At eBay
I did an in-depth build versus buy analysis, where I spoke with 9 or so vendors in the space, as well as the internal team who would theoretically build an in-house solution. The result of this was a document which I reviewed with all relevant stakeholders (architects on the platform side, budget folks, as well as the experimentation group), explaining the situation, what I learned from each vendor, and a suggested path forward. The choice before us was “get something great now” or “build something more deeply integrated”. This wasn’t an easy choice given eBay had a highly developed experimentation platform and a vendor choice, while great from a flagging perspective, would make the transition from flag → experiment more difficult.
At Axel Springer
Had a chat w/ Tom Halank at Axel Springer as part of a job interview and they had a principled way of looking at build vs buy when doing platform engineering.
First, it’s not static. While it may have made sense to originally do it, times change and the industry catches up.
Second, they had a few interesting questions:
- Can we create competitive advantage by building it ourselves?
- How quickly will this speed up our delivery?
- Are we brave enough to deprecate things? see Technical Debt