The Data Scientist, sexiest job of the 21st century. Is your corporate pulling power sharp enough?

Anyone wanting to attract data scientists to join the company must be ready to answer some searching questions from their candidates. That’s the advice from ?rsted’s Head of Analytics, Jan Voetmann.

In an Inform Online conversation for Commodity, Jan stresses the importance of preparing what he calls ‘the company story’.

when you as a potential employer approach these people, what works is to have a strong story about why you would like them to join your company. Because that's the first thing they will ask. They will say to you, why am I here? They need a good, strong story that they can believe in.

These data scientists will also demand to know what are the tools you propose to put in their hands, Jan continues. Other key questions from their point of view: What are the data you will give me access to? What's the mountain you will give me to climb? And who will I be attempting to climb with, which team?

That's what they care about: the challenge.

Data scientists are motivated more by the task offered to them than by the company. They love the data and the science. They actually don't care too much about things beyond that. What they want is to be part of a strong team composed of great people they can learn from, be challenged by, discuss with. Clearly they want to have a lot of data. And they want to have some problems to solve which were unsolved until this point, so they can build something new that hasn't been built before.

Therefore he recommends to companies to ask themselves:

why is it important to solve the problems that you have in your company? What difference could solving them make to your company?

Someone with the profile of a data scientist has a mindset and a toolbox that can be applied across almost any industry and geography. When you're hiring a data scientist, therefore, it means the competitive set you are looking at as an employer is not just your own industry or even your own geography. It actually is quite wide. So it's important to also recognize that really good data scientists take very seriously how they invest their time.

They’re quite curious. Their ambitions are not necessarily directed towards career or salary, but they're very ambitious with wanting to create as much value as possible through their hands when working with the data.

Quite often companies haven't really thought through how they will empower and enable data scientists to work and to deliver that value.

For me, if you look at the way a company makes money, there are two ways in which data science can impact that quite visibly. It’s about either increasing revenues or reducing costs. By enabling the company to capture more revenue or making the cost base more effective. You can have a data scientist who builds the most amazing model that can potentially create a lot of value in revenue or contain costs. But unless the model travels to the front line and is implemented, there is no value. The data scientists are fully aware of that.

Especially in larger companies, they will expect to hear how you will make sure that the value they create in fact reaches the bottom line.

Quite often, Jan continues, the initial conversations around something that will become a data science product begin with an individual trader who has a problem to solve and just wants an answer. However, whatever is built may be applicable also to other traders because it's not a support function.

I mean, data science is not a capability that you just deploy to help individuals with their individual decisions. If you really want to harvest the value, you've got to think of it as software, as something built in a way that is scalable. This means that you've got to think of it design wise as something where, yes, you start off with a concrete problem from an individual trader.

..start off with a concrete problem from an individual trader.

But how do you take the core of that, the true crux of that problem and rephrase it in a way that 10 other traders would agree it’s important to find an answer to it? And then that's the problem you solve.

That is interesting work, if you're in data science. Yes, it requires a lot of collaboration. I think this is where people are often surprised --- that a data scientist who is to build something that solves your problem wants to co-create that with you. For this to be a good outcome, a good product, it requires a lot of active collaboration. And it's iterative. It's not a ‘here are the requirements, now go build’. That's not the model that works best or to attract and retain the best people.

I think this is a direction in which most prospective employees are moving. Data science is not just a nice, it's a need.