100,000 Pages High at TUI Group

TUI lived on low-margin metasearch traffic. I designed one destination page and the rules to generate 100K+ of them, and organic traffic grew 4× where they went live.

  • Senior Product Designer
  • Design Systems
  • E-commerce
  • Multi-market

In short

Team

  • The head of e-commerce and me at the start, creating and pitching the base concepts in a Lisbon hub that grew to about 10 people across product, SEO (search engine optimization), data, back-end and machine learning, front-end, and design.
  • An outsourced team for the checkout flow and our first in-house engineers, hired exclusively for the PageBuilder project.
  • Me, the designer: destination template design, generation rules, UX and UI, front-end QA (quality assurance), and the Tagus design system.

Problem

TUI’s existing website relied on metasearch for revenue: a skinny slice of the market, a dangerously homogeneous customer base, and our thinnest margins. The new website needed organic traffic, and competitors already ranked high with destination pages.

What I did

I designed PageBuilder: one destination template and the rules for a generator, fed by geo and place data, that built a page for every real place where TUI could sell a stay. And we built Tagus from scratch, the design system that kept every page and booking flow coherent.

What changed

An experiment for 100K+ pages for 175 destinations in Portugal and Spain, with the idea to expand to 5 markets, each with a complete booking flow. The result: organic traffic grew more than 4× and metasearch dependence fell.

A site nobody found unless we paid

We had a new TUI e-commerce website in the pipeline to replace the legacy one, but no way yet to bring real people in. The existing website relied heavily on metasearch traffic to survive: a skinny slice of the market and a dangerously homogeneous customer base.

We knew that to reclaim better margins, we had to get off metasearch and bring in more organic traffic from people already in a traveling mindset. Competitors already did that, with destination pages that helped those people get to know a place and choose what to do, where to stay, and what’s worth seeing.

So we framed our hypothesis: with millions of destination pages indexed by search engines, we would increase traffic and drive more sales from new types of customers, more drawn to the travel experience itself, and a more natural fit for TUI Group.

One template, millions of places

Our small Lisbon hub was meant as an ideas lab for TUI’s e-commerce. We pitched concepts driven by instinct and the courage to try different ways of doing things, looking for disruptive products.

So, with this problem at hand and the leeway the company gave us, we rolled up our sleeves and designed one template and the rules to fill it in. We called it PageBuilder. We wanted a flexible system of fast-loading static sites, with content generated from geo data plus information and images about each place. The output still had to be appealing and competitive: pages that inspire people to book a stay somewhere beautiful.

This project was never meant to be local. The template had to survive different cultures and markets for a site that could grow globally. As a team dedicated to discovery and experimentation, we wanted to pitch a unified TUI Group website built on Tagus, the design system we were creating from scratch.

We grounded the work in what travelers needed: a nice destination, a comfortable hotel, memorable experiences, and recommendations to make the most of the stay. We benchmarked Expedia, Booking.com, Agoda, Airbnb, and all the other usual suspects and adapted Baymard’s e-commerce findings to travel, so we could set our guardrails for good travel UX.

Blue sky, then down to earth

We tested wireframes with real users, reading the heatmaps and their unbiased interactions, and used them as a base for the metrics we would later track on the live pages. From critique sessions and usability tests, we crafted the vision interface: blue-sky thinking about what we would love to build. It was ambitious: weather forecasts, bookable activities, and more of what travelers want from a destination page.

Then we set out to build it and hit limitations that forced us to adapt. We iterated in code prototypes and cut each page to what it could answer well in a first version: what the place looks like, where it is, where to stay, and what to visit. The ambition still paid off: in some cases we pushed the limits and built composite back-end systems to support the idea and honor the vision designs.

Tagus, a design system with global ambitions

None of this would hold together without a design system, so Tagus was one of the first things we did: we designed and built it from scratch, in parallel with the vision designs. It allowed us to keep the generated pages and the booking journey—from inspiration to checkout—cohesive across the different markets, with every component documented for designers and developers.

Just 100,000 pages in one click

Wired end to end, the proof of concept generated 100,000 pages, a first step toward the millions of pages we wanted to target in our hypothesis across 5 markets: Portugal, Spain, Brazil, Italy, and Malaysia. Each page was a real place with TUI offers and a complete booking and payment flow. An outsourced team built the checkout flow, with front-end QA from me to keep it true to the design, while our first in-house engineers were hired to work exclusively on PageBuilder.

The plan was to open all 5 markets. We launched in Portugal and Spain first, 175 destinations, with the other 3 markets to follow, after experimenting: A/B tests on the order of the content, session recordings to watch real interaction, and a chatbot to keep hearing from people.

Setting the standard for TUI Group

Our analytics team measured organic traffic growing more than 4× in Portugal and Spain, and our dependence on metasearch engines dropped significantly, saving money from the ads. The work left reusable assets too: a geo-data lake, a content generation tool, and cleaner data schemas for the legacy back ends we had to work with.

Because the design was global from the start, we took it on tour across TUI, presenting it to the Nordics, the UK, and Germany, home to TUI’s headquarters, to replicate it for the whole TUI Group—the ambition was high.

Looking back

I’d keep the same attitude: not be afraid to start from zero, look for radical new concepts that answer people’s needs, and build the structure that will lead us to that output.

What I’d change: bring better data to enrich the pages, pitch our ideas more often, push to build with our engineers from the start, and get more flexibility to create our own solutions. When the engineers finally sat next to us, we started to gain traction, and many stakeholders backed the project, but a change in leadership brought different priorities before we could continue.