How WMATA replaced fifty years of piecemeal signage with decision points, progressive disclosure, and error-targeted design, and how to run the same method on your own product

Baltimore, Maryland

The method here transfers to any product with a path through it. WMATA broke a station trip into eight decisions, placed information only where each decision happens, tagged every design change with the specific error it should reduce, and tested the system in virtual reality before printing anything. Run those four steps on a checkout flow, an admin console, an onboarding sequence, a warehouse floor, or a store aisle: map the decisions a person makes in order, cut each screen or sign down to what that one decision requires, name the failure each change targets, and measure that failure before and after you ship. Metro supplies the counterexample too. Fifty years of adding information one sign at a time produced a system where 39 percent of riders took a wrong exit in a single year.

My grandfather lived in Washington and I spent childhood summers with him. He drove Chrysler New Yorkers and nothing else, one after the next. Between those drives I rode Metro, and the stations looked like the future to me. They still do. Thomas C. Gale, who ran design at Chrysler, said good design adds value faster than it adds cost.

I do not work for WMATA and have not spoken to anyone there. I ride the system. Everything below comes from WMATA’s published materials, its July 2026 briefing to the Riders’ Advisory Council, and public reporting, listed at the end.

Metro opened in 1976 with five Red Line stations and nothing on the walls.

Harry Weese designed the coffered concrete vaults, and he wanted nothing to interrupt them. Massimo Vignelli, then at Unimark International, agreed to keep strips and panels off the vault face. He proposed a single freestanding object instead. The pylon was a square porcelain enamel column set on the platform at a diagonal, angled to face riders stepping off the train. It held the station name, the loudspeakers, the fire emergency box, the air vents, and the uplights that wash the vault ceiling. Its brown matched the bronze columns outside the entrances. Stripes and dots marked the lines. Helvetica ran up the sides.

The method under the pylon came from New York. In 1966, Unimark partner Bob Noorda walked the subway, traced passenger flow through several stations, and marked every point where a rider had to decide something. Each decision point got the information needed there and nothing else. Vignelli described the Washington answer in the same terms: one element, at the point of need, containing everything required. Unimark codified the system in the 1970 WMATA Manual of Graphic Standards. Lance Wyman drew the map and the icons in 1971.

Riders got lost. WMATA started bolting signs to the walls. Vignelli, in 2010, called the additions pollution.

Fifty years of drift

The system grew to 98 stations. New stations brought new architecture, new sign hardware, and new one-off graphic fixes. By WMATA’s own accounting, graphic solutions arrived piecemeal and hardware variability multiplied. In 2019 a platform contractor installed pylons at rebuilt Blue and Yellow Line stations using Helvetica Black instead of the Helvetica Bold the standards required. Metro replaced them at the vendor’s cost.

The deeper problem sat on the vaults. Track-side signs listed line colors and terminal destinations, and changing a track-side sign requires track rights: a work window, protection crews, trains held or single-tracked. When the Yellow and Silver lines changed service in 2025, the color and destination pairs printed across the system went stale. WMATA put small decals on its paper maps and told riders to trust the digital tools during the transition. The agency’s July 2026 briefing to the Riders’ Advisory Council opens on a slide stating that small service changes can cost tens of millions of dollars.

The rider data looks like this. A Metro survey from January and February 2023 asked about the previous twelve months:

  • 39% had taken the wrong exit at least once
  • 29% had gone to the wrong platform
  • 23% had boarded a train headed the wrong way

In the same period, 62% said Metro needed to do more to help people find their way. A March 2026 analysis of WMATA rider data by the MIT Transit Media Lab found that infrequent riders take longer total trips and miss more transfers. They take suboptimal paths inside stations.

The new system

The first signs went up at L’Enfant Plaza this month. WMATA expects the station finished by the end of October, five more stations within six months, and the full network within five years.

The core engineering decision is a separation. The new system splits the graphic layer from the physical hardware. Signs are assembled from a kit of modular graphic parts and applied to existing sign cases, with minor hardware changes where needed. WMATA scoped hardware out on purpose. Vault wall sign hardware will be standardized over the project so maintenance crews can swap a panel without first surveying what was actually built at that station.

I spend my working hours on systems where content and rendering live in separate layers so one can change without a redeploy of the other. WMATA just applied the same rule to a 98-station physical network. Service information is now data. The vault is the renderer.

Decision points and progressive disclosure

WMATA’s plan breaks a trip into eight steps: recognize the entrance, identify the line, find the accessible route, choose the platform, confirm direction and stations ahead, transfer, select the exit, and continue to the street, a landmark, a bus, or an airport. Signs go at the point where each decision happens. Information appears when that decision is live and stays off the sign otherwise.

This is Noorda’s 1966 method, restated with a UX term. Progressive disclosure comes from interface design, where the rule is to show the primary options first and push secondary ones a layer down. Kevin Lynch coined “wayfinding” in The Image of the City in 1960, and his framework treats a city as paths, edges, districts, nodes, and landmarks that people stitch together into a mental map. A station follows the same logic at smaller scale.

The behavioral basis is the Hick–Hyman law. Reaction time rises with the logarithm of the number of choices presented. A sign with eight messages takes longer to parse than a sign with two, even when six of the eight are irrelevant to the reader. WMATA’s briefing makes the same argument in plain language: more information takes longer to read. John Sweller’s cognitive load theory adds the working-memory side. Extraneous information spends capacity that the task needs.

The briefing tags each feature with the error it targets. Simplified layouts target wrong train, wrong platform, wrong direction, and wrong exit. Those are the four failure categories from the 2023 survey.

Choice architecture on the platform

Richard Thaler and Cass Sunstein define a nudge as a change in how choices are presented that shifts behavior without removing options or changing incentives. Every sign is choice architecture. The new system uses four levers.

Salience on exits. Exit information now appears as red type on a white field. WMATA says aligning with national standard exit signs makes the system safer in an emergency, and that the high contrast lets someone hunting for an exit find it and ignore the rest. Points of interest beneath each lettered exit now sit in a stacked list, which WMATA chose for scannability.

Landmarks over termini. Terminal station names give direction only to someone who already knows where the terminal is. “Ashburn” and “Branch Ave” are geography a visitor has to look up. The new directional helpers pair service information with places a visitor already knows: the airports, the National Mall, Union Station. Usability research calls this recognition over recall. The rider matches a known destination instead of retrieving an unfamiliar one from memory. The helpers have their own distinct graphic treatment so they register at a glance, and WMATA can change the landmark pairing station by station.

A confirmation step before commitment. The 23% wrong-direction figure is an error made at the moment of boarding. The redesigned Station Ahead Maps show every upcoming stop, where the old versions dropped some. They use a destination-first orientation, a stronger You Are Here marker redesigned from rider feedback, and a font size that scales with the hardware and the number of stops. WMATA’s briefing references Stockholm and Tokyo as comparisons. The map gives the rider one last check before the doors close.

Predictable grouping. At stations with multiple lines, rail discs and their terminal destinations now appear in consistent blocks. Each kind of information sits in the same position on every sign across the network. The grouping also means a service change like the 2025 Yellow and Silver adjustments no longer forces thousands of signs to be reprinted.

Accessibility

Several changes target specific impairments.

Color is no longer the only code. The exterior pylons are being redrawn to show rail discs instead of colored stripes. WMATA’s stated reason is that discs are more legible than stripes, especially for color-blind riders. According to the National Eye Institute, red-green color vision deficiency affects roughly 1 in 12 men and 1 in 200 women of Northern European ancestry. A stripe communicates through hue alone. A disc is a discrete shape with a boundary, and the new pylons pair it with entrance identifiers that match the lettered exit system.

Lighter type weight. Rail disc lettering moves from Extra Bold to Medium. Light letters on a dark or saturated background appear to bleed outward, an effect called halation or irradiation, and it gets worse for readers with astigmatism. Heavy strokes close up counters and blur together. Medium weight lets WMATA raise the point size without that bleed. It also ends a quirk where interlined disc pairs alternated between white and black text. The briefing notes Medium sits closer to the weight of the original generation of signs.

Separate visual and tactile characters. The 2010 ADA Standards for Accessible Design, section 703.5, allow an exception where raised tactile characters and visual characters appear separately. WMATA’s new tactile panels use that exception. The braille and raised text stay in place for blind riders, and the panel no longer has to double as a visual sign for everyone else. WMATA’s position is that this lowers visual noise for sighted and low-vision riders without changing anything for blind riders.

Standard character sizes. The modular kit uses standard type sizes from 3 to 6 inches. The briefing says this keeps ADA compliance consistent regardless of sign type or message.

Icons built to a standard. The new icon library follows ISO standards, which go through formal legibility testing. Icons sit on an even background, use a shared scale, and match their line weight to the typeface. Elevators, escalators, accessibility, and connections all get consistent symbols.

Information moved to the platform. Variable service information is coming off the vaults and onto platform signs. WMATA frames this as a maintenance win, and it is. It also puts the information physically closer to the rider. A low-vision rider can walk up to a platform sign. The vault signs hang over the trackbed.

Testing

WMATA says it evaluated the system with customer surveys, in-station paper prototypes, task-based virtual-reality testing, design stress tests, and observation during pilot installs. The VR model is navigable, and the briefing says it saved time and money over legacy testing methods. Task-based testing measures behavior: can a participant find the elevator, board the right direction, pick the right exit. It sidesteps asking people which sign they prefer.

The briefing includes a slide on measuring success across domains. WMATA has not yet published outcome data from the L’Enfant Plaza install.

Open questions

The rollout runs five years across 98 stations. For most of that time, riders will move between stations using two different visual systems, and transfer stations may display both at once during installation. WMATA already tells riders to expect temporary signs and partial replacements.

Progressive disclosure has a failure mode. If information appears only at the decision point, a rider who misses that sign has no fallback downstream. The design depends on sign placement being right at every station, and 98 stations have 98 different geometries.

The briefing lists two project goals in order. Reducing maintenance cost is first. Improving customer experience is second. For most of the twelve features, the same change serves both goals. The survey numbers from 2023 give a baseline, and a repeat of that survey after the first six stations would show whether the wrong-exit, wrong-platform, and wrong-direction rates move.

The redesigned pylons keep Vignelli’s brown enamel and trade the colored stripes for rail discs.