Turning a travel widget into a sales channel

Redesigning sherpa°’s embeddable widget for visa requirements on partner sites. A passive block of information became a sales channel, inside the space each partner allows.

Overview

Trip Element is sherpa°’s widget showing travel requirements by passport and destination. It lives on 50+ partner sites as an API component. I was the only designer, from research to dev QA, on a team with 1 PM, 2 frontend and 2 backend developers, and 1 QA. The goal was 10% more visa sales.

+35%Purchase likelihood in the A/B test

+41%Task success

“He played a lead role in the design of our travel requirements solutions and was instrumental in the development of new features for our products.”

Yana StepchenkoCreative Director at sherpa°

The problem

The widget told people they needed a visa, then offered a plain button with no destination, type or price. It informed them but didn’t sell.

It also had real limits: partner space, API boundaries, and regulations that kept changing, so it had to stay clear and trustworthy.

The approach

Turn the button into an offer

Better copy was the smallest change for partners, but copy alone can’t answer what it is and what it costs. A list of visa options was compact, but harder to compare.

I went with a card showing destination, type and price, moved first and open by default, with room for more data later. A card takes more space on partner sites and showing the price could put people off, so we won space back elsewhere and ran an A/B test before committing.

The widget on a partner’s page, before and after. The airline and data are invented.

Building it

Price and arrival date

Buyers need to know a visa will arrive in time, so each card shows when. The price sits next to it, before the click, which filtered out window-shoppers. People didn’t click more, they just bought more.

Price and arrival date before the click. Sample dates.

Passport switcher

Changing passport inside the visa section cluttered it, and some partners have strict space limits. Moving it to the header unblocked partner adoption without layout regressions.

The requirement follows the passport. Sample data.

What didn’t work

Value props

We tested value props above the cards, “Fast & simple”, “Private & secure” and “Expert support”, to explain why to buy through sherpa° rather than a government portal. There was no significant uplift, since users still preferred official channels, so we dropped them.

The value props we tested and dropped. Sample data.

Impact

+41%Task success, measured across flows on UserTesting.com

+35%

More likely to lead to a purchaseIn a 2-week A/B test with about 200 users, so I treated it as a strong direction, not a precise number. The design then rolled out to all 30,000+ users, and visa sales moved the same way, well above the 10% target.

Takeaways

What I’d changeTalk to partners earlier, because their limits shaped the layout.

What I’d changeTest value props on a smaller scale first.

A widget on someone else’s site has to earn every pixel.

Clear information can do more selling than persuasion.