Outcome-Aligned Teams Around End-to-End Journeys
Organize around journeys, not repos
Component teams optimize locally. Journeys die at the handoff. On a launch program with identity, registration, entitlement, and partner surfaces in play, every repo had an owner - and nobody owned "user completes the path." Org design and SIT Phase 2 seams are the same problem.
Here's what this piece covers: the leadership cost of feature-factory orgs, the Jul 2026 topology for a 20-person scope, how journey teams map to testing, the named pattern, and three habits before your next reorg.
Part of the Field Notes: agentic delivery series.
1. The leadership problem
A "web platform team" becomes a bottleneck without clear outcomes. Feature teams ship; seams fail in TEST. Chronic regression clusters on one boundary while three managers argue whose backlog owns the fix.
Wrong: "We reorganized for agility."
Right: "Four journey teams + Platform Council - what each owns, how testing maps, one PO per outcome."
The anti-pattern is the feature factory org - every team ships; nobody owns the seam.
2. What the program docs show
Jul 2026 topology in docs/team-topology/website-team-ownership.md (also referenced from the 2026-07-28 metrics draft). Scope: 20 people (16 developers + 4 QA), 4 delivery teams, 4 POs, Platform Council as a practice (not a fifth delivery team). External partners (CMS, identity provider, transaction platform) stay outside the 20 - called out explicitly so they are not "someone else's API" by accident.
3. What I built: four outcome teams
| Team | Success looks like | Shape |
|---|---|---|
| Site & Experience | Users discover, complete the primary action, track status | Two pods, one PO, one embedded QA |
| Account & Ownership | Logged-in users manage identity, owned items, history | One PO, embedded QA |
| Registration Service | Registration and entitlement data correct and integrated | One PO, embedded QA |
| Quality Platform & Release Engineering | Releases backed by trustworthy test evidence | One PO, embedded QA |
| Platform Council (practice) | Shared architecture stays consistent | Five part-time leads · no PO |
Registration split that prevents double ownership: Account & Ownership owns the website journey (forms, BFF); Registration Service owns API and data; identity provider owns tokens - Account owns how login feels on the path.
(def outcome-teams
{:site-experience {:owns [:discover :complete :track]}
:account-ownership {:owns [:identity :owned-items :history]}
:registration {:owns [:entitle :warranty-data]}
:quality-platform {:owns [:evidence :release-gates]}})
(defn seam-owner [from to]
;; Prefer the team that owns the destination node on the master flow
(first (for [[t {:keys [owns]}] outcome-teams :when (some #{to} owns)] t)))Testing maps to the same seams: Phase 2 shared workflows in the 63-workflow SIT catalog fail where topology is wrong. Wrong topology → untestable handoffs.
4. Named pattern: outcome-aligned teams
Outcome-aligned teams - organize around end-to-end journeys, not repo folders; journey teams deliver; Platform Council governs shared standards without becoming a delivery bottleneck.
| Artifact | Who uses it | Decision it supports |
|---|---|---|
| Team charters | EM, POs | "Who owns this seam?" |
| Pod split (Site) | Leads | Discover vs complete without two POs |
| Cross-team seam list | Integration | What SIT Phase 2 must prove |
| When-to-change rules | Staff meeting | Revisit when chronics cluster |
What broke: Platform Council that starts taking tickets becomes a shadow team. Split POs on one outcome recreate parallel truths from article 01. Revisit topology when chronic regression clusters on one seam.
5. What I'd do as a lead
Three habits - steal one this month:
- Draw team boundaries at journey seams, not repo folders. If the master flow node has no owner, the org chart is incomplete.
- One PO per outcome team. Avoid split accountability on the same user path.
- Revisit topology when chronics concentrate. Source-area heat from the recurrence pipeline is an org signal, not only a bug signal.
Further reading
- Field notes: agentic delivery - series intro
- Prior: Chronic regression
- Related: Team → seam → journey · Fifty pages, one map
- Next: Engineering delivery dashboard
- Planning anchor: team topology docs · Jul 2026
