Eric Colson (ex VP DS at Netflix and CAO at Stitch Fix) on Why 90% of Data Science Fails—And How to Fix It
Before jumping into the podcast today:
We have exciting news at Delphina! Our business is growing quickly and we are consequently growing our team, hiring actively across 5 roles:
If working with a world class team of engineers and scientists to build AI agents for data science sounds exciting to you, and you might be a fit for one of these roles, please don’t hesitate to apply! Now on to the podcast.
This week on High Signal, we sit down with Eric Colson—former VP of Data Science at Netflix and Chief Algorithms Officer at Stitch Fix—to discuss why so many companies fail to leverage data science effectively and how to fix it.
Why 90% of Data Science Fails—And How to Fix It
Most companies treat their data scientists as a support function—overwhelming them with requests instead of letting them drive impact. According to Eric, that’s a huge mistake.
In this episode, we cover:
Why You Can’t Treat Data Science as a Service Function
Many companies limit data scientists to answering business questions rather than shaping strategic initiatives. The best teams generate ideas—not just execute predefined tasks.The Hidden Value of Data Scientists: Cognitive Repertoires
Data scientists bring unique ways of framing problems, drawing on machine learning, statistics, and classic models like Markov chains, embeddings, and the newsvendor model. Recognizing the right framing often matters more than the specific tools used.Why You Must Scale Experimentation, Not Just Models
Most experiments fail—but that’s not a problem if you run enough of them. Companies that embrace trial and error, rather than rigid planning, uncover the biggest wins.Decoupling Algorithms from Engineering Enables Impact
When data teams rely on engineering to deploy models, iteration slows down. Separating algorithm development from infrastructure lets data scientists experiment rapidly and test new ideas without waiting in line for resources.
🎬 Check out the clip where Eric explains why most companies fail to leverage data science effectively.
Listen to the full episode or find us on
And don’t forget to subscribe!
Catch Up on High Signal
Stay tuned for more conversations with industry leaders about navigating the challenges of data science and AI. If you missed our recent episode with Chris Wiggins, Chief Data Scientist at the New York Times, be sure to check it out!