7.5 Agile Project Controls and Artifacts
Key Takeaways
- Agile control comes from transparency and frequent inspection (information radiators, events) rather than formal change control boards
- Prioritization techniques: MoSCoW (Must/Should/Could/Won't-have), value-vs-effort matrix, and Weighted Shortest Job First (WSJF = Cost of Delay / Job Size)
- Definition of Done is a team-wide quality standard for every increment; acceptance criteria are story-specific conditions set with the Product Owner
- Agile-friendly contracts (Time and Materials, fixed-price-per-Sprint, Money for Nothing/Change for Free) replace rigid fixed-scope contracts
- The Sprint Retrospective is the primary continuous-improvement mechanism; formats include Start-Stop-Continue, Mad-Sad-Glad, 4Ls, and Sailboat
How Agile Controls Work
Predictive projects control change through a change control board (CCB) and formal change requests. Agile projects achieve control differently — through transparency, frequent inspection, and continuous adaptation. Change is not gated; it is welcomed into the Product Backlog and reprioritized. The exam reward is recognizing that an agile answer rarely says "submit a change request to the CCB" — it says "add it to the backlog and let the Product Owner prioritize."
Information Radiators
Information radiators are highly visible, low-effort displays that broadcast status so the team reduces meetings and reports:
| Radiator | Purpose |
|---|---|
| Task/Kanban Board | Status of every work item by stage |
| Burndown Chart | Remaining work in the current Sprint |
| Burnup Chart | Completed work against total scope |
| Velocity Chart | Velocity trend across Sprints |
| Cumulative Flow Diagram | WIP and flow across all stages |
| Impediment Board | Active blockers and their owners |
Principle: Radiators embody transparency. If a question contrasts "detailed weekly status report to executives" with "a visible team board," the agile-preferred answer is the radiator.
Backlog Prioritization
MoSCoW
| Category | Meaning |
|---|---|
| Must have | Non-negotiable; release fails without it |
| Should have | Important; a workaround exists |
| Could have | Desirable; nice to have |
| Won't have (this time) | Explicitly deferred this release |
The second "o" in MoSCoW is just filler; the real categories are M, S, C, W. The exam often tests that W = Won't have this time (deferred, not deleted forever).
Value vs. Effort matrix
| Low Effort | High Effort | |
|---|---|---|
| High Value | Quick wins — do first | Big bets — plan carefully |
| Low Value | Fill-ins — when convenient | Avoid — not worth it |
Weighted Shortest Job First (WSJF)
SAFe's economic prioritization: WSJF = Cost of Delay ÷ Job Size (duration), where Cost of Delay = User/Business Value + Time Criticality + Risk Reduction/Opportunity Enablement. The highest WSJF goes first, which mathematically favors short, high-value jobs.
Definition of Done vs. Acceptance Criteria
These are frequently confused — and frequently tested.
| Definition of Done (DoD) | Acceptance Criteria | |
|---|---|---|
| Scope | Team-wide; every story/Increment | One specific story |
| Created by | The Scrum Team | Product Owner (with team) |
| Question it answers | "Is it built to our quality bar?" | "Does this story do what was asked?" |
Example DoD: code peer-reviewed; unit and integration tests pass; coding standards met; documentation updated; no critical defects. Example acceptance criteria (often Given/When/Then): Given a valid cart, when the user clicks checkout, then they reach the payment page; given invalid payment data, when submitted, then an error displays. A story must satisfy both its acceptance criteria and the DoD before it is Done.
Agile Contracting
Fixed-scope, fixed-price contracts fight agile's welcome-change stance, so agile-friendly arrangements share risk:
| Approach | How it works |
|---|---|
| Time and Materials (T&M) | Pay for time/effort; scope stays flexible |
| Fixed price per Sprint/iteration | Fixed cost per Sprint; scope negotiated each Sprint |
| Graduated / incentive | Bonuses or penalties tied to delivery |
| Money for Nothing, Change for Free | Customer may end early and keep unused budget; in-budget scope swaps are free |
Continuous Improvement
The Sprint Retrospective is the engine of improvement, but learning also flows from Daily Scrums (impediments), Sprint Reviews (product feedback), metrics analysis (velocity, cycle time, defects), root-cause analysis, and Lean Kaizen events. Improvement items often become Sprint Backlog work so they are actually done.
Retrospective formats
| Format | How it works |
|---|---|
| Start, Stop, Continue | What to begin, end, and keep doing |
| Mad, Sad, Glad | Surface emotions about the Sprint |
| 4Ls | Liked, Learned, Lacked, Longed for |
| Sailboat | Wind (helps), Anchors (slows), Rocks (risks), Island (goal) |
Rotating formats keeps retrospectives fresh and combats "retro fatigue," but the output is always the same: a small number of concrete, owned improvement actions the team commits to next Sprint.
Control Through Transparency, Not Gatekeeping
The deepest CAPM theme in this section is that agile control is built in, not bolted on. Because the Increment is inspected every Sprint Review, defects and misalignment surface within weeks rather than at a final gate. Because the board and burndown radiate status continuously, a slipping Sprint is visible the same day, not in next month's report. Because the backlog is reprioritized every Sprint, the project stays aimed at the highest-value work as the world changes.
This is why an agile-correct answer to "how do we keep this project under control" emphasizes frequent inspection points, visible information, and empowered reprioritization — not a heavyweight approval chain. The Product Owner controls what is built (value and order); the Developers control how (technical decisions and the Sprint plan); the Definition of Done controls the quality bar. Separating those three loci of control is a frequent exam distinction.
Acceptance, Done, and Release — A Worked Example
Suppose a story reads, "As a shopper, I want to apply a promo code at checkout so that I get my discount." Its acceptance criteria might be: a valid code reduces the total correctly; an expired code shows a clear error; a blank code leaves the total unchanged. The team's Definition of Done then layers on universal quality: peer-reviewed code, passing unit and integration tests, updated documentation, and no critical defects. The story is only Done when it satisfies both — its specific acceptance criteria and the team-wide Definition of Done.
Even then, release timing is the Product Owner's call, since meeting the Definition of Done makes an Increment usable but not automatically deployed. Keeping acceptance criteria, Definition of Done, and the release decision distinct is exactly the kind of layered control CAPM rewards.
In the MoSCoW prioritization method, what does the "W" stand for and imply?
Which statement correctly distinguishes the Definition of Done from acceptance criteria?
Using WSJF, two features have equal Cost of Delay but Feature A is a small job and Feature B is a large job. Which is prioritized higher?
A stakeholder proposes a new requirement mid-Sprint on an agile project. What is the MOST appropriate agile response?