Lessons from BlaBlaCar

This blog post from BlaBlaCar shares some great lessons - not just for implementing data mesh, but for any platform team looking to build products for internal stakeholders.

In particular I like how they talk about building generic products. It might take a little more upfront effort, but the solution can then be reused many times without further effort.

As I wrote yesterday, this was one of the realisations I had when coming up with data contracts. If the contract described the data well enough we could easily build generic solutions for everything from backups to access controls to anonymisation and compliance.

I also liked how they favour flexibility over hard rules. You might feel something is best implemented centrally and more generally, but that’s not always possible. Rather than block that team that wants to go alone, encourage and support them. If it proves successful you can centralise later.

I like to use the word autonomy when discussing both of these areas. In a data mesh, or in any decentralised organisation, you want to support local autonomy so those closest to the problem are solving it in a way that provides the best value for the business. Keep an eye out for any gatekeeper or bottleneck that makes that difficult or slow and remove it.