Knowledge Transfer Before Moving On
Case study — handing over a platform I'd led the UI of, through structured sessions backed by written docs, a reuse strategy and a clean backlog, so the team could own it without me.
Context
When I moved on from Sky, the team was going to inherit the low-code workflow platform whose UI I’d led (Building a Low-Code BPMN Workflow Platform). A lot of how it worked lived in my head.
Problem
Knowledge concentrated in one person is a risk that only shows up when that person leaves. The specific danger was regression to old habits: without understanding why the editor worked the way it did, the easiest path for the next developer would be to bypass the model and add bespoke code. That would undo the reuse the platform had been built for.
What I did
I ran a series of structured knowledge-transfer sessions, each backed by written documentation the team could return to:
- Diagram modelling and adding a property. The standard behind the editor; where constants and typed models live; the step-by-step recipe; and the one rule that matters most: always update through the model’s API, never mutate directly, or you lose undo/redo and change events.
- Form building. How to choose a drag-and-drop form framework (anything whose layout is structured JSON can work), preferring ones that reuse our own components over fully custom builders.
- The API gateway. Registering endpoints, swapping environments with configuration, throttling (Rolling Out an API Gateway).
- Permissions. Route-level and feature-level checks, and why both the UI and the server must be audited and tested.
- Performance. Every input change executes against the model, so debounce inputs and show saving state.
- What next: a reuse strategy. Reusable form snippets, named sub-workflows for shared processing, and a feature catalogue of tested, approved pieces, so new use cases start strong instead of repeating the copy-paste era.
Alongside the sessions I tidied the backlog so the next priorities were clear and ready to pick up.
Outcome
The team took over with documented processes, recorded sessions, a clear reuse strategy and a backlog they could keep working through. The recipes meant the most common change, adding a new configurable property, no longer needed me.
What I’d do again
- Pair every session with a document. Talks are forgotten; documents get re-read the first time someone actually needs them.
- Teach the “why” alongside the “how”. The rule about updating through the model sticks only if people know what breaks when they don’t.
- End with direction, not just description. A reuse strategy gave the team somewhere to go, not just a map of where they were.
What I’d do differently
Start much earlier. The best knowledge transfer is continuous: documentation written as features ship, pairing on every new area, decisions captured in Architecture Decision Records. A handover should be a summary of what’s already written down, not the first time it’s written.
Related notes
Pair Programming · Design Review Readiness Checklist · Communicating with Stakeholders
- #knowledge-transfer
- #documentation
- #leadership
- #handover
- #case-study
↩ Linked from (2)
…end. Documentation written while the knowledge is fresh is cheaper and better. That lesson fed straight into Knowledge Transfer Before Moving On.
- Pair ProgrammingGrowing
…fastest way to hand over context, and the fastest way to learn what's confusing about our codebase. Related: Knowledge Transfer Before Moving On.
↪ Links to (6)
Related by topic
- Growingmetaphor
Deloads Are Tech-Debt Sprints
Calisthenics Lab - Growing
Digital Culture Is Built from Habits
Career - Growing
DORA Metrics
Platform engineering - Growingmetaphor
Hypertrophy and Skill Phases: Build vs Harden Cycles
Calisthenics Lab