We almost bought Auth0 for auth. It was the "smart" choice. We built it instead—in one quarter, for a fraction of the cost. Same with our database platform: "it's too stateful to automate" turned into fully self-service Postgres in 9 months.
I wrote up the ACT framework we use to spot bad assumptions before they become expensive commitments:
- Awareness: What am I taking for granted?
- Challenge: What if the opposite were true?
- Test: Timeboxed spike with clear success criteria
Plus the "Inversion" technique for design reviews (ask what would make the opposite decision correct) and the exact script I use to sell 2-day spikes to skeptical managers.
What assumption is your team holding right now that might be wrong?No one has commented on this post.