A redesign of the photo store used at tourist attractions – designed without research data and verified only after launch. More orders from fewer photos.
Two months after the new site went live, Lookout Amsterdam photographed 13.9% fewer groups than a year earlier. Online orders grew by 12.5% over the same period, and the average order value by 24.5%. Fewer people at the entrance, more money at the exit.
It is the only result in this project that truly means something – and further down I explain why I don’t credit all of it to myself.
01 Provide your email
02 Continue with WhatsApp
03 Account menu
04 Your order
GoPhoto started in Amsterdam with an idea about making photography in the entertainment business more efficient. Six years later the company operates in 18 countries and exports through the US, Asia and Europe, across more than 65 locations. It sells attractions a complete plug and play system: hardware, software, green screen technology and photo products. This is not one product for one market, it is the same sales mechanism running in Amsterdam, in Paris and on the other side of the globe.
That has a direct design consequence. GoPhoto is not a brand anyone knows. The visitor lands on a site carrying the branding of the attraction they just visited, and GoPhoto is reduced to a “Powered by” line at the bottom of the screen. Every location has its own visual variant, so the layout has to survive a swap of colours and logo without falling apart, while working for someone who has never heard of the brand before and will not hear of it after the purchase.
The new site launched first at A’DAM Lookout, GoPhoto’s flagship location. It is an observation deck in Amsterdam with over 800,000 visitors a year, of which the system photographs 80%.
There are two photo moments. Before boarding the lift, guests sit on a steel beam in front of a green screen while a capture tower with two cameras takes the shot, and automated editing swaps in a background within seconds. The second moment is at the top: “Over The Edge”, the highest swing in Europe, a hundred metres out over the edge of the building, shot with a mobile camera setup.
That explains the shape of the offer. The product is not one photo, it is a set of assets from two different attractions in the same place, which is where the swing video and 360° packages come from.
More important for the design is that the Photo Website does not operate in a vacuum. The retail area holds seven order stations where photos can be bought immediately, on a large screen, with staff on hand. The site picks up everyone who did not do that: they walked out, moved on, put the decision off.
On arrival a visitor gets a card with a QR code, scans it, and lands on a site where they can buy their photos.
People didn’t understand how to buy digital content from a place they had just visited. Support kept answering the same questions over and over. The site itself looked dated, which made the whole product feel unprofessional.
That was all we knew. There was no user research, no testing, no session recordings. The brief from the company was: increase sales by improving the “login” flow. Beyond device data we had nothing, so to a large degree we were guessing. I say this openly, because it changes the nature of every decision described below. These were bets, not conclusions.
One number did exist, though – and it turned out to be the most important one in the whole project:
The entry path is different from ordinary e-commerce. Nobody lands on this page from a search engine or an ad. You have to scan a physical card with a code – one you were handed hours or days earlier.
That created real friction: people lost the cards. The founder wanted to solve it with face recognition, so a photo could be found without any code. So I led the project with features in mind that didn’t exist yet: face search, home delivery of physical products, and possible photo customisation.
On top of that came multilingual support. The site detects the user’s locale and switches to their language. I designed it in four: English as the default, Spanish, Italian and German. A tourist who visited Amsterdam and scans the code back home shouldn’t have to switch anything manually.
01 Move closer
02 Registered faces
03 Scanning
04 Matched photos
Everything else in the product is just a route to a single view. The visitor scans the code, sees their photos, and within seconds decides whether this is worth anything. I designed that screen at 375 px only, in one column, with no separate desktop layout.
Photos keep their own proportions rather than being cropped into uniform tiles. A shot from the swing above Amsterdam and a group shot against the skyline have different shapes and different meanings, and a square grid crops away exactly the thing someone is looking at while deciding to buy.
Every unpurchased photo carries a watermark and a share button. That tension is deliberate: the watermark protects the sale, sharing dilutes it. I kept both, because a photo sent to friends before purchase is advertising for the attraction rather than a lost transaction.
The gallery is not a static list. Photos arrive from the location asynchronously, so the screen handles a “New photo available” state with a notification mid-browse. Between photos I also allowed for venue ad banner slots, for locations that use them.
Download all photos and videos you own for this photo code.
Download all
Nothing here yet
There’s nothing to show for this photo code yet.
Coupon applied: -100%
Adyen gateway
Sorry your payment failed!
Try again or
Purchase Confirmed!
Your photos are being prepared and will be sent to your email address shortly.
Are you sure you want to logout?
Get your photos in a high resolution
Download all your photos in high resolution (without any branding overlay). Image size: 1920 x 1280 pixels.
Copy the ID number and give it to our support team in help center.
1 pictureGophoto
AirDrop
Kelly Hu hu
Group2 PeopleMichael LastHanan
Instagram
Facebook
WhatsApp
Gmail
For any issues or questions, reach out to our support team at [support@example.com].
That a photo taken on site doesn’t have to end its life the moment you walk out of the attraction. That it can be delivered later, or a visit reminder can be sent – instead of counting on someone coming back to the site with a crumpled card in their pocket.
That belief set up everything else. If the moment of purchase is detached from place and time, the site is not a kiosk at the exit. It is something that has to work on its own – at home, on a phone, weeks after the fact.
The desktop slideshow view. It looked good and was ready in the first iteration: a large preview on the left, a thumbnail carousel, a toggle between slideshow and grid.
I threw it away because 95% of the traffic was phones. Maintaining a separate, richer mode for one-twentieth of the users meant two layouts to design, two to build and two to maintain. That time was worth more elsewhere.
Data from Lookout Amsterdam, two months after launch.
Two months after launch, compared year over year – so seasonality is taken out of the picture.
The new site went live together with a WhatsApp integration. Two changes at once.
How much of the +12.5% is the redesign and how much is the new photo delivery channel. I don’t have a clean measurement of the design alone, and I’m not going to pretend I do.
With fewer people entering the funnel, more orders of higher value came out of it.
