An Interview with Lukas Vermeer

We hear from's Product Owner about developments within data and analytics


Lukas is an experienced data science professional with a background in computing science and online machine learning for real time decision support. A strong advocate of "Evidence-Based Everything", he is forever learning and helping machines do the same. As a Data Scientist at, the world's leading accommodation website, Lukas is helping millions of users find their perfect destination and accommodation by building user-facing products using vast amounts of behavioural data. 

We sat down with Lukas ahead of his presentation at Predictive Analytics Innovation, London, on May 11 & 12.  

How did you get started in analytics?

I started as a Business Intelligence consultant, working for Oracle, supporting many different customers of several of Oracle's predictive analytics products.

After a few years, I realised most companies use data to generate pretty reports, but they are not really data-driven; rigorous scientific experimentation and basic things like hypotheses and control groups are extremely rare.

It seemed to me that I was probably not going to be able to change this culture from where I stood, so I changed jobs and joined one of the few companies,, that does have a very experimentation and data focussed culture already. I love it here.

Are there any recent innovations in the analytics data science community that you see as a ‘game changer’?

The biggest thing that I see changing is company cultures. The analytics and data science techniques we are using are not new, they are actually pretty old, but only recently have companies started realising that they can make a profit out of them.

The dinosaurs and incumbents that are not changing their culture are being outcompeted and will eventually die out. That will take time, but you can already see this happening.

What are the unique challenges facing that you are looking to solve with data science?

There are very few problems that are completely unique to us, and we apply data science to almost all our problems.

What will you be discussing in your presentation?

I'll talk about some of the challenges around running experiments in a business setting. We are not scientists, so we have somewhat different objectives, requirements and problems. This means we'll have to think about which techniques are useful and which ancient rules of Science we want to break and why. 


Read next:

Social TV: Cross-Channel Insights on the ShareThis Platform