We are hiring for several data positions.
In this hiring process, we have inevitably identified a very interesting pattern in our candidates' form responses. We ask about what they look for in a company and position, and by far the most common answer is “to be able to make an impact”.
This raised my attention not only because of the ordinariness of such an answer, but also for the fact that I feel really happy to be able to genuinely tell them that the position they had applied indeed offers the opportunity to create a lot of impact for the company.
But of course, for it to be that a common desire, we can assume that this is not the case for most places, and, given the oddity of such a trait, I wondered about what makes Findhotel stand out. What do we have here that allows data scientists and data analysts to make an impact?
After some thinking and a few conversations, I came to the conclusion that data professionals have great opportunities in real data driven companies.
So before talking about FindHotel, let me first present my vision of what defines a real data driven company.
As presented in the title, I personally find it misleading to consider a company to be data driven just because a set of people use data to make their decisions. I have nothing against charts or presentations (not always, at least), but I see that the best use of data is not to prove that some idea is right or wrong. Maybe this was a good definition 20 years ago, when the availability of tools was not comparable to what we have today. But in our time, with our tools, there is a lot more that data can offer.
Using data to make decisions is the first important step. In an ideal scenario you need to let data decide alone, as much and as often as you can. The goal is to give your business the power to react itself to the signals you capture in favor of your mission.
In an analogy where you want to build a windmill (we are located in the Netherlands, after all), the most basic way of being data driven is to measure the most common direction of the wind and setting a fixed windmill against that direction. A further step is to make it easy to rotate, so you can adjust the direction according to wind. Ultimately we want to build a mechanism that changes the windmill’s direction based on the current wind.
Ideally, decision makers are morphing into decision model makers, who apply their business knowledge combined with data techniques to make their components read the environment via a set of sensors and make the best decision all the time.
The techniques used do not need to come from the former decision maker, that person will often work closely with data scientists, data analysts and data engineers. They will work to define what data they need to track, how to connect the systems, and how to translate the business logic into a machine that receives a set of inputs and outputs decisions.
By doing this, a company will scale the impact of its business knowledge by performing more and better decisions, to quickly react to market changes, or even better, to act ahead on forecasts.
Having a bit more clarity on what I mean when I say real data driven, we arrive at the second point, why data scientists have big opportunities in data driven companies.
The most obvious reason, they are directly in touch with decision making automation, so they will be a key element in being data driven, the same way software engineers are a requirement for any tech company.
However there is also one hidden value. In addition to the sentence “to be able to make an impact“, I would say that any profession can create a lot of impact, from customer service to backend engineers, from specialists to high level managers.
In the end we need more than the opportunity to impact, we need to be valued by our impact. Value needs attribution of causality, it needs to know what caused what. One's impact is only properly valued if it is clear that the impact was caused by his actions.
Considering that, a data driven company provides an environment greatly favorable for data scientists, as not only offers a lot of opportunity for them to make an impact, but it is also rooted on performance metrics, making it crystal clear what caused what. From estimating a project potential, simulating the impact during model validation and using a controlled experiment during deployment.
Surely we can find many other reasons to justify the opportunities of data scientists in a data-driven company, but I believe these are enough to prove my point. Having said that, then we move to the next part.
Is Findhotel a data-driven company?
Findhotel is still a few steps away from being really data driven in every aspect it can be, we have plenty of manually set parameters and static decision inputs all over the place, but to be honest, we are in an even better phase in terms of opportunities. We are in the transition. We are fortunate to have found success in a couple of problems that we solved in a proper data-driven way, and now we want to explore the beauty of it in every way we can.
What Findhotel does provide is a supportive environment to transform it into a real data driven company. Here I’ll talk about what I see as the main reasons for that.
The first and obvious one: Good Data
There is no way to extract value from data if you don’t have it easily and reliably accessible.
In the same recruiting process, it is very common to see data scientists with AWS / Azure experience, building complex ETLs in distributed systems. They end up learning because they have to. This is an inefficient use of their skills, as they are not doing what they are good at, and there are much better people to do that. They are called data engineers.
We have a data infrastructure team. A group of people working to facilitate the collection and use of data for the entire company. This seems a bit basic, but I know many companies do not provide it and do not see the real value of having a good data infrastructure. At the moment we have as many data engineers as we have data scientists, and that is a very fair distribution, in my opinion.
I’m very happy to be able to say in case of any kind of data issues: “That is not my problem” as I trust that it will be solved by that team in a time very adequately prioritised.
The second one: Culture
In some way, everyone here is a data analyst. Here, analytics tools such as Looker and Datadog are not exclusive to labeled analysts. We use it as much as we use Slack, and much more than we use email. We are all very intimate with a few metrics, we all know them by name, initials and nickname.
People know their problems in the sense that they know the metrics they want to optimize and their bottlenecks. In this context it is much easier to identify where there is an opportunity to transform a relevant decision making process into a live system that will do it much better, both in scale and in quality.
This behaviour is not random and will not change, we require good analytical skills in all hires, no matter the position, and we keep pushing up their skills once they are in.
Combined with that, we have a lot of freedom to experiment, to try new techniques, to push boundaries. We are anyway in a very competitive market, so any kind of innovation can bring interesting outcomes.
Often people think of artificial intelligence as a highly intelligent being mimicking human behaviour, but to me, the best way to use it is as plenty of small systems each focused on doing really well one single operation, helping us in doing all of our jobs more efficiently. This is how we as a small team can compete with some giants of our market, if they have 100 times our number of employees, we need to be 100 times more efficient, and we are not afraid of aiming at such goals. For Findhotel, not being data driven is not an option.
If you are like us, always aiming for the best, not afraid of facing the hard problems, and of course, want to make an impact, have a look at our open positions and let’s get in touch.