Lock-in is a spectrum, not a binary
A managed database service with a proprietary query extension is a mild form of lock-in with a clear, bounded migration cost. A fully proprietary serverless orchestration model with no portable equivalent anywhere else is a much deeper form, with a correspondingly higher exit cost if the relationship sours.
Treating all lock-in as equally bad leads to either paralysis (avoiding every useful managed service) or, more commonly, accidental deep lock-in because nobody was tracking where on the spectrum each adopted service actually sits.
Price the benefit, not just the risk
The honest version of this analysis prices what the proprietary service is actually saving you — engineering time, operational burden, capability you could not easily build yourself — against the quantified exit cost from the earlier exit-cost modelling exercise.
Frequently, the honest number shows the proprietary service is worth the lock-in, decisively. The point of running the analysis is not to always conclude 'avoid lock-in' — it's to make the trade-off a deliberate, documented decision rather than an accidental one.
The decision that actually matters: can you see it coming
The real risk isn't lock-in itself — it's undocumented, unnoticed lock-in that nobody chose deliberately and that gets discovered only during a renewal negotiation or a failed migration attempt. Track adopted proprietary services explicitly, with their accepted lock-in level noted, as part of normal architecture governance.
This turns lock-in from a silent, accumulating risk into a visible, periodically reviewed portfolio decision — exactly the same reframing that works for technical debt.
The goal was never zero lock-in. It was never being surprised by the lock-in you already have.MNB Strategy Practice
Key actions
- Treat lock-in as a spectrum, not a binary to avoid entirely
- Price the benefit of the proprietary service, not just the exit cost
- Track adopted proprietary services explicitly as a governance artefact
- Make lock-in a deliberate decision, not an accidental discovery