Steve Domin, Founder & CEO @ Duffel – Rebuilding Travel Infrastructure, Surviving COVID, and the Future of API-First Companies

0:00Steve Domin:Welcome back to Riding Unicorns. Today's episode is with Steve Dommer, founder and CEO of Duffel. Duffel is building the infrastructure layer for modern travel, taking one of the most complex and outdated industries and revolutionizing it. And they've raised over 50 million from the likes of Index and Benchmark. So welcome to the podcast, Steve. It's great to have you on. I'd like to start by asking, what did you see in the travel industry that made you think that's broken and I need to rebuild it? When I used to work for a payments company in London called GoCardless, a couple of friends and I tried to build a product in the travel industry and really struggled to get access to inventory.

0:42and working at GoCardless, having experience as a software engineer with companies like Stripe and Twilio, I just couldn't imagine that it was going to stay the way the travel industry was. You have a few providers that have been around for a very long time, very antiquated. I mean, the first one was started in the 50s. It's a company called Sabre in the US. And, you know, it's the antithesis to Stripe in so many ways. You can't sign up online. You don't know what the product is. You don't have access to a test environment. The key insight for me was that I just couldn't see that it was going to stay that way.

1:19I think it would be a really informative thing for our audience to hear you talk about exactly what Duffel does and like what the world could look like in a post Duffel rollout estate. Yeah, absolutely. So the way we kind of view the world post-Duffel, as you said, is where you have a lot more access point for travel. So we're seeing both on the corporate and consumer side, you know, banks and spend management platforms are getting into travel and trying to have a deeper integration with their respective ecosystems. But our view is that this should be happening in a lot more places. I won't mention their name because they're still under wraps, but there is a company that's doing something really cool in the sales space and CRM space.

2:11And there's so many ties here between sales. You're going on a trip, you're traveling to visit customers and being able to see what could the itinerary be? Where can I stay? What are the flights to get there? And we're seeing every day now more opportunities for that to happen. So it's flights, hotels. We do flights. We do hotels. We have car rentals coming up and we will have another product as well later this year. So we try to cover as much of the travel spectrum as possible. So your journey with Duffel has been quite an interesting one. You started eight years ago. You've been through YC. You went through COVID, which as a travel company was, I imagine, a big deal.

2:51And we want to hear about some war stories and perhaps near-death experiences there. But tell us about the moments that have really challenged you as a founder. We started the company at the back end of 2017. We went for YC summer 18. At that point, we had signed up a few airlines as integration partners. We only had access to test environment on their side. So we get read-write access to their reservation systems, essentially. But at this stage, it was just test. And that's what got us into YC. And then we raised a seed round with a fantastic set of investors. Blossom Capital here in the UK. Jan at Index, Michael Moritz, Kim Aventures, and obviously YC as well.

3:33And came back to London, started building the company, and everything was looking great. We had more airlines, I think, on the back of the announcement that were interested to work with us. 2019, raised Series A and Series B from Benchmark and Index. That went quite quickly. we launched the API with live access in November 2019 around the same time we announced the series B and three months later there was not a single plane in the air so I think that was probably challenging moment number one we had a fairly small team at the time we had these new products and the market had entirely disappeared which is quite rare it is just nothing nobody wants to buy and during we ended up going pretty deep on flights and working with more established travel companies that had been around the industry for a while and were looking to replatform.

4:29What was difficult and we only realized later is that the signal we were getting on the product was just very slim. Like it was very hard to kind of extract any kind of meaning. Were we doing well, not well? The companies were trying to sell to when we started the company, new entrants that we were talking to Brex on the back of YC, a few banks and Challenger banks and neobanks. And they all gone away. They had all gone away during COVID. And so we ended up coming out in 2022. I still had a lot of capital, but the product just wasn't ready for these customers that were starting to think about travel again.

5:12and we had grown the team quite substantially at the beginning of 2022. So we had to size it down, which was challenging as a first-time founder. That experience was quite painful. All the while trying to figure out what is the product and where are the customers and having raised the capital that we had, there was definitely an expectation that revenue traction needed to accelerate in some ways. That has been tough. The post-COVID period was quite tough and realizing also how far away we were from the targets because we had kind of built for two years vacuum. It seems like there was some sort of dissonance between like, that's an insane round benchmark and index backing something where you shortly have to find that the demand you thought there was going to be kind of wasn't really there.

5:59Like, did something change in that period? Like, why wasn't there demand for this? We connect to suppliers, so to airlines and now hotels and other travel suppliers. During COVID, it was fairly easy to continue selling to them. But I guess the real driver of revenue for us are platforms that leverage our API, leverage our infrastructure in order to sell to their customers. We work with the likes of Rippling or Engine or Builds on the consumer side, and they are selling travel to their customers. And what we found is post-investments, none of these companies were interested in selling travel because of COVID.

6:40So while we continued going quite deep on the supply side and being able to acquire inventory, essentially, we just had no one to sell that inventory to. It's quite an interesting moat that you're building by building that inventory because it's not like building a sort of unified API for products and APIs where they're publicly available. like these are slow moving incumbents who I imagine are pretty tech backwards and who you know it can take years or certainly months to you know get even sandbox sort of access yeah yeah is that where you see your sort of defensibility long long term there's definitely a big piece of it getting access to the suppliers is incredibly difficult our longest airline contract was five years and a bit, I think, to come through.

7:31So a very, very long sales cycle. And eventually you get access to something and you spend another year on top building an integration. There's obviously a lot of optimism around building part of the integration now. I think that's going to trend towards zero over time, but there's still a lot of the integrations we do that are imperfectly documented. You have to talk to the one person in the distribution department of the airline that has been there for the last 55 years and know about this exact quirk that you're running into. So there's a tribal knowledge that's very hard to get to. And that's also a big factor in the defensibility.

8:13Granted, you can get access to the airline in the first place, which is also very hard, as I mentioned. So there's a huge piece of defensibility there.

8:22Steve Domin:Yeah, Stephen, I wanted to ask about to your end customers, the value is presumably having all of the airlines. So it must be quite difficult as a bit of a chicken and egg scenario there where providing access to one is great, but it doesn't really solve the problem that you're trying to solve. So how did you plan for that? How did you think about sort of aggregating the market in its entirety in order to really provide that value? And how challenging was that to sort of line them all up to make sure you did get them all over the line? There are a couple of problems that our customers face when they want to get into travel.

9:00First, as you mentioned, the inventory, the aggregation of inventory, and you can't have missing inventory. You can't be selling travel and saying, we have British Airways and Virgin, but no Delta, no United. So we definitely need to have near complete inventory. In order to solve that problem, we have more than 50 API connections across all verticals, work directly with suppliers. That was one of our strengths, especially on the airline side, as I explained. But we also eventually connected to aggregators, the companies that have been around for a while and that have that long tail of inventory.

9:40Especially in the case where there is no way to connect directly to a supplier. We kind of have to rely on those. And then there are a variety of suppliers across the ecosystem for hotels and car rentals that do work of aggregation, negotiate commercials with a particular hotel or a particular airline. And we connect to some of those as well. So on the hotel side, for instance, we work with Booking.com and Expedia and Priceline in order to get that very broad coverage of inventory. So that's on the inventory point. And then the kind of additional value for our customers is that we are essentially a full stack travel agency.

10:24So we disguised as an API, but we've set up agencies in multiple countries where we get the licenses, the accreditations. and then we're able to do support on the health of our customers, make sure that we and them get paid the commissions we are owed by the suppliers, settle. I mean, there's a lot of gnarly problems that travel companies have to solve and we try to take as many of them away as possible. That's really interesting. So is the vision almost that you're like the last travel agent the world will ever need? I mean, I love that. I love that, yeah. Okay. available APIs yeah exactly and you mentioned booking.com a couple of others there's some serious incumbents Amadeus, Sabi already mentioned Airbnb even like what are some of the strategies you're using or hacks we love hacks on riding unicorns you're using to circumvent or beat these players sometimes some tension because in the relationship because we there are partners and sometimes competitors, so supply partners and sometimes competitors.

11:35Where we operate, we don't feel as competitive in the sense that we really look at the API-enabled connectivity as our primary differentiator. And we focus a lot on developer experience. The other players in the space have different priorities. Sometimes they have a consumer business and a B2B business. already there's a bit of tension there with the b2b distributors like if you are a bank leveraging expedia's content expedia is somewhat competing with you but also they are a partner enabling the bank so there's always a bit of an interesting dynamic there and then a lot of these solutions are white label often so you get a portal that's pre-made and we found that for the customers we're going after technology companies it just doesn't work they need something that they can customize and deeply integrate into their ecosystem yeah super interesting and

12:35Steve Domin:balancing that sort of competition apart and thing must be challenging you mentioned that you you worked at go cardless and you're i think head of platform engineering there and that was the role prior to founding duffel and there are certain companies that become like talent factories where there's lots of people who come out of them who go on to start exciting businesses and i'd put go cardless in that category there are successful companies that are successful in their own right but don't really have that what is it that certain companies do that makes them these kind of talent factories and any founders listening what should they be doing to attract and develop exceptional people but also train them as future entrepreneurs?

13:19I feel very very lucky to have joined Bocales at the time I did. It was back in 2012, beginning of 2013. Small team but the density of talent there was exceptional. It obviously came from the founding team Matt, Yoroki, Tom, I guess Ari as well that was there quite early. They put a lot of emphasis on hiring very very smart people and just giving them big problems in the company. Lots of autonomy. I think for the first few years, we just ended up hiring more and more of those great people. And for a company that young, the interview process was quite extensive. I mean, after like a few calls, I spent an entire day there and I met everyone in the company.

14:06And yeah, I think at the time it was 10, 12 people, you know? So it was quite a long day of just being grilled by everyone on technical, on non-technical matters. And I think really seeking what is that person that's in front of me is doing that is non-conventional, you know, or they are going a bit further than most people would go in order to do that. It doesn't have to be something in payments or fintechs. They didn't really care about that. But yeah, I think there was a very strong emphasis on getting smart talent in the door. if we wind back the clock why do you think you are an entrepreneur the two ingredients that I think put me in the position that I have today and it came back I think throughout I want to say career but even before that is one having ideas about things you know wanting to things to be different and I think it applies both in terms of oh I want travel to work in a different way and let's go and start a company but also oh I don't like the way this is done at go cardless or we could do this new product or this new feature or any kind of role or when I was younger I started a non-profit when I was 15 I was setting up basketball tournaments and where we got some funding to set up a music studio and you know there was always they're like okay I want to do something let's try to make it happen and then the second ingredient was gathering people in order to do that and you kind of need to convince people to get on the journey with you whether it's a team in a company or when you're starting a company or when you're setting up a music studio you know like someone has to learn how to use Cubase in order to do recordings and things like that and I think that they are the two things that meant that I'm doing what I'm doing today.

16:02Super interesting. Do you think if the world looked like it does today when you were 18, do you think you would have become a sort of a tech entrepreneur sooner in terms of the tools that are available to you to start companies more easily? I mean, I grew up in a mid-sized city in France, 50 ,000 people. Things we take for granted today, knowing about how to start a startup, being able to work at a startup, getting that first time experience, knowing what Aker News is and Y Combinator or like they were very far away from me. And that's why I said I was lucky. I kind of got on the train with GoCardless.

16:40I knew about some of these things before because I was interested in them, but they were very far away from me. I think now there's so much content available just on YouTube, but even like startup school coming to a lot of cities now and startups are more distributed. So I feel like I probably got opportunities to do something sooner. YC probably would have been the thing, honestly, that would have gotten me on that path. If I had known about it earlier, I'd probably tried.

17:10Steve Domin:Do you think there's a counter to that though? Obviously, I have to have touched on some of the AI tools that make it easier to build things and launch things. But obviously, the cutting your teeth in a culture and an environment must have been incredibly valuable. So, you know, are we risking also people sort of doing one employee companies but not really getting the full education of actually being involved in a company? I guess the key question here in my mind is what does the future ends up looking because in some ways I wouldn't trade the experience that I had at GoCardless seeing the company grow the team grow how we went about thinking about processes and product and engineering, velocity, and all of these things, I think were hugely valuable.

18:00So while a part of me wished that I had known about YC earlier, I also kind of liked the way things happened and I wouldn't trade that off. But the counter in my head is that how many of these things are going to be needed by the new generation of entrepreneurs? Because maybe it won't be one person startup, But, you know, if it's 10 people, 20 people, then the experience that you need to get is very different. Yesterday or the day before, I was reading this article on Hyperliquid. I think they're doing 900 million in profits with like nine people on the team. And I was like, wow, you know, if this is the structure of companies going forward, who cares about getting the experience of scaling a sales team to 200 people?

18:49It doesn't matter anymore. So obviously a very different business. I'm being a bit facetious. But I think that's one of the key questions. What's going to be the makeup of teams going forward and what's going to be necessary? Steve, on that note, what is the shape of Duffel going forward? How are you guys using the tools that now are available to teams like yours to make things more efficient, help you move faster? As many companies, we are over-indexed on engineering today. And it's starting to spill out outside of engineering. And now there are more attempts by various teams to build agents for themselves in order to kind of do things while they're doing something else.

19:32On the engineering side, I spend a lot of time with our engineering managers. And my view is that we need to put a ton more emphasis on building the software that builds the software. The end product that we're building and integration to British Airways. as I said at the beginning of the show, isn't going to be as valuable, I think, in the long run. Assuming someone has the documentation and they can grab that person at BA that knows all the gnarly details, writing the code is going to become fairly trivial. But writing the initial integration is only 20 % of the problem. It's how do you maintain it?

20:11How do you scale it? How do you, that the integration continues improving? How do you upgrade the version of the API when BridgeShower is releasing a new version? And there's a lot now that can be done almost in a kind of a self-improving loop where we don't really need humans to be there and to say, oh, BA released a new version. Just go on and make the upgrade. Like ideally, that happens automatically. And we won't go as far as building foundational models. You know, I think that's kind of crazy. But the layer that sits on top of these models can do a lot of the orchestration between our monitoring service or exception tracking service, our logs, query database, tying it all together and deciding what's the next best thing we can do on the code base.

21:04The fastest company that get there will have a massive advantage over everyone else. Super interesting.

21:11Steve Domin:It's great to hear. Yeah, I mean, the way teams are being built now is changing so quickly. And every company is going to have to use AI. But as you say, it's not necessarily just about foundational models. It's about using it to drive efficiencies for the problem that you're solving. Steve, we're going to move on to our final two questions. The first is our future Unicorn prediction. So if there was another company that you'd seen maybe early stage that you thought could go all the way, who would they be? I would pick they. It's a CRM, a new CRM, AI native CRM. I've really fallen in love with that product.

21:49It's really changed how I interact with our sales team, how I can look at feedback. Maybe for context, I don't know if you're familiar with the company, so it might be helpful for me too. How do you spell it? It's a day.ai. Is it day? Day.ai. Day.ai, yeah. They are started by the former CPU of Upspot, if I'm not mistaken, and recently raised their Series A. The product is really good. It's essentially AI native CRM. So everything you want to do, moving an opportunity somewhere in the pipeline or getting an update about a customer, all is done through an agent or multiple agents if you want. This is the first time really where I found the conversational interaction with the product to be really natural and make sense as the default.

22:40So, you know, you're in pipeline review and you ask questions. Well, now I can almost do a pipeline review before the pipeline review meeting, or I can just go in today, talk to the agents. They're quite good at the proactive side as well. The agent will send you a briefing every day and say, you might need to help these people on these opportunities because things are slowing down. And yeah, I've been really impressed by the product.

23:04Steve Domin:Awesome. Thank you. And then our final question is our dinner party guest game. So if you could have dinner with any three people, who would they be? So there was one person that's immediately in my head I would love to have dinner with. And that's Travis Kalanick. If I ever get a chance, I've been a huge fan of the product first as a user from the very early days. And then I think seeing what they've built. And the reason I'm particularly interested is because so much of what they had to build was physical world infrastructure. and that's so true of Duffel as well. There's the bridge between deploying a city and we have to spin up an agency in the country and you have to support travelers on the go and obviously it's a lot of similar motions and I think they've executed so brilliantly that I'd love to take a few pages of this book.

23:58Yeah, 100%. Maybe it won't fit the requirements but it would be someone from some kind of luxury brand like Hermes or Chanel. you know like craft brands and i don't know if i would pick someone that's like in the weeds or someone like an executive or someone but definitely go build a list there and find a person that has that kind of okay let's put it all on the brand it's all about the craft not high volume let's do whatever we do really really well it would be very interesting and then the final person and That dream has sort of been fulfilled, but it would be an astronaut. And I was lucky. I was at an index event and they had the Garrett Reisman.

24:43I happened to sit next to him at dinner and I would do it again. I actually made me, I keep that on the list because there's some like fantastic stories that they have.

24:52Steve Domin:Yeah, nice. Well, you've got travel, brand and engineering. So I think that pretty much sums up. yeah exactly awesome well thanks for sharing that steve it's really great insight and yeah the whole episode is full of little bits of golden knowledge around what it takes to build a company sticking to a problem that you really want to solve fighting through difficult moments but ultimately you know ending up at a really interesting place with the right partnerships the right tag everything in place so congrats on everything you've done so far and thanks for sharing the journey and we wish you all the best with it going forwards thank you for the kind words and thank you for having me that's it for this week thanks very much for listening to stay up to date with the latest episodes please follow or subscribe on your favorite podcast platform we also have a newsletter called reading unicorns which is another great way to get every episode direct to your inbox please tell your friends about it and we'll see you on the next episode