Andy Garcia · freelance dev
FRENES
Back to the blog
UX & Pricing

Baymard Ran 463 Ticket-Buying Test Sessions. Testers Walked Away When the Fees Came Last.

October 19, 2026 7 min read

On September 24, 2026, Baymard Institute published the first public findings of a new research study on event ticketing sites. The numbers on the launch page: more than 2,400 hours of usability testing, analysis and research, 463 qualitative "participant/site" test sessions following the think-aloud protocol, and more than 2,800 medium-to-severe usability issues, distilled into more than 370 guidelines.

Sixteen sites were tested, in the US and the UK. Eleven marketplaces and resellers: Ticketmaster, StubHub, SeatGeek, AXS, Vivid Seats, TickPick, Gametime, Event Tickets Center, TicketWeb, Eventbrite and Broadway.com. Five venues and performing arts centers: Royal Albert Hall, National Theatre, Barbican, LA Phil and Fox Theatre.

Baymard's summary fits in one sentence: users "struggle with pricing communication and seat-selection transparency, leading to numerous abandonments during our test sessions."

I do not sell tickets. But I ship iOS apps with paywalls, and I built BidForge, an AI that reads French public tender files for small businesses. Both end on a screen where someone decides whether to pay. Ticketing is the extreme version of that screen, and Baymard says why: a live event cannot be returned. Its researcher puts it this way: "experience shopping is confidence-driven and must be designed for hesitation." Many software purchases are closer to that than to buying a phone case.

What was tested, and what is public

Think-aloud means participants say what they are thinking while they use the site. Baymard describes its testing as mainly in-person, one-to-one moderated lab sessions. Its methodology page says participants are told to behave as they would on their own, including abandoning a site for a competitor and going off-site to look for information.

Two details matter before reading the findings. The 463 are participant/site sessions, not people: one participant can test several sites, and the public pages do not say how many individuals took part. And the free article shows three takeaways and nine examples. The 370+ guidelines, of which 37 are rated "Essential", are sold through Baymard's research platform, along with audits. Everything below comes from what anyone can read. I have not seen the guideline texts.

1. The price comes last

At the LA Phil, on Android, a participant found no price in the showtime listings and had to go all the way to the seat selection page to learn the cost. Her question: "why would I hit 'Buy Now' if I don't know how much the tickets are?"

On StubHub, on iOS, a father wanted five tickets to PAW Patrol Live! for his family. To see a price, he had to select a date, enter a quantity in a blocking overlay, wait for the seat map to load and scroll the listings.

The sharpest example is at the Fox Theatre, on iOS, for "& Juliet". A participant selected two seats for a $360 total, then went through several pages, evaluating add-ons and booking parking. At the cart, she found $75 in ticket fees, a $3 parking fee and a $5 "Order Charge". She abandoned the cart to look for tickets elsewhere.

The arithmetic: the fees named in the article add $83 on top of $360 of seats, 23 percent. The ticket fees alone average $37.50 a seat. Her own summary of the fix is five words: "Bake it into the cake."

Baymard's general finding is that withholding mandatory service fees until the cart or order review "triggered intense sticker shock and eroded platform credibility." My reading: a late fee does not only raise the price. It changes the question in the buyer's head, from "is this worth it?" to "what else is this site not telling me?". The participant said it herself: "You know they're just nickeling and diming me."

2. The constraint comes too late

On Ticketmaster, a participant looking for "Nurse John" clicked through to a third event before discovering, on the partner page, that there were "No more tickets". At the Barbican, on iOS, another found a date for "A Midsummer Night's Dream", opened the seat map and saw only a few single seats. She spent almost four minutes repeating this for two performances, then said: "I think at this point I would just look for a different show."

Baymard's point is that sitting together is often non-negotiable for group buyers. If the site does not ask for the number of tickets before the showtime, section or seat pages, and does not flag "Sold Out" early, users load seat map after seat map to find scattered single seats.

Every product has its group of five. On a SaaS plan, it is the number of seats. For a buyer with data constraints, it is the region. For a small business reading public tenders, it is the eligibility conditions. A tender that a company cannot answer is a sold-out show, and the useful moment to say so is before the file is opened, not after.

3. The map does not answer

On TickPick, a participant looking for two tickets to Lakers vs. Heat clicked on section 110 of the zone map. The listings did not filter. She assumed the feature was broken, and with prices higher than on a site she had seen before, she left for a competitor. On AXS, on iOS, a participant looking for five tickets to a classical concert found only a static map and no view from the seat, and left to look for the same ticket on other platforms.

The positive examples are the instructive ones. On Gametime, price pins on the venue map opened a tooltip with a seat-view image: "That really allows me to stay on the page the entire time." At the LA Phil, tapping a seat opened a photograph of the actual view, and the participant bought her tickets.

The software equivalent of a seat view is a preview at the place of decision: a real screenshot of what the export looks like, sample output, a demo.

What this study does not show

The public article does not show how often any of this happens. It gives no frequencies and no conversion rates: we do not know how many of the 463 sessions ended at the cart because of fees. Baymard's methodology says each guideline is rated for severity and for frequency, in bands from "A Few" to "All", and that the two combine into an importance level such as "Essential". Those ratings sit behind the paywall, and a band is not a rate. The examples are illustrations that Baymard chose.

It does not size the effects, and Baymard says so in its own methodology. The aim of its testing "is not to arrive at final statistical conclusions", but to map the breadth of problems. It cites a binomial calculation: with 20 participants, a think-aloud study finds on average 95 percent of the usability problems that occur at a rate of 14 percent or more. The method is built to find problems, not to measure them.

A moderated lab session is not a buyer alone at home, the article does not date the sessions, and Baymard sells the detailed guidelines and audits. None of that makes the findings wrong. It means they are hypotheses for your funnel, not measurements of it.

For tickets, the law got there first

In the US, the Federal Trade Commission's Rule on Unfair or Deceptive Fees covers live-event tickets and short-term lodging. It took effect on May 12, 2025, the date set in the final rule published in the Federal Register on January 10, 2025, and the Federal Register listed no amendment to it as of October 19, 2026. A business that displays a price must disclose the total price, and more prominently than any other pricing information. The total price is the maximum total of all fees or charges a consumer must pay, mandatory add-ons included. Government charges, shipping and optional add-ons may be left out.

Because the article does not date the sessions or say which fees were optional, I cannot tell whether any of these screens predate the rule, and I am not saying any site broke it. What I notice is that the participant at the Fox Theatre asked, in her own words, for what the rule requires.

Outside tickets and lodging, that rule does not apply. The usability problem does.

What I would copy into any booking or quote flow

  1. Show the total at the first price. If a fee is mandatory, it is part of the price. On a paywall, that means the amount billed and the billing period on the first screen, not only a per-week equivalent.
  2. Ask for the constraint before the list. Quantity, seats, region, eligibility: ask once, then hide or mark what cannot satisfy it.
  3. Let people evaluate where they decide. A preview, a real screenshot or a demo on the same page as the price.
  4. Count the screens between the first price a buyer sees and the total they pay. In the Fox Theatre example, that gap held add-ons, parking and three fees.
  5. Then measure on your own traffic, step by step. Baymard hands you the hypotheses. Only your funnel can tell you their size.

A buyer who cannot return what they are buying wants three answers early: what it costs, whether it fits, and what they will get. Every screen that postpones one of them is a screen where they can leave.

Sources

A project like this one?

I design and deploy products like this. Let's talk.

Let's talk