3 Questions to Ask your Data Scientists

Data scientists have become some of the most valuable employees any business can have in its IT team. With demand for these skilled personnel far outstripping supply, competing for and securing the right talent - and then keeping them happy - needs to be a top priority for any firm.

But these professionals can only do their jobs well if they're given the right resources to work with, while maintaining a good working relationships with the business units that will supply the source data they use.

However, this may not always be the case, whether it's due to a longstanding attitude that IT teams are somehow separate from the actual operational units, or a lack of understanding among business units about what these teams are capable of. And if this means your data scientists feel undervalued or think their insights are being misused, they won't be happy.

That's why it's important to keep lines of communication with these professionals open at all times and to make sure they have everything they need. To help, here are three key questions you need to be asking your data specialists to ensure they’re properly supported.

Are you in tune with what [insert business unit here] are asking of you?

One of the biggest conflicts between business units and the IT teams is when there is a mismatch between what technology professionals can deliver and what business managers expect. Far too often, what a sales executive thinks is a simple question can turn out to be hugely complicated from an IT perspective, and this can lead to frustration on both sides if they feel they aren't being listened to.

At the same time, it can be easy for discussions between business leaders and IT teams to descend into complex jargon, where neither side is fully able to understand what the other is talking about. It’s best to leave talk of algorithms and data models aside and relate planning back to the real world. If a business unit has a hypothesis based on their own experiences, see what data scientists can do to prove or disprove it with their reporting.

It's also crucial that data scientists are clear with business units on what they can expect, and for business units to ensure they are asking reasonable, practical questions. If your data analytics professionals feel what's being asked of them isn't realistic, make sure they're able to communicate this to the business units, and ensure they can offer suggestions on how to improve.

Is the data you have good enough?

No algorithm can help you deliver the right results unless the information being fed into it is complete, accurate and relevant, and any errors or inconsistencies are likely to skew the results - as the old saying goes, garbage in, garbage out. But are you sure what's being provided to your data science team meets their expectations?

Many data scientists may be so used to spending a large portion of their time fixing and cleaning data, they take it as an unavoidable part of their job. Indeed, one survey showed these professionals spend 60% of their time cleaning and organizing their data.

But if you ask them where the problems are, you may be able to put in place processes to improve the quality of data before it ever reaches your data science teams, saving huge amounts of work for everybody.

Do you have the right context for your questions?

Big data analytics shouldn't take place in a vacuum. In order to get the best results, data scientists don't just need to know what answers their business units are looking for, they also need to know why the questions are being asked and what the rest of the business hopes to use that information for.

A business might want to know why it experienced a drop in sales in the last month. But data scientists won't be able to answer that unless they have some other crucial pieces of information about how the firm operates. For example, how exposed is the business to external factors such as political issues, exchange rates, or even the weather? How resilient is the supply chain to unexpected disruption? What impact will a key competitor's new marketing campaign have?

These are all crucial pieces of context that can change how data should be interpreted; but unless they are factored into the analysis at the right time and to the right degree, your data science team will only have half the picture. Therefore, it's vital to make sure they understand the full context when asking their questions.

Ultimately, conversations between data scientists and the business units they're feeding their insight to need to be clear and straightforward, and both sides must have trust in the other. Without this, a firm won't be able to make the best use of the results it gets, so make sure your specialists aren't feeling neglected and have all the support they need.