Lesson learned: How (not) to expand internationally
Context
Back in 2017, I had been working at Luckey (an Airbnb property management company) for about 1 year. We were experiencing significant growth in France, where we had launched our service in 20 new cities over the previous 6 months, managed a portfolio of more than 1,000 properties, and, above all, our acquisition cost was lower than ever, at under €30 per qualified lead.
We wanted to prove that our model could be applied anywhere, in other countries and languages, so we decided to try our luck in the second largest market in Europe: the UK, starting with London.
Spoiler alert: it didn’t work. Here’s why.
What we thought
The model was working so well in France that it didn’t seem like an issue to launch in the UK. The operations were essentially the same: a local manager to handle offline operations (cleaning, check-ins, maintenance, etc.) and an online team based in France, already composed of English-speaking individuals with tailored tools. It seemed like we just had to translate our communications from French to English, and we’d start generating revenue.
The challenge was to acquire our first properties. Again, it didn’t seem to be a problem: we had built a proven lead generation machine based on a funnel of ads, SEO, SEA, an optimized website, and a trained sales team. We were prepared to invest more if needed to gain a head start.
There were as many Airbnb listings in London as in Paris. We thought we could at least generate half as many leads (which we believed was a pessimistic estimate, considering we needed to grow brand awareness). Half as many leads as Paris was more than enough to get started.
I was Head of Customer Acquisition, and although I was informed shortly before the launch that we were expanding to London without much preparation, I was confident we could pull it off.
What went wrong
It soon became apparent that instead of generating half the number of leads as Paris, we managed only a fifth. More concerning was the cost—it was far too expensive. The advertising costs were significantly higher, and our conversion rate was lower than expected at every step of the flow.
The market was very mature. Too mature. Established players had significantly more funds than us (their recent funding rounds were 10 times higher than ours), increasing the barrier to entry to a level we couldn’t afford to invest in.
Let’s dive into each stage to understand:
Generating a simple visit to our website was costly. The cost per click on Google (CPC) for high-intent keywords like “Airbnb property management” was around £15 (€17.5), compared to €7 in Paris on average, due to fierce bidding competition. I tried adjusting the bids to target only the 3rd position on top keywords while keeping the highest bid for niche keywords, but even after precise optimization, it remained very expensive.
The website conversion rate was at 15% vs 25% in France, for two main reasons:
1. Customers knew there were other options in the market, so they would double-check before leaving their details. In French cities, we had a maximum of three decent competitors; in Paris, perhaps five. In London: 10, very visible, some with billboard ads on the metro.
2. We didn’t appear as local as others, despite our efforts. We had a British-reviewed translation of our website (Oxford English, not Texan ;)), but somehow, visitors could still tell we were a foreign company. That never helps. We underestimated the importance of localizing our website and sales content.
Leads were much less qualified: we rejected 75% of them, compared to 50% in France.
1. The market was so mature that the best properties had already been managed by rental managers for some time.
2. The remaining acceptable leads went directly to the competition due to brand awareness or simple referrals, skipping online searches. We were left with so-so properties, which generated significantly less revenue, and we had already lowered our criteria considerably.
Closing a lead was hard, with a conversion rate at 15% vs 30% in France. For every lead, we had at least three competitors who could demonstrate a presence of more than three years, a substantial portfolio, and a well-tested sales pitch. In these conditions, 15% was actually a good score.
Moreover, the listings generated less revenue. There was a rule in London that people could only rent their primary residence, and only for a maximum of 90 days, compared to 120 days in Paris.
Bottom line, a new customer would cost almost €2,000. In France, it was €100.
And the LTV (Life-Time Value) didn’t cover even a quarter of it. We might have accepted such an effort as part of expansion, betting on the fact that long-term acquisition costs would decrease with SEO and greater brand awareness, but we simply didn't have the budget.
Marketing cost only €10,000 per month across all of France. We had planned €10,000 per month for London alone, but it was absolutely insufficient. That meant acquiring five new properties per month, which was not satisfactory given the local staff we had to pay, not to mention that those properties were already unprofitable, and that’s an understatement.
Even if we had the budget, it would not have been worth it.
And we should have known it.
What we explored
Bidding on generic articles about ‘how to manage an airbnb property’ instead of high-intent keywords: I was already doing this in some French cities, and it proved effective. The cost per click was a lot lower, and often, the leads were less mature, meaning they were less aware of our competitors. We had more chances at conversions. However, the amount of traffic was quite low, and we could only generate two properties per month through this method.
Ads on the Eurostar train. There were TVs in the train between Paris and London, and we thought targeting expatriates was a good idea because they were more likely to be owners frequently leaving their homes. It wasn’t too expensive (still a €10,000 ticket, but what else could we do?). Well, it didn’t work—perhaps one or two leads. Offline advertising is primarily for brand awareness, and you should only focus on it when you’re big enough.
Flyers in mailboxes of targeted properties, including a coupon with benefits → we didn’t receive any calls.
Rental management fairs. Well, only competitors show up at fairs. They’re too expensive for what they offer—still good for engaging local team members, though, but it's for the long-term, never for short-term results.
Buying leads from competitors. But first, they would only sell less-qualified leads (obviously), and second, the price was still very high. Not worth it.
What we should/could have done differently
We should have tested the market beforehand. This means different things:
We should have bid on a few keywords to estimate the cost per click before launch.
We should have made calls with all players to understand their acquisition strategies and business models.
We should have created fake listings on Airbnb to test demand, prices, and estimate revenue.
We should have understood customer profiles through research and calls.
…
We could have launched in smaller UK cities before London. From our experience in France, acquisition costs were much lower in secondary locations (because of lower competition). But we thought that as a foreign company, it was harder to target such local destinations. Plus, we wanted to set up operations in the capital to better manage secondary locations. This choice wasn’t probably the best.
We could have bought a small local player: this would have provided an initial portfolio to understand customers, their needs, and allow us to get our hands dirty and gain confidence before expanding further.
I should have spent a month in London. I was operating from France the whole time, syncing frequently with the London team of city managers (who were very committed), but being away didn’t help—I only spent a weekend there. Despite the team reporting all the information, I would likely have seen other signals and tried other methods. Not to mention that working physically with the team would have helped in finding new solutions that we couldn’t discover just by syncing online.
Lessons learned
Why do you want to launch in this new country? Ask yourself this question and write down the answer. Oh, you may think you know the answer, but you might realize that the reasons are blurrier than you thought, and are perhaps not shared with your associates. Is it because it’s a huge business opportunity? To capture the market before competitors? To position as an international brand for customers and investors? For Luckey, London was essentially a symbol. We’d have Paris and London, the two top tourist destinations in Europe— boom. Of course, it was a bit more than that: we also wanted to prove that our model was replicable, which is a strict condition for a funded startup—but we could have done that somewhere else.
Never under-estimate the market differences in another country. We’re naturally optimistic and think that it’s similar. It’s true in a way: a company wouldn’t launch in a new country if its business model wasn’t applicable. But tweaks are ALWAYS necessary.
Play the devil’s advocate and ask yourself this question: what will potentially go wrong? List all answers. I repeat: list all answers. Test them for real: launch ads, observe, make conclusions. Define “death metrics”: if you reach them, it means you need to make crucial decisions about your business and either close or drastically change the strategy to avoid losing money for too long.
Try to kill your business model, like you would as a skeptical investor. Imagine you’re launching a totally new business, as if you didn’t already have operations in a country, like you were starting from the ground again. It may seem exaggerated, but it helps take a fresh perspective, crucial to uncover market specificities and foster true innovations.
Rome was not built in a day. It takes time to find your product-market fit, and a new country implies a new PMF—a similar but distinct process. To explore the metaphor: it’s like Rome when they conquered Gaul. Victory was not the only thing; they had to partner with local tribes, offer them benefits, grow trade, start building in existing cities, then create whole new cities. After a few decades, it felt like the Romans were always there.
Going back to business considerations, any marketing campaign will take at least three months to show results, and you need to launch several to iterate. It takes at least six months to have a reasonably stable machine, and that’s just for inbound. Brand awareness takes a lot longer. That’s something you need to reflect on when you launch: how long can you survive with little revenue? That ties back to the death metrics mentioned above. In our case, we could afford paying salaries for a year without revenue, knowing that we shouldn't probably have launched.
Epilogue: what happened in the end?
We ended up closing our operations after six months; it was too costly. We said that we would come back to London someday, which we did under a different business model and a different name, five years later. But that’s another story. ■