‘If you want to learn a language, you need to speak to a human’: Babbel branches out into travel bookings

Babbel is offering language-learning holidays through the acquisition of LingoVentura, a Berlin travel startup

Markus Witte, co-founder of Babbel, a language-learning app (Courtesy Photo)

Markus Witte can read Flaubert in French, but he can’t order his breakfast. Such are the perils of learning a language without ever speaking it.

“Eventually, if you want to learn a language, you usually speak to a human being,” says Witte. It's an unusual position, given Witte is a co-founder of Babbel, the highest grossing language-learning app. Did he really just say that to learn a language, you need to do it in person?

“Yes of course! There’s no denying that. But that doesn’t mean the app doesn’t make a lot of sense."

From busuu, to Babbel to the UK-based Memrise, language-learning apps have proved big business. DuoLingo claims to have the most users, at 300 million. Babbel doesn't release user numbers but says it signed up 1 million US users in the last year. Thanks to its subscription model, Babbel is the top grossing language app. Its US sales alone rose 140 per cent in the second half of 2017. Paying users rose by 40 per cent in the second half of 2017 in the US, up from 40,000 last August, overtaking Europe.

In a bid to help people get out and practise their skills with real humans, Babbel has announced that it will start offering language-learning holidays. It has acquired Berlin startup LingoVentura, which handles bookings for 200 language schools in over 100 cities and countries.

Babbel users will be able to find these language-learning holidays through a separate part of the app, Witte says. Backstage at TechCrunch Disrupt, a tech conference in Berlin, he added: “We don’t want to interrupt the learning journey, because it’s not an advertising tool, but we will make sure that if you want this kind of service you can find it.”

Babbel has come a long way since it was founded in 2007 by Witte and Thomas Holl, two software engineers with no experience in languages. “When we started out trying to build a better language learning website in 2007, pre App store, we had no idea what we were getting ourselves into,” Witte says. “We were three arrogant tech guys saying, this problem is already solved, how hard can it be?”

A free iteration of Babbel went online in January 2008 and attracted investors alongside startup funding from the European Structural Fund. But Witte quickly detected an important problem: “You couldn’t learn a language with the thing we had built.”

The engineers had built a vocabulary trainer and given it a user-friendly interface. They learned two things: “One, that you would never learn a language off that, and two, you would never make money off it.”

So they rebuilt the app from scratch and started asking people to pay. Babbel charges between €12.99 (£12) and €8.25 (£7.35) a month, depending on how long users sign up for, unlike its biggest competitor, DuoLingo, which is free.

“DuoLingo is free, and that’s their problem,” Witte says. “We’ve been there. It’s hard to make money with a free product. I’m not talking about becoming filthy rich, I’m talking about paying your freaking bills! And making sure your people can pay their bills.”

At the same time, Babbel rethought its approach to language learning. Rather than thinking of it as a tech problem, they began to think of it as a human problem. “You can’t substitute the hundreds of years of research and experience that there is in language learning with technology without even looking,” Witte says. “We did that, so I am totally guilty of it. We also made the humble move to say, we were wrong.”

Babbel now employs 700 staff, including many language experts. “Tech matters - all the founders are tech people,” Witte says. “We’re totally excited about machine learning and all the fancy stuff out there, but in time we realised that these are just tools. We can be excited about them, but it’s the way you use them that matters.”

Babbel originally set out to build an adaptive language learning tool, with exercises put together by machines. But machines aren’t as good at teaching languages as humans and they can throw out random phrases. “Randomness is what as a language learner is pretty off-putting,” Witte says. “So we use stories.”

Babbel is using tech to experiment with adapting stories without breaking them. The idea is for two people to be able to invest different amounts of time in learning a language and still have a consistent human story to guide them. The tech is used to adapt the programme to the speed of the learner.

The missing piece of the jigsaw for Babbel is the ability to help the user learn to speak, or to move beyond reading Flaubert and into ordering breakfast. But while Babbel has embraced language holidays, it has no plans to move fully offline and into language schools, Witte says. “We don’t want to get into the language school business, we are not disruptive, we don’t want to eat anyone else’s lunch,” he says. “Many of them are very small, family run and offer a very personal service. To replicate that personal element is very tough at scale and might not be what we’re good at.”

Instead, holidays offer a chance for app-users to try out simple conversations that they have practised on the app. Witte remembers using Babbel to learn a few phrases in Turkish before going to a restaurant to order a doner kebab. “They laughed at me, of course they did, but they appreciated the effort,” he says. “It felt good, and that’s the kicker that keeps you going. If you never speak the language, it’s hard to learn.”

He adds: “The app can bring you pretty far, but machines are not human and never will be. If there is one thing that makes us human, it’s language.”