Benchmark · November 2026 · Mobile

We tried to buy two tickets at 8 of the world's most visited museums. On a phone.

Same task everywhere: two adult tickets for tomorrow, stopping just before payment. Every step captured, scored on seven criteria weighted by the strength of the evidence behind them, and the weakest flows rebuilt as working prototypes.

3/8show the entry price before any choice
3median taps before a price appears
1requires an account to book
€4.6Ma year for every extra 1% of online bookings, combined
Start here

Four fixes with the strongest evidence

Recommendations first, the reasoning underneath. Numbers in circles link to the sources at the end of this report.

Fix 1

Show the entry price before asking anything else

Extra costs are the most cited reason to abandon a checkout, and a total that is not visible up front is cited by 12%. 12

Applies to: Musée d'Orsay, Rijksmuseum, Vatican Museums, Uffizi Galleries

Fix 2

Open the shop in the visitor's language

Three in four online shoppers prefer their own language, and 40% never buy from a site in another language. 5

Applies to: Musée d'Orsay, Vatican Museums, Uffizi Galleries

Fix 3

Let visitors book without an account

A required account is cited by 18% of shoppers who abandon a checkout. 3

Applies to: Uffizi Galleries

Fix 4

Keep a running total visible while choosing

Visitors should always see the state of their order. Not seeing the total up front is a documented reason to abandon. 26

Applies to: British Museum, Rijksmuseum, Uffizi Galleries

Most quantitative evidence comes from e-commerce checkout research. It is the strongest available, but it does not prove the same effect size for museums. That is what an ab-test on your own flow measures.

Ranking

Scores on seven criteria

MuseumScorePrice visible before choices
25 · Strong
Arrives in the visitor's language
20 · Strong
Running total while choosing
15 · Moderate
No account needed
15 · Strong
No unnecessary in-between steps
10 · Moderate
Controls not covered
10 · Principle
Relevant ticket types for an international visitor
5 · Weak
/100
1The Met
New York
2520151510105100
2MoMA
New York
202015151010595
3Van Gogh Museum
Amsterdam
2520815100583
4British Museum
London
2520015510075
5Musée d'Orsay
Paris
6101515010258
6Rijksmuseum
Amsterdam
0200151010055
7Vatican Museums
Rome
00815510543
8Uffizi Galleries
Florence
01000010222
How we score

Weighted by evidence, not by opinion

Criteria with stronger evidence carry more weight. Issues without a statistical or proven source are not scored as flaws; they appear as questions for the museum team.

CriterionWeightEvidenceHow it is scoredSources
Price visible before choices25StrongTaps until the entry price is visible: 0 = 25, 1 = 20, 2-3 = 12, 4-5 = 6, 6+ = 0. A price range without exact amounts is capped at 15.12910
Arrives in the visitor's language20StrongShop opens in English when coming from the English site = 20, lands in another language but switchable = 10, needs 3+ taps to switch = 0.5
Running total while choosing15ModerateTotal visible and updating during ticket choice = 15, only on a later step = 8, none before checkout = 0.26
No account needed15StrongTickets can be added without logging in = 15, log-in required = 0.3
No unnecessary in-between steps10ModerateScreens that add no decision (extra "Details" levels, redundant lists): 0 = 10, 1 = 5, 2+ = 0.4
Controls not covered10PrincipleNo floating element covers ticket or date controls in the screenshots = 10, covered = 0.6
Relevant ticket types for an international visitor5WeakShare of shown ticket types an international adult or family can use: 75%+ = 5, 50-74% = 2, below 50% = 0. Low weight: the evidence that more options lowers conversion is weak.12

One mobile session per museum (375 × 812) on 11 November 2026, from the official ticket page to the step before personal details. The Louvre and Prado ticket shops block automated visitors; we respect that and left them out.

Museum by museum

Every step, every finding

Scroll the screenshots sideways. Findings carry a source; everything else is a question, because a choice that looks odd from outside can be deliberate.

#1 · New York

The Met

100/100
Taps to first price 0Ticket types shown 4Visitors 6.0M (2025)13
met: prices upfront
prices upfront
met: date shortcuts
date shortcuts
met: postcode
postcode
met: types
types
met: subtotal
subtotal

What works

  • All prices in the first screen of text, before any choice
  • Sticky subtotal and Next from the first second
  • Quick date buttons: Today, Tomorrow, Another date
  • Location, date and tickets on one scrolling page with numbered sections
  • Next stays disabled until the order is valid

Findings with evidence

  • No issues that meet the evidence bar.

Questions for the team

  • The pay-what-you-wish rules for New York residents fill the first screen for everyone. Could non-residents see a shorter version?
  • A postcode box sits above the ticket counters. Could it move below them for visitors who do not need it?

Each extra 1% of online bookings is worth about €1.2M a year at USD 30 per adult ticket.

Open the concept on your phone
#2 · New York

MoMA

95/100
Taps to first price 1Ticket types shown 6Visitors 2.7M (2024)13
moma: calendar
calendar
moma: prices
prices
moma: total
total
moma: passes
passes

What works

  • Three-step progress shown from the start (Tickets, Info and payment, Receipt)
  • Prices next to every ticket type right after picking a date
  • Sticky order bar with live total and Continue
  • Passes and special cases moved below as info cards, not as extra counters
  • No time slot needed, which removes a whole step

Findings with evidence

  • No issues that meet the evidence bar.

Questions for the team

  • A sticky announcement takes about a seventh of the screen on every step. Does it need to stay visible during booking?
  • Long eligibility text sits under each ticket name. Would an info toggle keep the rows compact?

Each extra 1% of online bookings is worth about €513k a year at USD 30 per adult ticket.

Open the concept on your phone
#3 · Amsterdam

Van Gogh Museum

83/100
Taps to first price 0Ticket types shown 4Visitors 1.8M13
vangogh: cookie
cookie
vangogh: tickets
tickets
vangogh: two adults
two adults
vangogh: date
date
vangogh: sold out today
sold out today
vangogh: slots with total
slots with total

What works

  • Prices on the very first screen, under 18 free shown as EUR 0.00
  • Seven languages one tap away
  • Time-slot grid shows the order total on every slot
  • Extra ticket types collapsed behind 'Show more tickets'
  • Clear rules up front: non-refundable, valid only for date and time

Findings with evidence

  • No total on the ticket step; it first appears on the time slots 2
  • "From €25" without saying what makes the price higher 1
  • Today is pre-selected and turns out to be sold out only below the calendar 6
  • Dashed slot borders carry a meaning that is not explained 8
  • Only the active step label is readable in the five-step progress bar 6

Questions for the team

  • Floating chat and Next buttons overlap the ticket dropdowns on mobile. Is that known?

Each extra 1% of online bookings is worth about €322k a year at EUR 25 per adult ticket.

Open the concept on your phone
#4 · London

British Museum

75/100
Taps to first price 0Ticket types shown 7Visitors 6.4M (2025)13
british: cookie
cookie
british: visit
visit
british: cookie again
cookie again
british: calendar list
calendar list
british: days
days
british: slots
slots
british: donations first
donations first
british: free below
free below

What works

  • Free entry stated on the first screen of the visit page
  • Sold-out time slots are hidden instead of disabled
  • Clear 'Book now' on the visit page

Findings with evidence

  • About 30 full-width time-slot buttons sit above the ticket choice and push it far below the fold 4
  • Price fields are drawn as input boxes but cannot be edited 7

Questions for the team

  • Four paid donation tiers are listed above the free ticket. Is this to raise donations, and has it been measured against a donation step after booking?
  • Seven ticket types are shown for a free entry. Which ones do most visitors use?
  • Cookie consent is asked again on the ticketing domain. Can the choice be shared between the two domains?
  • Visitors pass a queue page on a second domain. Is the queue needed outside peak times?

Entry is free, so the value of a better flow lies in donations, visit planning and data, not ticket revenue.

Open the redesign on your phone
#5 · Paris

Musée d'Orsay

58/100
Taps to first price 5Ticket types shown 4Visitors 3.8M (2024)13
orsay: french landing
french landing
orsay: home en
home en
orsay: product step1
product step1
orsay: long text
long text
orsay: calendar
calendar
orsay: slots
slots
orsay: price total
price total

What works

  • Clear table with rate, amount and quantity, plus a total row in tickets and euros
  • Calendar legend covers today, available, selected, no slots and sold out
  • Explicit 4-step progress indicator
  • Anti-fraud notice warns about fake ticket sites

Findings with evidence

  • The English ticketing link from the main site lands on the French shop 5
  • About 1,600 characters of conditions come before the calendar; the first price sits below them 2
  • Admission tickets sit two levels deep: a product list, then the product page 4
  • Selecting a time reloads the page and repeats the same text block 4

Questions for the team

  • The fake-website warning takes about a third of the first screen. Would one line keep the same effect?
  • Nine products sit under "Admission tickets". Which ones do most visitors need?
  • Free admission rows show "Sold out" while the text says eligible visitors can queue without booking. Which message should visitors follow?

Each extra 1% of online bookings is worth about €420k a year at EUR 16 per adult ticket.

Open the concept on your phone
#6 · Amsterdam

Rijksmuseum

55/100
Taps to first price 6Ticket types shown 7Visitors 2.5M (2024)13
rijks: landing
landing
rijks: seven types no price
seven types no price
rijks: date
date
rijks: first price
first price

What works

  • Cookie choice with a clear decline option
  • Countdown keeps the basket reserved for 30 minutes
  • Large, tappable steppers and calendar cells

Findings with evidence

  • The first price appears on step 3, after tickets, date and time 2
  • No running total while choosing tickets 2
  • The step indicator is cut off on a 375px screen 6

Questions for the team

  • Seven ticket types include two Dutch lottery cards. How many international visitors use them?
  • The same selling points appear twice above the ticket choice. Is the repetition intended?
  • Free entry under 18 sits behind an info icon. Could it be stated next to the ticket?

Each extra 1% of online bookings is worth about €438k a year at EUR 25 per adult ticket.

Open the redesign on your phone
#7 · Rome

Vatican Museums

43/100
Taps to first price 6Ticket types shown 2Visitors 6.9M (2025)13
vatican: home
home
vatican: language
language
vatican: home en
home en
vatican: calendar
calendar
vatican: sold out
sold out
vatican: results
results
vatican: first price
first price
vatican: fee total
fee total

What works

  • Clear two-tier pricing once reached (full and reduced)
  • Total bar with Proceed button on the ticket step
  • Max visitors per product stated (1-6)

Findings with evidence

  • The ticket shop opens in Italian; switching to English takes 4 taps 5
  • The first price appears on the fourth screen, after date and visitor count 2
  • The €20 ticket price excludes a €5 per-person reservation fee, shown as a separate line 1
  • Visitors are counted twice: a number field first, then dropdowns per ticket type 4
  • Red calendar days have no legend explaining what red means 6

Questions for the team

  • Unavailable products are listed by default. Is that to show the full offer, or would hiding them by default work better?
  • The total shows €10 in fees before any ticket is selected. Is that intended?

Each extra 1% of online bookings is worth about €1.2M a year at EUR 25 per adult ticket.

Open the redesign on your phone
#8 · Florence

Uffizi Galleries

22/100
Taps to first price 6Ticket types shown 5Visitors 2.7M (2024)13
uffizi: cookie
cookie
uffizi: tickets page
tickets page
uffizi: shop italian
shop italian
uffizi: catalog
catalog
uffizi: calendar legend
calendar legend
uffizi: slots
slots
uffizi: login wall
login wall

What works

  • Calendar with a legend for available, sold out, not online and not on sale
  • Time slots show remaining places
  • Clear cap of 10 tickets per order

Findings with evidence

  • Tickets can only be added after logging in to an account 3
  • Coming from the English site, the ticket shop opens in Italian 5
  • Two levels of "Details" before the actual entrance ticket 4
  • The first price appears after date and time are chosen 2
  • Remaining places are bare numbers in brackets, without a label 8

Questions for the team

  • Is the log-in required for a reason, such as named tickets or preventing resale?
  • Group headset rules are shown to individual visitors. Can they move to the group flow?
  • Cookie consent is asked again on the ticket shop domain. Can the choice be shared?

Each extra 1% of online bookings is worth about €555k a year at EUR 29 per adult ticket.

Open the redesign on your phone
Redesign

One ticket kit. Any museum.

One set of components in Figma and one prototype engine, applied with each museum's colours and products. The three flows with the most evidence-backed issues got a deeper redesign; the Rijksmuseum redesign is our earlier case.

Redesigned ticket screens for Vatican Museums, Uffizi and British Museum
Same components, three brands. Built in Figma with a brand mode per museum.
Vatican Museums

Opens in your language, counts visitors once, includes the booking fee in the price.

Try it
Uffizi Galleries

Guest checkout instead of a log-in, English shop, straight to the entrance ticket, places left in words.

Try it
The British Museum

A test variant: the free ticket first and the donation as one optional step afterwards, to compare against the current layout.

Try it
Ticket kit components
The kit: progress, ticket row, date chip, time slot and total bar, each with its states.
Business case

What one percent is worth

We do not predict how much a redesign will improve bookings. We show what each extra percent of online bookings is worth, so the museum can weigh that against the cost of testing.

MuseumVisitors / yr 13Adult ticketPer extra 1% / yr
Vatican Museums6.9MEUR 25€1.2M
The Met6.0MUSD 30€1.2M
Uffizi Galleries2.7MEUR 29€555k
MoMA2.7MUSD 30€513k
Rijksmuseum2.5MEUR 25€438k
Musée d'Orsay3.8MEUR 16€420k
Van Gogh Museum1.8MEUR 25€322k
British Museum6.4MFree entryn/a
Assumption, with source

70% of visitors book online, based on a 2021 survey of over 7,000 people by Tiqets, a ticket seller 11. A vendor survey, so treat it as an indication.

Assumptions, without source

One adult ticket per booking (concessions and free entries make the real value lower), and exchange rates rounded on 11 November 2026. How many percent a redesign adds is not known until it is tested.

How this report was made

Agents for the legwork, a designer for the judgement

An AI agent pipeline walked every flow, captured the screenshots, built the Figma kit and the prototypes, and tested them on an iPhone viewport. The criteria, the evidence rules and the review are by Design Narrative. Findings without a source were turned into questions rather than presented as flaws.

Sources

Where the reasoning comes from

  1. Baymard Institute Statistic (e-commerce)
    40% of shoppers who abandon a checkout cite extra costs (fees, shipping, taxes) as the reason. Source
  2. Baymard Institute Statistic (e-commerce)
    12% abandon because the total cost was not visible or calculable up front. Source
  3. Baymard Institute Statistic (e-commerce)
    18% abandon because the site required creating an account. Source
  4. Baymard Institute Statistic (e-commerce)
    17% abandon because the checkout was too long or complicated. Source
  5. CSA Research, 2020 Statistic (consumer survey)
    76% of online shoppers prefer information in their own language; 40% never buy from sites in other languages (8,709 consumers, 29 countries). Source
  6. Nielsen Norman Group, heuristic 1 Design principle
    Visibility of system status: keep users informed about what is going on. Source
  7. Nielsen Norman Group, heuristic 4 Design principle
    Consistency and standards: elements that look alike should behave alike. Source
  8. Nielsen Norman Group, heuristic 6 Design principle
    Recognition rather than recall: make information visible instead of making users interpret it. Source
  9. EU Unfair Commercial Practices Directive, Art. 7(4)(c) Law (EU)
    An invitation to purchase must state the total price; adding mandatory fees later is likely misleading (Commission guidance). Source
  10. UK DMCC Act 2024 and CMA guidance Law (UK)
    Prices must include all mandatory fees; optional extras such as donations need express consent, no pre-ticked boxes. Source
  11. Tiqets survey, 2021 (via American Alliance of Museums) Vendor survey
    Over 70% of 7,000+ respondents planned to book museum and attraction visits online. Source
  12. Scheibehenne et al., 2010 Meta-analysis
    Meta-analysis of 50 experiments: average 'choice overload' effect close to zero. Source
  13. Wikipedia, list of most-visited art museums Reference
    Annual visitor numbers, latest year per museum. Source
Working together

From benchmark to better bookings

Ticket flow audit

Your flow against this benchmark, on mobile and desktop, with a ranked fix list and sources.

  • Two weeks
  • Scorecard and screenshots
  • Business case on your numbers
Redesign sprint

Audit plus a redesign of your ticket flow in your brand, ready for your developers.

  • Three to four weeks
  • Figma kit and working prototype
  • Ab-test plan
Conversion partner

Ongoing tests across ticketing, memberships and the shop.

  • Monthly test cycle
  • Reporting on bookings, not clicks
  • Handover to your team

Scope and pricing are set per project after a short intake.

Talk to Design Narrative