Operating notes.
SEP 2026A schedule is not a dependency map. →
Dates answer “when.” Dependencies answer “what has to be true first.” A date attached to unresolved identity, access, source data or ownership is not a plan; it is a guess with formatting.
SEP 2026Go/no-go should be boring. →
The final launch conversation should confirm evidence, not discover whether the team is ready. Readiness should already be observable before the meeting starts.
SEP 2026Write down the decision, not just the meeting. →
Meeting notes preserve conversation. A decision record preserves the choice, rationale, owner, assumptions and conditions that would justify reopening it.
SEP 2026RAID is useful only when it changes behaviour. →
A risk register matters when a risk changes sequencing, scope, ownership, monitoring or a stakeholder decision. Otherwise it is administration.
Technical notes.
SEP 2026200 OK is not the outcome. →
A successful request proves that one layer accepted something. It does not prove identity, profile data, eligibility, journey behaviour or customer outcome.
SEP 2026SDK vs REST is usually an ownership question. →
Client behaviour should usually originate where the behaviour happens. Server-authoritative outcomes should originate where the business transaction becomes true.
SEP 2026Events describe moments. Attributes describe state. →
“Customer booked Paris” and “customer has three lifetime bookings” are both useful, but they are different shapes of information and should not be modeled interchangeably.
SEP 2026The system boundary is where I start debugging. →
If browser, web app, API, CRM and marketing platform are involved, I want to know exactly what leaves each system, what the next one expects and what evidence exists at the handoff.
SEP 2026Design the exit before the message. →
For any triggered journey, ask what business event should make the message irrelevant. The exit condition is part of the business rule, not an afterthought.
Noteworthy.
A small list of material I think is worth revisiting. Books link to an official publisher or author page; the practice items open into my own notes.
Thinking in Systems by Donella Meadows
A useful reminder that fixing the visible output is not the same as changing the system that produces it.
The Goal by Eliyahu Goldratt
Good for thinking about constraints, flow and why optimizing every local team can still make the whole system worse.
The Checklist Manifesto by Atul Gawande
I like the central idea: expertise does not remove the need for simple controls when complexity exceeds human memory.
High Output Management by Andrew Grove
Still useful for leverage, operating cadence and thinking about management as a system rather than a personality.
API documentation before implementation
Read identity, limits, error semantics and the data contract early enough that they can still change the architecture.
Synthetic test users
A reusable set of test profiles across identity, consent and lifecycle states makes end-to-end validation dramatically easier.