Capstone · Overview

The Worked Capstone

One initiative, documented twice — once by a lead who reasons from evidence, once by a lead who reasons from plausibility — with a reviewer's annotations explaining every difference in terms of forces, evidence and consequences.

Mission tie-in: the course's promise is judgment, and judgment is only visible in artifacts. This capstone is where the whole course becomes one thing you can show someone.

What the capstone is

Lessons 0039–0050 taught the initiative-leadership artifacts, one per lesson, using a single running case: a retailer's returns-resolution initiative. Lessons 0051–0060 then designed the system from scratch — subdomains, model, boundaries, state, invariants, components, contracts, data, style, release — on the same case.

Both sequences produce artifacts. Neither shows you the whole thing at once, with the reasoning that runs between the artifacts, the decisions that changed under evidence, and the numbers at the end. That is what the capstone is: the complete packet, worked end to end, with its costs and its misses visible.

It exists because the same fourteen artifacts can be produced at two completely different levels of quality, and the difference is not visible from a distance. Both packets here are professionally worded, internally consistent, and complete-looking. Only one of them could survive contact with evidence. Learning to tell them apart on sight — in your own work first — is the capability this course is for.

How it joins Phase IV to Phase V

The two phases are usually taught as separate skills — “leading an initiative” and “designing a system”. They are not separate. Each initiative artifact is the input to a design decision, and each design artifact is only defensible because of the initiative evidence behind it. The capstone's job is to make those joints visible.

How to use it

The order matters. Reading the strong packet before writing your own produces a good imitation of someone else's reasoning; writing first produces your own.

1. Build yours first. Use your own initiative if you have one, or the returns case if you do not. Blank versions of all fourteen artifacts are in the template pack. Lesson 0050's capstone exercise gives a 90-minute working-session version if you want a first pass quickly; the full packet is a weekend.
2. Self-review before the reveal. The capstone assessment has a self-review pass to run before you look at the worked version — that comparison is the most informative one you will get, and you only get it once.
3. Read the strong packet. Read the connecting prose, not only the tables. Follow the “Traces to” line under each artifact.
4. Read the weak packet. Then write down the three sentences in it you would have accepted without question a year ago. Almost everyone finds some.
5. Read the reviewer notes, which put the two side by side per artifact and name the rubric dimension each difference touches. Finish with the ten-step comparison and the self-review checklist at the end.
6. Score with the rubric, twice: once on your own reading, once after the annotations. The gap between the two scores is your next practice.
7. Turn it into evidence. The portfolio guide covers anonymising a real initiative, the executive summary, the technical-review version, and how to narrate your own contribution in an interview.

The strong packet

Read it → The complete initiative and system-design packet: fourteen artifacts, every one traced to an outcome, assumption, driver, risk or rollout decision. Two decisions change under evidence, two dissents are recorded with the conditions that would prove them right, three options are compared and two rejected in writing, and the closeout reports one outcome met, one missed by 0.6 points, and the packet's own weakest piece of evidence.

The weak packet

Read it → The same initiative, documented as a competent-but-junior lead would: outputs framed as outcomes, assumptions stated as facts, a to-be flow with no current state, a flat backlog organised by layer, non-functional requirements that no design could fail, a C4 “container diagram” that is a technology inventory, a status column any caller can set, one option presented as a decision, a launch to 100% behind a flag, a weekly status reporting velocity, and a closeout that ships and moves on.

It is deliberately not a straw man. Every page in it would be approved in a great many organisations, and several of them contain things the strong packet also has — a feature flag, a dashboard, an accessibility item, an ADR. The differences are not presence and absence; they are whether each artifact can be wrong.

Comparing them

The reviewer notes → pair the two packets artifact by artifact. Each pairing quotes both versions, explains the difference in terms of the forces at play, the evidence available and the consequences that follow, and links to the rubric dimension it is scored on. Judgments about taste are excluded on purpose: if a criticism cannot be phrased as “this artifact does not let a reader make decision X, and here is what happens when the evidence changes”, it does not appear.

The page ends with a ten-step routine for comparing your own packet, and a 34-line self-review checklist you can run before anyone else sees your work.

Scoring, templates and portfolio