
Too many orgs treat improvement like a side quest: an external campaign, fueled by OKRs and pep talks, bolted onto the real work. That’s the “analytical model” Alisson Vale warns about, optimization pushed from the outside. The result? Stop-start initiatives, context switching, and flow entropy: queues grow, lead time widens, quality dips, and everyone is “busy” but value trickles. If improvement depends on a special project, it dies the moment budgets tighten or priorities shift.
High-performing teams adopt the systemic model: improvement is an intrinsic capability of the system, not a sporadic event. We design for it. That means: visible flow (Kanban boards with aging WIP), explicit policies (entry/exit, classes of service), WIP limits that force finishing, and tight feedback loops (deployment telemetry, incident reviews, customer signals). Improvement is time-boxed, funded, and measured like any other service we provide. The culture shift is moving from “projects to improve” to “a system that improves by design.”
Improvement in the operating model:
(1) Reserve 10–15% capacity for continuous improvement, protected slack, not a best-effort.
(2) Stand up a weekly 30-minute Flow & Learning review: inspect aging WIP, blocked items, arrival vs. departure rate, and one leading indicator (build health or change failure rate).
(3) Run one small experiment per week using OODA Loop: hypothesis, expected impact on lead time/quality, owner, 2-week check.
(4) Codify three non-negotiable policies (DoR/DoD, expedite criteria, handover rules) and make them visible.
(5) Install WIP limits per class of service and enforce with pull, not push.
(6) Publish a team service catalog (what we do, SLAs/SLOs, how to request work) to cut upstream noise.
(7) Instrument telemetry dashboards the team owns. If a metric can’t trigger a decision this week, it’s vanity.
You should see fewer carry-overs, shorter queues, and calmer sprints; lead time and blocked time go down, and you’ll see a measurable drop in change failure rate. Your role isn’t to run a crusade; it’s to architect a system that compounds learning. Start Monday: ring-fence improvement capacity, publish the three policies, and schedule the first Flow & Learning review. Decide, explain, and go, then let the system do the heavy lifting.