The cost of handoffs between teams

Every handoff has a cost.

Yesterday I wrote about handoffs in the context of data. That was partly as a nice introduction to today’s note which I think is even more interesting - handoffs between teams.

Every time you handoff a task to another team you’re paying a cost. A cost in terms of context lost, a cost of managing further dependencies, a cost of distance between the enabler and the outcome.

We see this in data all the time. For example, you might have data generated by a product engineering team, extracted by a data platform team, refined by a data engineering team, modelled by an analytics engineering team, and presented by a business analyst.

The business analyst is so far away from where the data is generated that they have little context on how it was. It’s also been changed so many times on its journey to them that they don’t know who to ask to gain that context. They may be close to the analytics engineers, but have little context on the rest of the pipeline.

The product engineering team is so far away from the value generation that they have little context on why the data is important. Therefore they have little incentive to increase the quality or dependability of the data. They may be aware a data platform team is running some service to extract their data, but those data platform engineers don’t have much more context on the value either - it’s all just data to them.

Reducing these handoffs would help increase collaboration between those who generate the data and those who create business value from it, while reducing costs and wastage.

This isn’t a new idea, which is why many in the software engineering world advocate building product teams. Maybe we just need to add a well-rounded data person to those teams.