FitSD — Quickstart
TL;DR — The fastest path from “what’s this?” to “we’re using it.” Written for one person on a small team who wants a win this week, not a rollout plan.
Adopt FitSD by running one real piece of work through it, not by reading it. Five moves, about a week of calendar time and an afternoon of real effort.
The shape: two gates (is this worth doing? at Gate 1, is it ready to build? at Gate 2), then a Service Acceptance check that it’s proven ready to run.
Before you start (5 minutes)
Section titled “Before you start (5 minutes)”- Pick one real piece of upcoming work: something net-new (a new service, tool, or capability, not a routine change to something you already run). A tool to stand up, a service to replace, a capability to add.
- Grab the forms from
capabilities/solution-development/: the Idea Brief (FSD-FRM-00), Gate 1 (Outline Proposal), Gate 2 (Solution Design), and the Service Acceptance Record.
1. Write the Idea Brief — half a day, tops
Section titled “1. Write the Idea Brief — half a day, tops”One page (there’s a template, FSD-FRM-00). What’s the problem, what would “done” look like, what’s roughly in and out of scope. Don’t design anything yet; you’re just getting enough on paper to decide whether it’s worth a proper look. (A public worked example, run end to end, is in the works.)
2. Take it to Gate 1 — half a day to write, one ~30-minute conversation
Section titled “2. Take it to Gate 1 — half a day to write, one ~30-minute conversation”Fill the Gate 1 form: the driver, a rough value/risk read, T-shirt effort, the options you can see, and the decision you’re asking for. Put it in front of whoever says yes or no. They proceed, park, or reject. That decision, made on a page instead of in a corridor, is the first thing FitSD buys you.
3. Design it at Gate 2 — half a day to a few days, once Gate 1 says go
Section titled “3. Design it at Gate 2 — half a day to a few days, once Gate 1 says go”Design it: requirements, an architecture sketch with security in it, and the bit teams skip — how each acceptance criterion in your SAC baseline will actually be met: documentation, tested backup, access, monitoring, support, cost. If you can’t yet say how you’ll back it up or who’ll support it, you’ve found the problem now instead of in six months.
4. Build it — your normal change process, no extra FitSD time
Section titled “4. Build it — your normal change process, no extra FitSD time”FitSD hands the build back to however you already ship. It doesn’t replace your change process; it just made sure you knew what you were building, and why.
5. Accept it into service — an hour to walk the record
Section titled “5. Accept it into service — an hour to walk the record”Before you call it live, walk the Service Acceptance Record. Each criterion: met, with evidence? A tested restore, not a configured one. An alert that actually fired. A runbook someone other than you could follow. When it passes, it’s a product with an owner, not a thing you’ll quietly carry alone forever.
That’s the loop
Section titled “That’s the loop”You’ve run one service through the front door. If it helped (the Gate 1 conversation went better, or the acceptance check caught something), do it again with the next piece of work. That’s adoption: one project at a time, never a big bang.
Next action: pick your one piece of net-new work and open the Idea Brief template (capabilities/solution-development/FSD-FRM-00). That’s move 1.
Later: the maturity self-check (once built) will tell you where you stand and what one notch better looks like. Until then, the capability cards in capabilities/ show what else FitSD covers and what to point at for each.