How I turn priorities into software
Start with the business problem. Pick a useful scope. Stay responsible through delivery and production use.
Decide what matters now
- AskWhat needs to change, who depends on it, and what a useful first result looks like.
- SplitWhat is needed now from what can wait.
- ReleaseCut the work into releases that each deliver a complete, usable workflow.
RavenClip
Unattended publishing first
Then I reshaped the product around its paying customers.
Make the technical choices fit the context
What I weigh
- Performance
- Data correctness
- Failure recovery
- Cost of maintaining the system
A pattern has to solve a problem we actually have.
StoragePal
Cache what customers share
Shared catalog data can be cached, but each customer’s cart stays separate.
SPS
Remove offline support
Every warehouse had internet. Less code to maintain mattered more than scanning without a connection.
Use AI with clear responsibility
AI helps me
BuildInvestigateTest scoped changes
Stays my responsibility
PrioritiesPricingArchitectureAcceptance
On every change
- Write down constraints
- Challenge proposals
- Ask for proof that the result works
Keep the team informed and follow the result
During the work
DecisionsProgressProblems
I make them visible. Specs and notes give the team enough context to carry work forward.
After release
ErrorsStaff feedbackCostsReal use
I stay involved, and these decide what to improve next.
The rule
When the evidence changes, I change the plan.
Have a product to build or a system that needs attention?
Tell me what needs to change