Blocker Matrix - Fix What Unblocks the Most Flows
Tier by flows unblocked, not ticket age
Cert week: integrated TEST is red, every team claims their ticket is P0, and FIFO plus loudest voice decide the order. You clear easy single-flow bugs while entitlement still blocks registration, detail, and critical-path complete. Launch standups need leverage, not ticket age.
Here's what this piece covers: why FIFO fails under deadline, what the 2026-08-01 blocker matrix measured, priority tiers and focus order, the named pattern, and three habits for Monday publish. Same math as spread triage - rank by contexts (here: flows) cleared.
Part of the Field Notes: agentic delivery series.
1. The leadership problem
Without flow mapping, you fix cosmetic tickets while the critical path stays broken. Phase 2/3 SIT cannot prove team → seam → journey until tier-1 blockers clear.
Wrong: "Prioritize blockers."
Right: "Aug 2026 matrix: P0/P1/P2 by journey leverage - recommended focus order for cert, not FIFO."
The anti-pattern is loudest-voice prioritization - random fix order under deadline.
2. What the matrix shows
Derived 2026-08-01 from the integrated-environment status deck into docs/cubsit-blocker-priority-matrix.md. Method: each deck row linking a ticket to a blocked feature/flow = one link; dedupe per distinct flow name.
| Metric | Value |
|---|---|
| Unique blocking tickets | 27 |
| Total blocked flow links | 40 |
| Top ticket | CUB-1066 - 5 flows |
| Tickets blocking 2+ flows | 7 |
3. What I built: tiers and focus order
Tier 1 (3+ flows): CUB-1066 (5), CUB-1054 (4), CUB-1062 (3).
Tier 2 (2 flows): PTC-4786, PTC-5231, GLSCM-4686, CUB-1065.
Tier 3: single-flow blockers - still ordered by surface (critical path before cosmetic).
Recommended focus:
- CUB-1066 + CUB-1054 - entitlement / warranty and pending-registration paths ripple across product registration and critical-path complete (9 combined links).
- CUB-1062 - promotions across listing, detail, and cart surfaces (PDP / PLP / Cart as technical labels).
- Multi-flow tier-2 tickets, then single-flow - prioritize default address/email on complete among singles.
from collections import defaultdict
def build_blocker_matrix(deck_rows):
# Each (ticket, flow_name) once - leverage = distinct flows blocked
links = defaultdict(set)
for ticket, flow in deck_rows:
links[ticket].add(flow)
return sorted(
((t, len(flows)) for t, flows in links.items()),
key=lambda x: (-x[1], x[0]),
)
def tier(flow_count):
if flow_count >= 3:
return 1
if flow_count == 2:
return 2
return 3Cross-cutting clusters matter: fix the capability (extended warranty, pending registration, promotions), not one deck row in isolation. Same idea as registration/CRM handoffs - the seam is the unit of work.
4. Named pattern: blocker matrix
Blocker matrix - tier tickets by flows unblocked, not age or volume; publish leverage order for cert week.
| Artifact | Who uses it | Decision it supports |
|---|---|---|
| Priority tiers | EM, program | What ships Monday |
| Focus order | Integrators | Sequence across teams |
| Reverse lookup (flow → tickets) | QA / SIT | Triage a red journey |
| Cluster notes | Leads | Fix once, clear many rows |
What broke: deck drifts and the matrix goes stale. Unticketed blockers (service not yet integrated, SSO not ready) still hold progress - list them so FIFO theater does not hide them. Declaring "launch ready" without flow green criteria recreates parallel truths.
5. What I'd do as a lead
Three habits - steal one this month:
- Publish the blocker matrix Monday; rederive when the status deck updates. Stale leverage is worse than no matrix.
- Assign one integrator to cross-ticket flows (registration / entitlement path), not per-team silos racing FIFO.
- Define "flow green" before declaring launch ready. Ticket Done ≠ journey proven - same bar as integration chain tests.
Further reading
- Field notes: agentic delivery - series intro
- Leverage cousins: Spread triage · Chronic regression
- Prior: Engineering delivery dashboard
- Related: SIT 63 workflows · Green APIs, broken handoffs
- Next: Token compression and the Index Protocol
- Planning anchor: blocker priority matrix · 2026-08-01
