The Consent Banner Decision: How a Travel App's Fix for One Country Broke Another
Archi4 min read
Contents
The Consent Banner That Worked Everywhere Except Where It Mattered#
In early 2024, a mid-sized travel booking app (call it Wanderlane) faced a familiar problem: its consent banner was a mess. Users in Germany saw a different banner than users in France, and both differed from the one shown in Brazil. The legal team had stitched together regional fixes over three years, and each new privacy law added another patch.
The company decided to simplify. One banner, one set of code, served to every user outside the United States. The reasoning was straightforward: a single consent flow is cheaper to maintain and easier to audit. The engineering team estimated three weeks of work. Legal signed off after a two-day review.
Eight months later, a European data protection authority issued a decision that made that choice expensive. The regulator found that the company's banner did not meet the requirement for explicit, informed consent because the "Accept all" button was visually dominant and the "Reject" option sat behind a second click. The fine was modest by headline standards, but the real cost came from the forced redesign and the six months of lost time.
The Decision That Seemed Obvious#
The choice to unify the banner was not reckless. Wanderlane's head of privacy engineering argued that a single codebase would reduce bugs and make it easier to prove compliance across markets. The alternative, keeping separate flows, meant maintaining four versions of the same component, each with its own edge cases. The team had already seen a bug where the French banner failed to load for users on older Android devices, and nobody noticed for weeks.
Unification promised consistency. It also promised something subtler: a single place to make changes when regulators updated their guidance. That promise turned out to be the trap.
What the Regulator Actually Said#
The decision, published in late 2024, focused on the banner's design rather than its underlying data practices. The regulator argued that the visual hierarchy nudged users toward acceptance. The "Accept all" button was green and full-width; the "Reject" link was grey and small, placed below a paragraph of text. The company had tested this design internally and found it increased consent rates by 18 percent. That test result became evidence against them.
The regulator did not claim the company had sold data improperly or ignored deletion requests. The problem was the moment of choice. Users were not given a fair opportunity to say no.
Wanderlane's lawyers had reviewed the banner against the text of the law. They had not reviewed it against the regulator's prior decisions on visual design, which had been accumulating for two years. That gap between legal text and enforcement practice is where the company lost.
The Rebuild, and What It Cost#
The company had to redesign the banner for all users in the affected region. That meant new code, new translations, and a new round of user testing. The engineering team spent five weeks on the rebuild, plus another three weeks on a staged rollout that could be paused if regulators objected again.
The bigger cost was organizational. The privacy team now reviews every user-facing component that touches consent, not just the legal text. That adds two days to each release cycle. The company also hired an outside firm to audit its consent flows in six markets, a process that took four months and produced a 60-page report.
None of these costs appear in the fine. The fine was 0.4 percent of annual revenue in the region, a number the company could absorb. The redesign and the new review process are permanent overhead.
The Lesson for Other Teams#
The failure was not a failure of legal analysis. It was a failure to treat enforcement decisions as part of the requirements. The company's legal team read the statute; they did not read the regulator's published decisions on banner design. Those decisions were public, and they were specific.
A second lesson concerns internal metrics. The 18 percent lift in consent rates was a success by product standards. By regulatory standards, it was a signal that the design might be coercive. Teams that measure consent rates without also measuring rejection rates are measuring only half the story.
What Wanderlane Changed#
The company now maintains a single banner codebase, but it is configurable by region. The default configuration is the most restrictive: equal-sized buttons, no pre-checked boxes, and a clear path to reject. Regional teams can adjust only within limits set by the privacy office.
The company also added a step to its release checklist: for any component that collects consent, the team must cite the specific regulatory decisions it has reviewed. That citation requirement has slowed releases, but it has also caught two potential issues before they shipped.
The decision to unify was not wrong. The decision to unify without reading the enforcement record was. That distinction is the difference between a project that saves money and one that creates a new line item for the legal budget.
More articles