Given that we’re data scientists, you might think that engineering or algorithms or logic lies at the heart of what we do, but while those are all important components, they’re simply tools for the job.
Over the years, we’ve learned that data science is a creative process, and the magic that makes it work - the linchpin we place at the heart of everything we do - are relationships.
It’s how you tell data scientists apart
Data science is something of a black box to those who aren’t data scientists. Many business people who outsource it don’t have access to an in-house expert qualified enough to quality-check the logic behind an algorithm. And, if asked to differentiate between a mediocre model and a spectacular one, most people wouldn’t know where to start.
So when you hire a data scientist team, what should you be looking for? Well, chances are you’re already paying attention to the right details: how they manage their relationship with you.
On face value, qualities like timeliness, attentive listening, or clarity of communication seem like nice-to-haves. They’re good qualities in a human but are they necessary in a data scientist? What do they have to do with the quality of their technical work?
You might be surprised by the answers.
It’s indicative of a number of other qualities
For a start, a well-run business tends to attract and retain the best people. Sure, you might find a few geniuses scattered around elsewhere - but if their expertise is trawling through a filter of poor communication, you won’t get the best outcomes - and any data scientist worth their salt is unlikely to stick around in such a company for long.
On that note, be wary of any data scientist who uses their technical ability as a smoke-and-mirrors excuse for poor communication. There’s a quote attributed to Albert Einstein that goes, “If you can't explain it to a six year old, you don't understand it yourself.”.
We’ve seen that to be true in our industry. Typically, if a data scientist is reluctant to explain an issue, they probably don’t fully understand the problem. Whereas if a data scientist is open to questions, and clear in the answers they give, they likely have a strong grasp of how to apply their technical knowledge to the issue at hand.
It’s fuel for brilliance
Data science is the art of helping humans make better decisions, and to do that well, a data scientist needs to know the humans they’re helping.
That means they should take their time and ask good questions. A data scientist can’t rely on experience or engineering methods to get the best outcome. They need to be attentive so they can get beneath the surface of a client’s needs. Otherwise they might overlook vital context and create a brilliant solution to the wrong problem.
At Unai, we also make a point of being attentive to one another. We want to take the initiative with our team members, asking for their insight and noticing when they have issues. That’s part of what makes us a tight-knit team, and it makes us better at what we do.
It’s equal relationships that work wonders
Of course, you can’t verify our internal culture very easily. Unless you want to spend a week listening into our internal comms, you’ll have to take our word for it. Though of course, when it comes to trust, we can’t demand it from you.
Clients quickly notice how attentive we are to them because it’s possible to demonstrate attentiveness. Trust is another matter. You can’t demonstrate or demand trust, it has to be earned.
And ironically, we can only really prove it by the decisions we make that customers never see.
Because it can be such an unknown quantity, data science can make some companies nervous. Sometimes they might insist on micromanaging us.
And it’s here that we insist on pushing back.
Many data scientists are not brave enough to place expectations on their customers. But we’ve learned that good, long-term relationships cannot be built on a power imbalance. While we’re technically working with a supplier-customer dynamic, to do the best work, and to create the best partnerships, we have to relate on an equal footing.
In the long term, it’s how you get the best outcome possible
For this reason, we try not to pitch potential clients a long list of promises (we will do this and that, etc). Instead, we will detail the part we will play and the part our potential client will play.
Not only that, we’ll also provide feedback and hold our customers to account when they overstep the boundaries.
It might be unusual, but it works.
Ultimately, you’re not just outsourcing expertise, you’re hiring a team. One that does the hard work to prioritise communication. One that goes the distance even when it’s not noticed, and one that takes the time to listen, to care, and to create a strong team culture that will generate the best work possible.
After all, that’s the right kind of home for any expertise, data science included.
To find out more about making the most of your data science project, check out our five common mistakes to avoid and don’t hesitate to get in touch.