Vimcar Logbook · Ash Rahal

Onboarding completion by month 3: 88.9% from 80.7% in month 1
Support tickets: −24% within 90 days
Auto-categorisation adoption, iOS: 74% 65.1% on Android, 12 weeks
Drivers using the app: 120,000+

Who Business drivers
People who drive for business most of the month. Fleet managers track company cars on the web.
What Dongle + app
The dongle captures trips; the app shows them for categorising.
Why An annual tax report
A logbook their tax advisor and the tax office will accept.

Owned iOS, Android and web
End-to-end design across the three apps.
Did Strategy to A/B test
Design strategy, technical architecture with engineers, prototyping, A/B testing.
Led People and decisions
Mentored junior and mid designers. Presented to product and engineering leadership.

Part 1 · Onboarding

People sent the dongle back before the trial ended

>20%

of new customers abandoned onboarding and returned the dongle within the 30-day free trial. Many contacted support about setup.

Setup spans a parcel, a car and a phone, in five steps.

Designed in collaboration with the Vimcar team.

I planned several flows with the mobile team, ran 12 interviews with a research partner, and A/B-tested onboarding concepts.

Figma prototype map showing two onboarding flows, Flow 5 and Flow 6, as long rows of connected screens. — Two of the onboarding flows I prototyped and tested, Flow 5 and Flow 6.

Add the vehicle
Connect the hardware
Set reminders

The new onboarding. State sequences from my prototype recordings. The last screen tells drivers to allow 5–10 minutes after their first drive before looking for trips.

Onboarding screen: Do you have your hardware? with buttons I have the hardware and I don't have it.

Decision 1

Ask before assuming the hardware is plugged in

Options: Assume the dongle is plugged in and start the sync, or ask first and branch.
Chose: Open the hardware step with “Do you have your hardware?”. Then scan or type the dongle code, with short install videos for the dongle and the box.
Why: Setup problems were driving support contacts, and returned dongles ended trials.
Gave up: One extra screen at the start of setup. A shorter flow wasn’t worth the support contacts.

Reminders step 5 of 5: warning that uncategorised trips become private after 7 days, reminder time 6:00 PM, and Monday to Friday selected.

Decision 2

Turn the 7-day rule into a reminder

Options: Leave the 7-day rule in the help docs, warn drivers after they’d missed it, or set reminders during setup.
Chose: A reminders step that states the rule up front: uncategorised trips become private after 7 days. Drivers pick the days and the time.
Why: Trips lock 7 days after the drive, a legal requirement. Missing the window costs drivers tax savings.
Gave up: Another step in setup, plus a notification permission request that some drivers decline.

Onboarding step 4 of 5: Your Addresses with home and work address, and Smart Address Recognition switched on. — Step 4 of 5 in onboarding

My Address settings: address recognition switched on, with a radius slider set to 500 m. — Later, in settings: radius 500 m

Decision 3

Collect addresses during setup, not after

Options: Ask for addresses during setup, or let the app learn frequent places from trip history over time.
Chose: Step 4 of 5 asks for home and work addresses, with address recognition on by default. Drivers can change it later in settings, including the radius.
Why: By the time the trip syncs, the app already knows the addresses, so drivers confirm instead of typing. It sets up Part 2.
Gave up: A longer setup, in exchange for useful suggestions from the first trip.

Onboarding completion, month 1: 80.7% after launch (BI)
Onboarding completion, month 3: 88.9% BI report
Support tickets: −24% BI report

Part 2 · Trip logging

An automatic logbook that still made people type

15–30 min

The dongle syncs 15–30 minutes after a trip ends. Drivers recalled destinations from memory and typed partner, company and reason by hand.

In Maze usability tests, 63% of participants hesitated to categorise right after a trip.

“I wish I could just swipe and be done.”

Test participant

Fahrtenbuch trip list with dates, start and end addresses, and a status icon per trip. — Before: the manual path.

Trip details: map, start and end address, and four buttons: Business, Commute, Private, Mixed. — Before: the manual path.

The goal was to collect context up front, so drivers confirm instead of type.

Turn on, get suggestions
Fix it on the map

Decision 4

Address snapping: suggest, then confirm

Options: Categorise trips fully automatically, or suggest and let the driver confirm.
Chose: When a trip ends near a saved contact, the trip shows “Address categorized”. Its details ask “Is this a business trip with Karl Wagner and Vimcar GmbH for the purpose of payment collection?” One tap fills the form.
Why: Drivers had been typing from memory. I tested address recognition concepts with 20 users in an A/B test.
Gave up: Full automation. Drivers still tap once, because the logbook is a tax document and the driver has to stand behind each entry.

Logbook list where the top trip carries an “Address categorized” label. — After: address snapping.

Trip details with a banner asking whether this is a business trip with Karl Wagner and Vimcar GmbH for payment collection, with a Confirm link. — After: address snapping.

Maze results: a bar chart on whether the app gave too little or too much information, a rating distribution, two heatmaps of taps on trip screens, and written answers. — 20-user A/B test. Maze results comparing address recognition concepts: information load, ease of navigation, tap heatmaps and written feedback.

Outbound · categorised by address snapping

OfficeCustomer

Return · now suggested as well

CustomerOffice

Research showed drivers wanted the trip back (B to A) categorised too, not just the trip out (A to B).

Business trip details after confirming: partner Bob Baumeister, company Bauhof, reason Return trip. — Drag to compare. The same form, before and after one tap.

Business trip details before confirming: partner, company and reason empty, with the return-trip question at the top. — Drag to compare. The same form, before and after one tap.

Decision 5

Return trips: the second leg is a question, not a form

Options: Copy the outbound category to the return trip silently, or ask.
Chose: Detect the return journey and ask “Is this a return trip from Bob Baumeister?”. Confirming fills partner, company and the reason “Return trip”.
Why: Users told us they wanted return trips auto-categorised.
Gave up: Silent automation. Asking costs a tap, but a wrong silent guess would end up in a tax record.

Swipe a row

Decision 6

Swipe to categorise, for everything else

Options: Swipe to categorise, or hold to categorise. I tested both in Maze.
Chose: Swipe won in testing and became the main action: swiping a row reveals Business, Commute and Private without opening the trip. Tapping still works for drivers who don’t swipe.
Why: A test participant asked to “just swipe and be done”.
Gave up: Swipe hides the actions until you discover the gesture, so the tap path stayed for drivers who never swipe.