Books / Learning comic
Program Management - Shopping Portal Delivery
Theme: Business & Management
This sixteen-chapter sequel turns an online-shop design into a worked delivery programme. Maya connects product goals, people, dependencies and evidence while Asha and Dev remain the two developers, supported by clearly bounded part-time roles. The recurring Desk Lamp travels through capability mapping, ordered stories, acceptance examples, estimates, sprint boards, reviews and retrospective experiments. Product feedback changes later work without rewriting previously agreed acceptance. Provider uncertainty, payment recovery and restoration tests expose the difference between completed stories and a releasable customer journey. The illustrative six-week target becomes a no-go decision, a transparent Week 8 reforecast and a controlled pilot. Final reconciliation accounts for every purchase attempt, separates email from payment state, assigns operating owners and shapes the next backlog. The collection makes programme management visible through connected decisions rather than status colour.
Start reading →View in the interactive libraryAbout this book
This collection follows the delivery of the online shop established in the system-design comics. Maya coordinates the programme; Priya orders product work; Asha and Dev remain the two developers; Ravi and Sam provide coordination, quality and operational support. Ada represents the retailer. Product42, the £29.99 Desk Lamp, connects staff maintenance to discovery, basket, reservation, payment, order status and confirmation. The scenario is illustrative, using records and events to explain delivery practice without turning supporting role names into implementation capacity. The chapters map seven capabilities and twelve feature families into a thin whole-shop release. Stories are vertical behaviours with feature parents, acceptance examples, developer tasks and links to later evidence. ADM-01 supplies a product-entry example: valid data saves, invalid values leave the record unchanged and unauthorised updates fail. Story-specific acceptance is separated from a Definition of Done. Estimates include verification and integration. A conditional roadmap identifies required inputs, dependency owners, needed-by points and closure evidence instead of treating the six-week ambition as a guaranteed date. Sprint 1 selects staff access, product entry, catalogue, typed filters and deployment against fifteen focus-days. An image-key response defect demonstrates why working screens do not prove integration. The developers reproduce, repair and retest it before completing the path. Ada's review then requests an inline price explanation, recorded as FB-01 rather than retroactively changing acceptance. A retrospective schedules shared boundary examples before coding. Sprint 2 uses sixteen focus-days to address that feedback, stock, private baskets, atomic reservation and sandbox payment initiation. Concrete tests cover trusted repricing, cross-session privacy, competing buyers, rollback and stable payment attempt keys. A provider dependency remains blocked until matching sandbox responses are verified; an email reply alone is insufficient. Review adds clearer Pending guidance to the own-order story and keeps all twelve feature families visible. Sprint 3 compares a remaining-work range extending beyond capacity with a feasible selection: verified payment, private order status, durable confirmation and bounded recovery investigations. Early contract checks improve visibility without producing claims of productivity gains. The integrated happy path succeeds, but a crash after provider success leaves the original order unresolved, and a completed backup lacks restore proof. At Week 6, Maya recommends no pilot and Ada agrees. The next sprint preserves scope and safeguards, estimates remaining work and explicitly reforecasts a conditional Week 8 launch. Recovery is repaired around the persisted attempt key, and an isolated drill restores and replays all 127 expected records in eighteen minutes against agreed criteria. The candidate revision then receives integrated, privacy, performance, rollback and readiness checks before a bounded pilot is authorised. Named operating owners and stop conditions support gradual rollout. The pilot's fifty attempts remain accounted for through reconciliation, ultimately producing forty-six paid, three declined and one cancelled without charge; confirmations are tracked separately. The final review records both the schedule miss and the delivered portal without claiming causal business uplift from a small sample. Search feedback, measured performance questions, support ownership and linked lifecycle evidence guide the next backlog, completing a continuous loop of planning, delivery, learning and operation.
Published chapters
- 01 - Map the Whole Shop
- 02 - Turn Capabilities into a Backlog
- 03 - Make Stories Testable
- 04 - Connect the release plan
- 05 - Plan Sprint 1
- 06 - Move the board
- 07 - Demo and listen
- 08 - Improve and plan again
- 09 - Build the purchase path
- 10 - Review the unfinished journey
- 11 - Make the next bet smaller
- 12 - Integrate without guessing
- 13 - Show the evidence
- 14 - Change the forecast honestly
- 15 - Release the thin portal
- 16 - Measure and choose what is next