all without the touchpoint. Attribution can be computed from a single platform's data; incrementality needs a comparison group and a source of truth outside the platforms.

The other 20 percent is the whole job

The artist's brand sells most of the tickets. Nobody serious disputes that. The job of a campaign is the rest, and the job of measurement is to say how much of the rest it moved, with a number that does not come from the platform being measured.

If you want to know what your campaigns are adding above the baseline on your next tour, book a demo. One event, clean numbers, then you decide.