Meet the team: Marketplace

We caught up with team-founder Philipp Monreal, to learn a bit more about what Marketplace does, and how they do it.

Life at trivago: Hello Philipp, thanks for taking the time to talk. Firstly, can you give us a little bit of background about the Marketplace team? How long has the team existed and how many people are you now?

Philipp Monreal: It’s a very new team, and certainly a new idea to structure it in the way it is. We’ve been around close to 2 years, we have around 50 people now, broken up into smaller sub-teams. Considering it was a 1-man team 2 years ago, only me, we have grown quite a bit!

Lat: That is some growth in less than 2 years! When you first set up the team, on your own, what was your mission?

PM: Well, the mission from that day to this day is exactly the same, Marketplace is very easy in terms of understanding the general mission. trivago’s marketplace is doing the same thing as anywhere else. We try and connect demand and supply, in a slightly abstract way. In a global way; you have people who buy and people who sell. You have one party interested in buying something, in this case hotels, and another interested in selling something (hotel rooms), and our job is to bring those together.

Lat: So, how do you go about doing that?

PM: Well, from a user-perspective we want to make sure they can view a comprehensive search, so they have everything available and they know what they are getting. We need the user to trust that they know what they will find, and get what they are looking for. As well as that, we also want to make it easy from an advertiser’s perspective; once they join, to start getting traffic and to control getting what they want. Obviously advertisers are very different; a huge OTA has very different needs to a small hotel chain for instance, so regardless of their different capabilities, they should still be able to join the marketplace and create a successful campaign there. So, on an abstract level that is the mission and has been since the beginning and will be for some time. Its simple and logical.

Lat: So, there are really just 3 parties involved; user, advertiser and trivago. trivago is like the middleman keeping everyone happy?

PM: Exactly, the three different participants are all after the same thing – the “transaction”. If a user books something, then that proves they are happy with trivago, because that’s why they came there in the first place. This also means the advertiser sold something, so they are happy with that too. And if the advertiser sold something then they have some money, and can give something back to trivago. So, there are 3 parties involved who all want the same thing, we just all look at it from different angles.

Alex and Yiannis from Markeplace, hard at work.

Lat: So that’s the abstract version, what does a normal day look like for a member of the Marketplace team? What are you looking at when you come into the office?

PM: Well, we are a cross-functional team made up of engineers, data scientists, comms people and product owners etc. We have different roles, but I would say what we all have in common is that we are all a bunch of data nerds.

Lat: That is what I think of when I think of Marketplace. haha.

PM: We are data nerds, because what we do all day is work on algorithms and data-driven decision making, all the people on the team, this is what we’re all really passionate about. I guess, a typical day… hmmmmmm…. so, in general we work in cross-functional teams, but each owns a specific topic, for instance, lets pick something that is easy for everyone to understand “Exposure algorithms”.

Lat: Okay, good pick…

PM: It’s not difficult I promise! Basically, the team worries about which hotel or advertiser do we show which user, and where. Why do you see this hotel on top, why do you see this hotel below, why are certain advertisers above others etc. So, in that cross functional team you have POs, data scientists, Java backend guys. How we work is that basically we try, from a strategic perspective, to identify where  we see opportunities to improve something, where do we see weakness that we need to deal with, things that aren’t working properly. My job as the team lead is to bring to the team a big vision of where we are going, then we have different processes of how we filter that down. We have our yearly mission, quarterly planning and then sprint cycles of 2 weeks. In the sprints, the teams are very self-determined; they get together, discuss what they are trying to achieve, what ideas they have, and how  they prioritize etc. Basically, the team is doing that on their own, so a typical day would be to discuss what we’re going to do on that day, and what we need to finish this week.

Lat: So, the 50 people in marketplace are made up of how many teams? 5? Do those teams have a similar, “cross-functional” make-up?

PM: Yes exactly, in general they work in different topics but they work cross functionally in a very similar way. You can break the Markeplace team almost down into 3 different functions or pillars. We have ‘Market Dynamics’, which is user-facing and based around what do we show where and how the auction model works. Then there is ‘Bidding tools’ which is advertiser-facing, and we ask the question “How best can we enable advertisers to interact with the marketplace?” For example, we have a product that enables advertisers to find their ideal bid. Then, the third “pillar” is ‘Advertiser Data and Metrics’, which provides the foundation for the other groups ie. what data do we need to understand what is going on? What should we be looking at? What data do we need to present to advertisers? So, all these teams operate in a similar way, but they all use different technologies. So, a “typical day” in the team would be exploratory data analysis, which is a fancy way of saying whats going on? Just looking at the data to check if everything is healthy.

Lat: When you say exploratory, does that mean you’re looking at numbers and suddenly you have “a-ha!” moments, where you see patterns you weren’t looking for or..?

PM: That does happen, yes. Most of the time the biggest insights start with “hmmmm this looks funny, then why does it look funny?” Then you start investigating! We also run a lot, a lot, a lot of tests! I am pretty confident that we run the most tests per person for any team at trivago, I think we run as many tests as hotel search with only a quarter of the people. So that is something we do every day. We always try to work in fast and small iterations and always trying to incrementally improve.

Lat: Obviously at trivago, we have regular team events. In the last 2 years, what has been your favourite Marketplace team event? What do data-nerds do for fun?

PM: My favourite team event hasn’t happened yet, but I keep advocating for it. We have a very democratic way of everybody making a suggestion and then voting on it, but my suggestion is never voted for! I would love to do dinner in the dark!

Lat: Okay, last question,; can you sum up the Marketplace team in 3 words?

PM: Ummmmmm… (silence…) Data-driven Decision Making.

Lat: So, data-driven counts as one word?

PM: Oh, cool, so I have another word?

Lat: No, I was being sarcastic, you had four, right? Or does data-driven count as one? Being hyphenated?

PM: You’re the words expert.

Lat: You’re the numbers expert!

PM: Let’s just say, Marketplace can be summed up in 3, or 4 words; Data-driven Decision Making.

Lat: That sounds statistically accurate.

Lee Jones
Lee Jones

Lee Jones is a Conceptual Copywriter and Employer Branding Specialist at trivago.

Related posts