
(Overview)Seaya sells week-long diving trips to Malaysian islands, paid upfront by Australians who have never been there. I asked 35 people what goes wrong when they book a hotel, watched four bookings up close, and built a page that answers what people actually stop on.
- Client
- Seaya Holidays
- Sector
- Travel
- Year
- 2022–2023
- Role
- Product designer
- Scale
- 4 island resorts
- Status
- Shipped
The colours, off the mark
- SunGold#F0B448
- WaterShallow#6CCCD8
- WaterMid#6CC0D8
- WaterDeep#60C0D8
- WordsTeal#306C78
Type
Demi carries the name. Light does the letterspacing under it.
Nobody sends money to a stranger for an island they have never seen.
A Seaya package costs between $775 and $1,340 a person. You pay before you fly, to a small agency you found on the internet, for a resort you can only see in photographs.
There is no shopfront and no second chance. Every doubt a person has, every question they would put to an agent across a desk, has to be answered on the website, because there is nowhere else for it to happen. Get it wrong and nobody complains. They close the tab, and you never find out why.
Seven people asked the survey for more photographs, and three said why in the same breath: the pictures are too generic to tell the rooms apart. One asked for no stock imagery. Another wanted images of the actual room. A woman in a session put it harder: “you have to have loads of photos, and if you didn’t I’d be very dubious. I wouldn’t go there because they hiding something.” Photographs are not decoration here. They are how you compare, and how you decide whether anyone is telling the truth.

A company nobody has heard of has to look like somewhere you would actually go.
The logo was mine too, and it had the same job as the site.
So the mark is a place rather than a monogram: a map pin, with a gold sun dropping behind three layers of water. The name does half the work already, sea and see, and the mark finishes the sentence.
It sits in the header of every page, which is the first thing anybody judges and the last thing they remember.
Gold for the sun, three turquoises for the water, one deep teal for the words. A palette you can picture before you have read anything, which is what you want when the thing being sold is a week somewhere warm.

So I stopped guessing, and watched people try to book.
I asked 35 people about the last hotel website they used. Six of them had failed to do what they went there for.
Then I watched two people book on two competitors’ sites, four sessions in all, because Seaya had nothing to test yet, writing down what they did and said to the second.
The form was never the problem. Not knowing was. One woman clicked a calendar arrow six times to reach her month, waited fifty-one seconds for a page, could not work out an add-on called Double bed, and still finished with “I don’t know if it’s Free Cancellation. I’m not ready to go through to payment yet.”
So I put the answers in front of the form. Every resort page carries all four room types, every package, every price and a plain list of what is included, in one view, with a booking button on each room. Nothing that decides the sale sits behind a click.

Doing the research is the easy half.
I find out what actually stops people, then argue for the design that answers it, including when the answer is less elegant than the one everybody already likes.
Here the tidy option was a step-by-step wizard, and every instinct in the category points at one. The right answer was the page you could read across, because the sessions showed people withholding a decision rather than getting lost.
Holding the line on what the research found is the half that gets the thing built.
The reasoning runs both ways. The photographs are specific because seven people asked for better ones. The rooms sit side by side because comparing them was the survey’s most common complaint. The change-of-mind window sits beside the price because a woman got all the way to payment and stopped, over exactly that.

Four resorts, every package in one view, nothing held back for the checkout.
On each resort every room type, every package length and every price sits in a single view, with a booking button beside each room. Telling rooms and packages apart was the survey’s most common complaint, and it is the one thing this page is built to make easy.
I handed the site over at launch, so there is no analytics account behind it and no conversion figure. The four numbers beside this are from the research, not from the result. They are what I actually know.
- People surveyed
- 35
- Did not finish
- 6 in 35
- Bookings watched
- 4
- Longest wait
- 51 sec

I ran the research, designed the journey and the logo, then built the front end. Seaya set the prices, and the resorts set what each package held.
What I did
- The survey and the four booking sessions
- Flows, interaction design and the visual design
- The logo and the identity it carries
- The front-end build
What the team did
- Seaya chose the resorts and set the prices
- The resorts set what each package includes
01 / 06
Use the left and right arrow keys to move through the images.
(The call)Everything on one page, rather than the tidy step-by-step booking wizard the category expects.
A wizard tests beautifully in a lab, because each screen on its own is simple. It was wrong here, and the sessions said why. People were not confused, they were unconvinced, and they wanted two rooms and two prices in their head at the same time. A wizard hides the thing you are comparing behind the step you happen to be on. The woman who said “I like when it’s all on the one sheet, more like an Excel sheet with the picture, it’s just easier to compare the prices” was not asking for a better flow. She was asking for a table she could read. The wizard would have scored better on first use and sold fewer holidays.
Website






People gave up most often because they could not tell a room from a package. That is the size of thing a booking flow is actually lost to.
The research was done on competitors rather than on Seaya, which is why it was possible before there was anything to test, and on hotel booking rather than dive booking, which is the one transfer I trusted rather than tested. The next honest step was running the same method on the built page. I did not get to it then, and it is the first thing I would do with it now.





