Adioso: making the travel site people want, even if it’s impossible



This time two years ago, things looked pretty grim for us at Adioso.

We were out of money, we’d lost all our team members other than myself and my co-founder Fenn, and we were emotionally and physically exhausted. As our first investor and mentor Paul Graham put to us so bluntly, we’d failed to reach a milestone with the money our investors had given us, and we looked like just another failed startup.

It was a bitter fall from our heights of two years earlier when we’d progressed from the Y Combinator Winter ’09 batch, buoyed by the sense that our product was one of the most exciting in our batch, at least to our fellow YC startup founders and the younger, more travel-savvy investors.


What had gone wrong?

Sure, as company founders we hadn’t measured up. We weren’t great at fundraising, and we didn’t raise enough money to do a whole lot. We weren’t able to get the best out of our team members. Our attempts at building the technology we needed to power the product had fallen short. In desperation, we wasted time and money on marketing stunts.

But all that was peripheral to one key fact. The product we’d envisioned was actually impossible to build.


We had no idea it would be impossible when we started.

When we started prototyping the concept in 2007, it didn’t seem such an insurmountable task.

We just wanted a travel site that gave us a good answer to questions like “best flight to anywhere in Southeast Asia next month for 10 to 15 days“. So we built it, and it worked really well.

At least it did for us, from our home city of Melbourne, Australia, on the low-cost airlines we were happy to fly with. We started promoting it to travellers in Australia, and they were ecstatic. “The travel search site I’ve always wanted!”, people excitedly told us.

So we felt like we were winning. Even before applying to Y Combinator, we were familiar with Y Combinator’s mantra “Make something people want”, and we felt we were doing just that.

So when we were accepted into Y Combinator, we thought all we had to do was keep growing it so everyone else in the world could use it, and within a few months… boomtimes!

Two years later, with our numerous attempted hacks to get around the unavailability of flight data doing little to generate any traction outside Australia, the message was becoming clear; there was no possible way of making this product work.

It wasn’t just us. No one else had succeeded in building anything remotely like our product either; not even the biggest, best-funded, most PhD-flush companies in the entire travel world.


What made it impossible?

The main thing was that there was no way of sourcing all of the world’s flight data in a way that was adequately complete, accurate or up-to-date.

For our product to work, we needed a complete database of all the flights, from all the airlines, with fares and availability updated in near enough to real time.

It turned out that you just could’t get that kind of data, from anywhere, at any price.


What to do?

People still really wanted our product. We knew that. The feedback was never “this is stupid, no-one wants this.” It was more like “this is awesome, I totally want this but it’s useless as it doesn’t have the airlines or destinations I want.”

We had all kinds of ideas ourselves, and suggestions from investors and advisors about what we should do to make the company work.

But every idea had the same flaw: we knew it wasn’t what people wanted.

The site we’d been building from day one was the thing people wanted. That was the only thing we were willing to keep building.


We just decided to keep on building it anyway.

Whilst it might have been impossible to successfully build our product in 2010 or 2011, we refused to believe it would be impossible forever.

So we decided to keep going, working on the things that were within our control.

We built a new flight search technology platform, which we called Wingtip, that had the features and performance and scalability that our earlier attempts had lacked and that still no other travel site offered.

We kept developing our user-interface, continually iterating to come up with something that was more beautiful and delightful to use than any other travel search site.

And we would experiment with marketing and distribution, validating ideas for a commercial platform to make Adioso not just something consumers want, but something airlines and travel marketers would want too.


Here it is.

It was 18 months ago that we made the decision to keep building, and as of the last couple of months, Adioso has finally started to become the product that Fenn and I always aspired to create.

We now describe Adioso like this: travel search that works the way you’re thinking.

The video at the top of this post explains a bit about what it does.


“But it’s so slow!”

Actually, the bit that Adioso controls is very fast. Our Wingtip engine enables us to answer fight search queries faster than pretty much any other travel site – if we have flight data pre-loaded into it.

That thing about not being able to get hold of a complete database of flight fares and availability? It’s still our biggest constraint.

So where we can’t currently get the data to pre-load into Wingtip, we do realtime lookups on other large travel websites. That’s slow, especially when you have to do dozens or hundreds of searches to deliver a single set of results.

But we didn’t want to let slowness stop us from being idealistic about building Adioso the way we’d envisioned it.

Just like modems and personal computers, mobile phones and now 3d printers, you start with a slow, barely-working first version, then improve performance from there. But the product has to exist before you anyone will support you in making it faster.

Oh, and even if it’s slow, it still beats the endless hours many people currently spend searching numerous permutations of destinations and dates on conventional travel search sites.


Flight data is still a problem, but at least now it’s our only major problem.

Two years ago, we didn’t have much going for us. The inability to obtain flight data was a problem, but it was just one among many.

Now we have a team of amazingly talented engineers and designers, a killer technology platform, and a user-interface that improves noticeably every week.

And people no longer say “Adioso would be awesome except it doesn’t have the airlines and destinations I want”. It has all the airlines and destinations now.

Now people say things like this…

Facebook comment


HackerNews Comment


Facebook comment


Things may be about to change.

For the first time, we’re seeing signs that the flight data we need is becoming available, in the right format, at a cost that might be economical for us.

So now we have a new challenge: to establish deals that get the availability of accurate flight data increasing, and its cost decreasing in such a way that Adioso can become continually faster and better.

If we can make that happen, we might finally succeed in making the travel site people want.