There's a version of consulting where the goal is to become load-bearing. Undocumented decisions, clever abstractions only the author understands, and an ongoing relationship that exists because ending it would be frightening.
It's lucrative and we think it's a bad trade. Clients who can leave tend not to.
What we do differently
- Boring technology by default — TypeScript, React, Postgres, the things your next hire already knows
- Architecture decision records written as decisions are made, not reconstructed at the end
- Runbooks with the specific commands, not a description of the commands
- Your engineers in code review from week one, not week ten
- A named owner on your side for every system before we finish
The test
At the end of an engagement we ask a client engineer who wasn't in the build to deploy a small change to production using only the documentation, with us in the room but silent. Whatever they get stuck on is the real gap, and we fix it that week.
It's an uncomfortable hour and it has never once failed to find something.
Why this is good business anyway
Almost all of our work comes from repeat clients and their referrals. The mechanism is simple: a client who owns what we built and understands it comes back for the next thing, and tells a peer.
A client who feels stuck comes back too, but only once, and they don't refer anyone.