# Initiative artifact pack Thirteen artifacts for leading a software initiative from an ambiguous brief to a measured, operated outcome. Each file has two halves: - a **blank** template — the fields, with one line of guidance each; - a **completed** example — the returns-resolution initiative used throughout lessons 0039–0060, filled in consistently across all thirteen files. Copy the whole directory into your repository next to the code the initiative changes, or copy one file when you only need one. A printer-friendly rendering of the same pack, with each blank template on screen and each completed example behind a disclosure, is at [`../templates/index.html`](index.html). ## The files | # | File | Answers | Lesson | | --- | --- | --- | --- | | 1 | [`initiative-frame.md`](initiative-frame.md) | Why act, and what counts as success? | 0039 | | 2 | [`assumption-map.md`](assumption-map.md) | What might be false, and what would that cost? | 0040 | | 3 | [`workshop-agenda.md`](workshop-agenda.md) | How do we make a decision with a room full of people? | 0040 | | 4 | [`workflow-and-eventstorming-board.md`](workflow-and-eventstorming-board.md) | How does the work really happen today? | 0041 | | 5 | [`story-map-and-example-card.md`](story-map-and-example-card.md) | What is the smallest useful end-to-end path, and how does it behave? | 0042 | | 6 | [`architecture-driver-table.md`](architecture-driver-table.md) | What must the design be good at, measurably? | 0043 | | 7 | [`c4-review-checklist.md`](c4-review-checklist.md) | Do the diagrams expose the decisions, and does every boundary have a driver? | 0044 | | 8 | [`state-and-contract-review.md`](state-and-contract-review.md) | Which transitions are legal, who owns each fact, and what does failure mean? | 0045 | | 9 | [`option-table-spike-brief-and-adr.md`](option-table-spike-brief-and-adr.md) | Which option, on what evidence, at what accepted cost? | 0046 | | 10 | [`delivery-and-dependency-map.md`](delivery-and-dependency-map.md) | In what order do we retire risk, and who else must deliver? | 0047 | | 11 | [`team-charter.md`](team-charter.md) | How do six to eight people act as one team? | 0048 | | 12 | [`rollout-contract.md`](rollout-contract.md) | How do we expose this safely and get back if we are wrong? | 0049 | | 13 | [`weekly-review-and-closeout-checklist.md`](weekly-review-and-closeout-checklist.md) | What changed, what next, and when is it actually finished? | 0050 | ## The traceability test The pack succeeds by coherence, not by weight. Pick one line and trace it all the way through. A broken link is either unnecessary work or an unmanaged risk. > **outcome → assumption → driver → decision → rollout** Worked on the returns case: | Step | Where it lives | The line | | --- | --- | --- | | Outcome | `initiative-frame.md` | Median time from request submitted to decision communicated: 72 h → under 4 h for eligible domestic returns, with the pilot open by 30 September. | | Assumption | `assumption-map.md` | A3: "the carrier creates a label fast enough to sit inside the request that accepts the return." High impact, weak evidence — a vendor's own documentation. Bought a four-hour spike. | | Driver | `architecture-driver-table.md` | D1: through a 30-minute carrier outage at peak, keep accepting and deciding eligible requests, and create 99% of queued labels within 15 minutes of recovery. | | Decision | `option-table-spike-brief-and-adr.md` | ADR 0003, superseding ADR 0002: persist the label intent in the accepting transaction, drain it from a worker, make the carrier call idempotent on the intent id, expose "label pending". | | Rollout | `rollout-contract.md` | The disable path that ADR 0003 created is rehearsed before every exposure stage — and the first rehearsal found it did not actually work. | The full chain in the lessons runs longer — outcome → assumption → slice → driver → boundary/contract → decision → milestone → test/telemetry → rollout gate → continue/change/stop — and the five-step version above is the short form you can run in a review without notes. Two directions to run the test in: - **Forwards.** Take the outcome. Can you reach a rollout gate without leaving the pack? - **Backwards.** Take any artifact — a milestone, a contract, a meeting on the cadence — and ask which outcome, assumption, driver, risk, or rollout decision it supports. If the answer is "we always do this one", delete it. That is the rule that keeps the pack from becoming document bureaucracy. ## How to use the completed example It is one coherent initiative, not thirteen independent samples, and it is the same initiative worked at greater length in the [capstone packet](../capstone/strong-packet.html). The same numbers appear in every file: 72 hours, under 4 hours, 41% → 31%, 40 requests/second, 800 ms, 50 referrals a day, fraud loss of 2.4 per 1,000, $3.42 per resolved return, 30 September. If you change one in your own copy, change it everywhere — that propagation is the discipline the artifacts exist to enforce. The example deliberately includes things that went wrong, because a packet without them is not a model of anything: - a **four-hour spike that killed the team's preferred option** — and killed a decision that had already been written up and accepted as ADR 0002; - a **falsified assumption** that reshaped the outcome statement (policy was 63% codifiable, not the 90% everyone assumed) and another that reshaped what the result could claim (guest orders were 29% of demand); - an **inconclusive result reported as inconclusive**, rather than quietly read as a pass; - **recorded dissent that survived the decision** — Sam's, on ADR 0003, with the condition that would have proved him right, checked at closeout; - a **changed decision** taken on evidence rather than under deadline pressure; - a **missed target reported as a miss**: contacts fell 23% against a 25% target, 0.6 points short, and the packet says so instead of rounding; - a **guardrail that held on a measure too under-powered to detect a breach**, named as such; - **debt left behind on purpose**, with its interest, its owner, and a repayment trigger that had already fired. ## Related - The narrative version of the same material: [Initiative Playbook](../reference/initiative-playbook.html). - The same initiative worked at full length, with an intentionally weak counterpart and reviewer annotations: [the capstone packet](../capstone/strong-packet.html). - Turning your own filled-in pack into portfolio and interview evidence, including how to anonymize a workplace initiative: [Portfolio guide](../portfolio/index.html).