data science

How your data scientists become bottlenecked

Data scientists are worth their weight in gold but they can become ineffective and bottlenecked if they don’t know the customer, the business problem, or the data itself.

You can throw a team of 5 highly paid data scientists at a SaaS churn prediction problem, but unless they can put themselves in the customers’ shoes, understand what levers the business can pull, and explore the data freely, their results won’t be effective.

There are three areas data scientists/analysts commonly lack which hold them back:

  • Customer empathy

  • Stakeholder empathy

  • Product Knowledge

Customer empathy is when you feel the customers’ stories and motivations. You understand how they think and what drives them. You gain this by physically (when possible) sitting down with the customer or users and listen to them. You should be able to answer, “how do my customers measure the value of my product?”

You should also be able to speak the language of the customer, even as a data scientist.

Stakeholder empathy is knowing how to implement your results. Data Scientists themselves are not the ones executing a business change, they are the ones illuminating an insight and giving a recommendation. Therefore, it is extremely important that you know your audience and how they will act upon your data.

This can be the Product Manager that you support, the General Manager that runs the product, or person that you’re ultimately driving the data deliverable to.

Like customer empathy, you sit down with this person and see what makes them tick. You can do this by sitting in work group meetings, all hands, strategic meetings, etc.

Product Knowledge is knowing the ins and out of the product. What are all of the features in the product? What can users do? What is a typical work flow like?

You gain product knowledge by

  1. Actually using the product

  2. Putting yourself through the hard training.

What training do the power users go through? What about the admins? The sales people? It may seem like overkill, but I’ve implemented this as a requirement for me and my team going forward.

You’d be surprised how data scientists are a hammer looking for a nail, but even worse, they don’t know what a nail is. They sometimes do analytics and modeling with zero business context.


It’s too easy to become removed from the product when you’re sitting comfortably in an SOMA office on the caffeinated end of a python notebook.

For execs, ask your team leads when was the last time their analysts got hands on with a customer.

For team leads, create opportunities for your analysts to get out of the office to build empathy.

For data scientists, give yourself a rating 1-5 in each of the three areas. Then create a plan to uplevel +1 in each rating.