WSDOT Ferries

The 26 moments where 2 million commutes fall apart every month.

Role: UX designer: moderating, notes and affinity synthesis
Team: 4 HCDE 517 researchers at UW
Timeline: Winter to spring 2025

The one-minute version

The problem

Riders weren't failing because features were missing. They were failing because the app spoke the system's language, not theirs.

What we ran

We ran a moderated iOS usability study on the WSDOT Ferries app with 10 Washington residents across 9 task scenarios, from a quick capacity check through to buying a ticket.

My ownership

1Moderating the P9 and P10 sessions
2Notes on the P4 session
3The study plan and the affinity-mapping synthesis
4The interaction map, research plan and recruiting, with the team

Outcome

First
Moderated usability study ever run on the WSDOT Ferries app
26
Friction points documented, clustered into five patterns
2 of 9
Task scenarios failed outright. Both were the highest-stakes ones: buying a ticket, and finding an inter-island route.
5
Prioritised recommendations delivered with the report

The long version, about five minutes

Context

For some riders, this app is the road home.

The WSDOT iOS app serves Washington State ferry riders navigating Puget Sound. We evaluated the Ferries section: Route Schedules, Vessel Watch, Buy Tickets and Reservations. The question was where friction lives between the app's structure and the decisions riders are actually trying to make.

Riders aren't a uniform group. Some are daily commuters; some visit once a year. For island residents, there is no alternative.

2M+

Ferry users the app serves

“Live on Orcas so ferry is our primary travel to the mainland.”
Participant, interest form

Method

We mapped the app before we tested it.

Testing screens tells you which screen is confusing. Mapping every path first tells you which confusions matter, so the nine tasks we wrote put the most consequential routes under real use rather than sampling the interface.

1. Map
Built a complete interaction map of the Ferries section: Route Schedules, Vessel Watch, Buy Tickets, Reservations.
(opens in a new tab)
Interaction map of the WSDOT Ferries sectionOpen
2. Research plan
9 task scenarios across 4 task groups: capacity, infrequent routes, ETA and alerts, and ticketing.
(opens in a new tab)
The study plan documentOpen
(opens in a new tab)
The moderator guideOpen
3. Recruit
10 Washington residents, iPhone owners, all with prior ferry-app experience. Daily commuters through to annual tourists.
The recruitment screener responsesNot linked, to protect participants
4. Moderate
Online sessions with a moderator and a note-taker. Participants on their own iPhones, plus a PC for screen sharing.
(opens in a new tab)
The session sheetOpen
Session notesNot linked, to protect participants
5. Synthesize
Notes into a spreadsheet: time, completion, errors, quotes and paths. Then an affinity map, built by task.
(opens in a new tab)
The affinity mapOpen

Session notes and screener responses stay closed, to protect participants.

The sessions

What happened when people actually used the app.

Times are from P2's full session sheet. The qualitative outcomes draw on affinity-map clusters across all 10 sessions.

Task 1.1Completed
Capacity check
Anacortes → San Juan Islands. Is there space for the car?
Most completed. Capacity bar interpreted inconsistently.
Time on task: 28s
Task 1.2Completed
Next departure
You're going to miss your ferry. When's the next one?
Quickly completed. “Pretty easy.”
Time on task: 17s
Task 2Struggled
Inter-island route
Friday Harbor → Shaw Island schedule.
Failures common. “Couldn't find Friday Harbor.” Backtracking.
Task 3.1Mixed
In-progress ETA
Family on the Bainbridge ferry. When does it arrive?
Most went to Vessel Watch first. Some succeeded there; some backtracked.
Time on task: 18s
Task 3.2Mixed
Future schedule
Bainbridge → Seattle, one week from today.
Longest task. Users went to Reservations first by mistake.
Time on task: 54s
Task 3.3Completed
Save route
Save Bainbridge → Seattle for quick access.
Largely completed. Save mechanism didn't match expectations.
Time on task: 33s
Task 3.4Mixed
Check alerts
Any alerts for Seattle → Bainbridge?
Expectation mismatch dominant. Riders wanted alerts tied to saved routes.
Task 4.1Mixed
Explore Vessel Watch
Port Townsend → Coupeville. What helps trip planning?
Cameras and live tracking valued, but information felt scattered.
Task 4.2Struggled
Buy a ticket
Tickets for you, two friends, and a car.
Most problematic task. External redirect; vehicle-size copy created anxiety.

Findings

Five patterns the sessions kept producing.

Twenty-six friction points came out of the sessions. These are the five that changed how the team talked about the app.

a.The ticket-buying flow breaks user expectations entirely.
b.Vessel Watch became the first instinct, even for tasks it wasn't built for.
c.Inter-island routes were nearly impossible to find.
d.Riders expected information to flow across sections; the app keeps it siloed.
e.Drive-up space indicators were visible, but inconsistently understood.

a.Buying a ticket

The ticket-buying flow breaks user expectations entirely.

High severity

When riders tap Buy Tickets, the app redirects them to an external mobile site (wave2go.wsdot.com). The handoff is jarring. The vehicle-size language makes it worse: red warnings that most standard compact cars do not qualify for the smaller bucket, arriving at the exact moment of purchase intent.

The Buy Tickets redirect to the external wave2go site

Evidence

Task 4.2 was the most consistently problematic across participants. Failure clusters dominated: expecting tickets to be in the app, frustration, non-conventional completions, pricing confusion.

What should change

High priority

Bring ticket purchase natively into the iOS app, or at minimum set the expectation before the redirect. Simplify the vehicle-size language: the red “most compact cars don’t qualify” warning creates anxiety exactly when riders are trying to commit.

b.Vessel Watch

Vessel Watch became the first instinct, even for tasks it wasn't built for.

High severity

Riders went to Vessel Watch first when looking for ETAs, future schedules and trip-planning information. The live map was valued, but the label and the information architecture didn't communicate what each section is for, so riders used the most visual one as a fallback.

Evidence

Tasks 3.1 and 4.1. Multiple participants opened Vessel Watch first and then backtracked; “Vessel Watch first” was the dominant initial-navigation cluster.

What should change

High priority

Reconsider “Vessel Watch.” Either rename it for what it is, live vessel locations, or expand it to cover the tracking and schedule jobs riders already bring to it.

c.Inter-island routes

Inter-island routes were nearly impossible to find.

High severity

Friday Harbor → Shaw Island had the highest failure rate. Participants who succeeded leaned on their own knowledge of San Juan Islands geography rather than on the app's structure. The route hierarchy didn't match how riders think about island-to-island travel.

Evidence

Task 2, failure and confusion clusters. P2: “couldn't find Friday Harbor.” The common detour was into Vessel Watch and Reservations before backing out.

What should change

High priority

Surface inter-island routes at the top level, or add a route search that doesn’t require riders to already know how the San Juans are organised. As it stands, succeeding at Task 2 is a geography test.

d.Siloed sections

Riders expected information to flow across sections; the app keeps it siloed.

Medium severity

The app splits Departures, Vessel Watch, Buy Tickets, Reservations and Cameras into discrete tabs. Riders expected them to talk to each other. A saved route should carry its alerts, and a future-schedule lookup shouldn't start in Reservations. The separation forced unexpected detours.

The Departures, Cameras and Vessel Watch tabs

Evidence

Tasks 3.2 and 3.4. Riders went to Reservations expecting a schedule, and in the alerts cluster expectation mismatch was the dominant theme, not findability.

What should change

Medium priority

Consolidate around a route-centric trip view that pulls schedule, capacity, reservation status and alerts for one route into one place, instead of asking riders to navigate four sections to assemble a single decision.

e.Drive-up space

Drive-up space indicators were visible, but inconsistently understood.

Medium severity

Some participants read “53 more spots” as remaining capacity; others as the total. The green progress bar was noticed, but its meaning wasn't universally clear. The fine-print disclaimer that vehicles in line for the tollbooth aren't included in the estimate was almost never read.

Evidence

Task 1.1 was largely completed, but interpretation observations recurred across participants. Capacity-bar meaning was the dominant secondary cluster.

What should change

Medium priority

Add a short legend on the capacity bar stating what the count means and that vehicles queued for the tollbooth aren’t included. The disclaimer exists today, at a size almost nobody reads.

Outcome

The first moderated usability study ever run on the WSDOT Ferries app.

First
Moderated usability study ever run on the WSDOT Ferries app
26
Friction points documented, clustered into five patterns
2 of 9
Task scenarios failed outright. Both were the highest-stakes ones: buying a ticket, and finding an inter-island route.
5
Prioritised recommendations delivered with the report

All friction points, all recommendations, and every session note.

Read the full report

Reflection

Riders arrive with one job: get on the right boat.

The app is built around what WSDOT publishes: schedules, vessels, reservations, tickets. The strongest finding wasn't any single bug. It was the consistent gap between the app's section structure and the route-and-moment mental model riders bring to it.