Besttables: Building the Restaurant Booking Platform TripAdvisor Later Acquired
Ruby on Rails Specialist and team lead at RUPEAL, Besttables (built by RUPEAL) · 2009 – 2011 · Lisbon, Portugal
In short: At RUPEAL, a Lisbon Ruby on Rails software house, I helped win the proposal to build Besttables, a real-time restaurant booking platform, and led the team that built it with founder and CEO Ricardo Sécio. After about 18 months of development it launched in April 2011 and went from 70 to 250 restaurants in its first four months. The platform became the core of a business that grew to more than 800 restaurants in Portugal and Brazil. In 2015 TripAdvisor acquired Besttables and made it TheFork Portugal and TheFork Brazil.
| Product | Besttables, free real-time restaurant booking (web, iPhone, Android) |
| My role | Ruby on Rails specialist and team lead at RUPEAL, the technology partner |
| Build | ~18 months of research and development, launched April 2011 |
| First months | 70 → 250 restaurants, about 2,000 visits a day |
| Outcome | Acquired by TripAdvisor in April 2015, became TheFork in Portugal and Brazil |
What was Besttables?
Besttables let people book a restaurant table online, for free, with instant confirmation: "Reserve a sua mesa. Simples. Gratuito. Confirmado." Restaurants received a free management system to run their bookings, and paid a commission for each guest seated through the platform.
In 2011, online booking was almost unknown in Portugal. Investors later estimated that about 2% of restaurant bookings in Portugal were made online, against 20–30% in the United States.
What was the challenge?
A booking platform only works if both sides trust it:
- Diners need a real confirmation. If a confirmed table isn't there, the diner never comes back.
- Restaurants need a tool they actually use. A table is only bookable online if the restaurant keeps its availability up to date during a busy service. The back office had to be faster than the paper book on the counter.
So the core problem was real-time availability: one source of truth for every table, shared between the diner's screen and the restaurant's floor.
What did I do?
1. Won the build. RUPEAL competed to be Besttables' technology partner. I worked on the winning proposal and then led the RUPEAL team that worked day to day with Ricardo Sécio, the founder and CEO.
2. Designed with restaurants, not just for them. We built the product with chefs and front-of-house managers from Lisbon restaurants including Olivier, Alma and Feitoria. Their service routines shaped the booking flow and the restaurant back office.
3. Built it iteratively. Instead of a long specification followed by a big delivery, we shipped in short agile iterations, tested with real restaurants, and changed direction when they told us something didn't work. Besttables' own launch page credited RUPEAL's technology team and this way of working.
4. Built the real-time booking engine and the restaurant back office. Ruby on Rails, PostgreSQL and jQuery: live table availability for diners, instant confirmations, and a management system for restaurants, plus search and maps for finding a place to eat.
5. Launched under real load. The beta went live in March 2011, and the public launch in April 2011 coincided with Lisboa Restaurant Week, with bookings for the event taken through Besttables. That is a demanding first week for any booking system.
What were the results?
- Fast adoption. 70 restaurants by the end of the launch month and 250 by July 2011, with about 2,000 visits a day.
- Growth beyond Lisbon. Porto and the rest of Portugal by 2012, then São Paulo, Brazil.
- An acquisition. In April 2015 TripAdvisor bought 100% of Besttables. It had more than 800 restaurants, about 20 employees, and was profitable in Portugal. It became TheFork in Portugal and Brazil. At the time, the CEO credited the "superiority of the technology platform", built by Portuguese developers in Lisbon.
What did I learn?
Two-sided products are won on the supply side. Diners come for choice, and choice comes from restaurants that trust the tool. Building the back office with real restaurants mattered more than any feature for diners.
Iteration beats specification for new markets. Nobody in Portugal knew what online restaurant booking should look like in 2011. Short cycles with real users found the answer faster than any document could have.
Good engineering outlives your involvement. I left RUPEAL a few months after launch. The platform we built carried the business for four more years, to an acquisition by the global leader in its category.
Building a two-sided product from scratch? Work with me.