Senior Product Designer · Booking.com Flights

Selling the seat, not the square.

A 10-week vision for how Booking.com Flights sells seats. I led it end to end, from the research and workshops to the roadmap, and took it to leadership, who adopted it for 2026/27. The team is shipping it, one experiment at a time.

Sep to Dec 2025

The seat selection page in the vision: one tab per flight, the flight's duration and airline, a card for each passenger, a seat map key, and the seat map with paid seats filled and free seats lighter. Four handwritten notes mark what changed: introduce tabs, one per flight; display flight duration and airlines; improve passenger selector visibility; and refine progression CTAs.

The seat selection page as it was: a grid of identical squares, with the price given only as a range in the legend.

The problem

82% of customers see a seat map. 11% pay for a seat.

Seat markups had just made seats a revenue line, and airline partners were telling Booking the seat map was below market standard.

82% Customers shown a seat map Offer rate

22% Free and paid combined Overall attach

11% Customers who pay for a seat Paid attach

78.3% Leave their seat to chance No selection

The problem · What customers saw

A grid of identical squares, with a price range in the legend.

Every seat looked the same. Prices appeared only after you picked a seat. Nothing showed legroom, where the front of the plane was, or which traveller you were choosing for.

The seat selection page as customers saw it: a grid of identical squares, with the price given only as a range in the legend.

The insight

Customers weren't refusing to pay. They couldn't see what they'd be paying for.

What a seat is worth

I paid a bunch of extra money, but can't find out what seats have legroom.

It's hard to tell the difference between premium seats

Explain what seats are the extra legroom seats

Reading the map

The plane map was not clear and it was confusing to choose seats.

when we went to select our seats the layout of the plane made no sense. very confusing!

Sitting together

We want to sit together. Cannot get any information about this.

Knowing what to do

Better filters & explanation of how/when to pick seats

The seat selection process is confusing and terribly aggravating.

Handpicked customer feedback on seats, word for word.

How I know

I checked it three ways.

Our product

The Matrix Audit: 13 missions across desktop, mobile web, iOS and Android.

Families with an infant got no seat map at all.

The market

The best airlines already sell the seat itself.

Singapore's 360° view, United's side-by-side comparison, Emirates' clickable legend.

The data

Couples are more willing to pay for a seat.

They make up 29% of free seat selections, but 39% of paid ones.

How I know · Our product

I designed the Matrix Audit and facilitated it as a cross-team workshop.

I ran it to get a clear picture of today's seat experience before designing the next one. In the room: 2 group product managers, 6 product managers, 4 engineering managers and over 8 designers. They paired up on 13 missions, each a different mix of platform, traveller and trip, to find what stops people paying for a seat, where the app and the web disagree, where complex trips break, and what we could fix straight away.

The Matrix Audit board: five groups side by side, each with its two or three missions and the screenshots and notes captured while running them. Handwritten labels name each group's platform: desktop web, desktop web, mobile web, iOS app and Android app. A note points at one mission card: one of 13 missions.

What it all pointed to

The audit, the market and the data pointed to three gaps.

1 · Architecture

The page guided no one.

No zones, no sense of front and back, no clue which traveller you were choosing for.

2 · Value

Nothing explained why one seat costs more.

Standard and extra-legroom seats looked the same.

3 · Decision

Everyone got the same map.

A family with an infant and a solo traveller saw the same experience, whatever the trip.

The call

Fix them in the order they depend on each other.

I designed and facilitated a cross-team ideation workshop, then set the order. You can't show value on a page nobody can read, and you can't recommend a seat whose value you can't show.

1 · Architecture

A page you can read

2 · Value

Each seat's value, shown as you choose

3 · Personalisation

A seat chosen for the trip

Value · The quick win, shipped straight from the audit

Prices up front lifted paid seat bookings by 3.6%.

One part of the Value gap didn't depend on a new page. The audit kept hearing that you only knew a seat's price once you'd selected it, so I pulled it out and shipped it straight away.

Impact +132 a day More bookings with a paid seat

−3.9% Clicks on Back. Overall bookings held

The seat selection page. The seat map runs from row 1 to row 32, with a price label above the rows: €10.00 on the paid rows and Free on the others. One price label is ringed, with a handwritten note: new, the price on the seat map.

The seat selection page before the change. The seat map runs from row 1 to row 32 with no prices on it: a darker blue marks the paid seats, and the legend gives only a range, €7,00 to €20,00.

The vision · 1 of 3

First, a page you can read.

Context, controls and the map each get their own zone. Filters, a clickable legend and plain-language search all land on one map.

Impact +22.8% More bookers chose a seat

+32.9% More bookers paid for a seat

+1,182 a day More bookings with a seat

−11.9% Skips on the seat screen

249k Visitors in the experiment, over 10 days

The new seat selection page: a tab per flight, the flight's duration and airline, a card for each passenger, a seat map key, and the seat map from row 1 to row 33 with the aircraft named and its front, toilets, galley and exits marked. Handwritten notes mark what changed: introduce tabs, one per flight; display flight duration and airlines; name the aircraft and mark the front; toilets; galley; improve passenger selector visibility; refine progression CTAs; paid seats filled; mark the exits by the row; free seats lighter.

Hovering a seat shows its value before you choose it: seat 3D, €30.50, front row and aisle seat, with wifi, a USB charger, a meal and in-seat entertainment, and a button to select the seat. Two handwritten notes mark the card: show where the seat is, and show what comes with it.

The vision · 2 of 3

Then, each seat's value at the point of choice lifted paid seat bookings by 3.85%.

Pitch and legroom on hover. Photos and a 360° view in full details. The richer seat data needs a new content integration, so the roadmap carries it as its own line.

Impact Seat value on hover +3.85% More bookers paid for a seat

+114 a day More bookings with a paid seat

352k Visitors, 9 days. Bookings held

Cabin details 2,304 People opened it in one week

The cabin details panel, open over the seat map: Your onboard experience, Boeing 737 MAX 8-200, a photo of the cabin with two travellers seated, 35 inch legroom and wide seats, and the flight's amenities: free wifi and USB ports included, food and drinks and in-seat entertainment for a fee. Four handwritten notes on it mark what the panel does: name the aircraft, show the cabin itself, legroom and seat width, and list what's on board.

The vision · 3 of 3

Then, a seat chosen for the trip.

Quick Start asks how you want to sit. Solo travellers choose by intent, couples and groups stay together, and families get a child beside an adult by default.

The seat selection page with a Quick start card over its right side, hanging past the edge of the screen. The card says we can find seats together for everyone, asks how would you like to sit, and offers three choices, each with a price: keep everyone together, in seats 1A and 1B, group by family, and best value. Two buttons follow: Preview and Select seats. Two handwritten notes beside the card say: recommends seats, without the stress of searching, and so couples and families sit together.

Landing it

I proposed this vision to leadership. They adopted it.

I prototyped in Figma AI on BUI, Booking's production design system, took the work through design reviews, and presented working concepts, not sketches, to the Flights leadership team and to design leadership. It became the seat direction for 2026/27, with the roadmap in dependency order.

H1 2026 · Architecture and value

Layout, legend, filters, seat details on hover, the seat content integration

Mobile web exploration runs alongside.

H2 2026 · Personalisation

Personalised and cohort-based seat selection, assisted selection

Upsell nudges and availability messaging start at the end of the year.

2027 and beyond

Seats as a route to cabin and fare upgrades

What happened

The vision I led is shipping, one experiment at a time.

Three experiments are now live for everyone. Each one led more bookers to pay for a seat.

2026/27 Adopted as the seat direction Flights leadership

+3.6% Prices on the seat map +132 paid seat bookings a day

+32.9% The new seat layout +1,182 seat bookings a day

+3.85% Seat value on hover +114 paid seat bookings a day

What I'd change

Next time, customers see it before leadership does.

This time the experiments did the testing, after the vision was adopted. Five or six customer sessions before the leadership review would have brought that proof forward.

Back All work Moyinoluwa Olufisayo Next Customers wanted to buy. There was nowhere to sell. Extras Vision · 2024 to 2028