How To Turn Around A Failing Data Team

Tales From Consulting

A few years ago, I heard the same observation from several data and business leaders.

“The default state of data teams is failure.”

This was in the early 2020s, when many data teams felt like they had ballooned in size. Everyone wanted to be data-driven and cheap money made it easy.

Fast-forward to 2025 and the landscape looks very different. Companies are running leaner. Many have intentionally shrunk their data teams and, in some cases, lean more on external partners instead of adding headcount.

As a consultant, I’m often brought in when a previous team has disbanded or when leadership wants to turn around a struggling data environment. Across these engagements, I’ve seen recurring patterns, root causes that explain why some data stacks and teams fail to deliver, and what it really takes to bring them back on track.

This is based off of my talk at Big Data London.

Picture Source - Thank you Szabolcs Magyar

Root Causes of Data Team Failure

There are many challenges data teams and leaders face. But it’s not just about picking the wrong technology. Many data leaders find themselves leading data teams with minimal coaching after being an IC. Others get pulled into being a catch all team where they have to manage not only the reporting but automated processes that might be better suited for a different team while also having to figure out how to lead AI initiatives for their companies.

All that said, here are some common causes of data teams failing.

Lack Of Ownership

Lack of ownership can show up in many forms. I liked the way it was described by

Alex Ewerlöf

in his article 6 Archetypes of Broken Ownership. In which he references that ownership depends on three elements: mandate, knowledge, and responsibility. If any one of these is missing, true ownership breaks down.

Consider a leader who has the power to make a decision (mandate) but lacks deep understanding (knowledge) and isn’t accountable for the consequences (responsibility). They make a call without consulting the team.

The result?

• The team delivers the project but morale sinks

• People stop raising concerns because they feel their input doesn’t matter

• Attrition follows as talented engineers look for healthier environments

Other ownership gaps signs:

• Key assets, tables, dashboards, critical scripts, become orphaned when the original owner leaves

• Technical leads are responsible for data pipelines but can’t influence upstream data quality

• Analysts are tasked with metric definitions but can’t enforce consistent usage across departments

The truth is, like many things, lack of ownership by itself doesn’t usually cause a data team to fail. Instead, it’s the build-up of multiple issues.