Operating model
How I work
Most projects do not fail on effort. They fail because ownership is unclear, dependencies are invisible, and decisions wait on the wrong person. This is the sequence I use to prevent that, and the operating system underneath it.
- 01
Align
Agree what we are actually delivering, and who decides.
Objective
Establish a shared definition of the work, its constraints and its decision path before any schedule is built.
Activities
- Clarify objectives, deliverables and success criteria with the client and internal leads
- Identify every stakeholder and what each one needs to see
- Name the accountable decision-maker for each deliverable
- Surface constraints early: budget, fixed dates, capacity, approvals, localization
Outputs
- Scope definition
- Stakeholder map
- RACI chart
- Constraint list
Tools
- RACI chart
- Stakeholder map
Leadership value
Most schedule slippage starts here. Naming the accountable decision-maker at the beginning removes the single most common cause of waiting.
- 02
Structure
Turn the intent into a plan that survives contact with reality.
Objective
Build an integrated schedule and resource plan that reflects real capacity and real dependencies, not optimistic ones.
Activities
- Build the integrated schedule across all workstreams
- Plan resource against actual capacity, including planned freelance support
- Map dependencies and identify the critical path
- Set the reporting rhythm and escalation route before they are needed
Outputs
- Integrated master schedule
- Resource allocation matrix
- Dependency map
- Risk register
Tools
- Integrated timeline
- Resource allocation matrix
- Risk register
Leadership value
A plan built on real capacity can absorb a surprise. A plan built on best case turns every surprise into an escalation.
- 03
Run
Keep the work moving and keep everyone able to see it.
Objective
Maintain delivery momentum and a single source of truth that teams, partners and leadership all read from.
Activities
- Track production status against plan and surface slippage early
- Run the reporting rhythm: what changed, what is at risk, what needs deciding
- Coordinate across time zones and languages
- Maintain budget visibility against plan
- Protect focus time for creative and production teams
Outputs
- Weekly status report
- Production tracker
- Project health dashboard
- Budget status view
Tools
- Project board
- Production tracker
- Health dashboard
Leadership value
Reporting is not administration. Done properly it is how a team stays able to act without waiting for permission.
- 04
Adapt
Absorb change deliberately instead of silently.
Objective
Handle scope change, risk and resource conflict as explicit decisions with stated costs.
Activities
- Price every change request in days and resource before it is agreed
- Present the trade-off rather than absorbing it
- Reallocate resource across concurrent work when priorities shift
- Escalate on unclear ownership, not only on missed dates
Outputs
- Change request log
- Revised schedule
- Updated resource plan
Tools
- Change request workflow
- Risk register
Leadership value
The difference between a project that flexes and one that quietly degrades is whether change is priced or absorbed.
- 05
Close
Land it properly, then make the next one easier.
Objective
Complete delivery cleanly and convert what happened into something the team can use again.
Activities
- Run final approval and delivery readiness checks
- Close out budget, purchase orders and invoicing
- Hand over assets and documentation
- Run a retrospective that produces owned actions, not observations
Outputs
- Delivery checklist
- Final budget reconciliation
- Handover documentation
- Retrospective actions with owners
Tools
- Delivery checklist
- Retrospective framework
Leadership value
A retrospective without a named owner per action is a conversation. Closing properly is what compounds across projects.
Project operating system
What I hold, on every project.
These eleven areas are the standing responsibilities. Each one has a document behind it and a named owner, so nothing depends on someone remembering.
- 01Scope
- Defined at the start, and re-agreed explicitly whenever it changes. Undocumented scope is the most expensive kind.
- 02Timelines
- One integrated schedule everyone reads from, with dependencies visible rather than implied.
- 03Staffing
- Resource planned against real capacity across all live work, including planned freelance support.
- 04Budgets
- Committed, spent and remaining tracked per workstream, so variance surfaces while it is still correctable.
- 05Risks
- Logged with an owner and a trigger point. A risk without both is a worry, and worries do not get managed.
- 06Approvals
- Approver named in advance, approval windows built into the schedule rather than assumed to be instant.
- 07Dependencies
- Mapped across workstreams and across concurrent projects, because shared resource makes them cross-project.
- 08Change requests
- Priced in days and resource before agreement, so trade-offs are chosen rather than absorbed.
- 09Stakeholder communication
- Proportionate to influence and involvement. Nobody important is surprised; nobody uninvolved is buried.
- 10Escalation
- Route agreed at kickoff, triggered by unclear ownership rather than by a date already missed.
- 11Postmortems
- Structured as what happened, why, and one owned change per theme. Otherwise nothing improves.
Operating artifacts
The twelve documents behind this.
Every artifact below is an original construction built for this site. None reproduces client data, and none is a screenshot of a third-party product. Names, projects and figures are invented to show structure.
Project board
Answers 'what is actually moving right now'. Work sits in one visible state each, so nothing progresses in someone's head. The count per column is the early warning: a swollen review column means approvals are the bottleneck, not production.
Review holds four items against one in production. The bottleneck is approval, not capacity.
Integrated campaign timeline
Shows how workstreams overlap and where they hand off. The value is not the bars but the dependency lines: they make it obvious which date cannot move without moving three others.
Localization starts before production ends. Moving the production end date moves launch.
Resource allocation matrix
Reads capacity across people and projects at the same time. Solves the most common resourcing failure: assigning someone who is already committed elsewhere, discovered only after the commitment is made.
| Person | Taxi | Eats | Retail | Brand | Load |
|---|---|---|---|---|---|
| Art Director A | 40% | 30% | — | 20% | 90% |
| Copywriter B | 20% | — | 50% | 40% | 110% |
| Producer C | 30% | 30% | 20% | — | 80% |
| Freelance D | — | 40% | 20% | — | 60% |
RACI chart
Removes the question 'who decides this'. Most schedule slippage is not slow work but work waiting on an unnamed approver. One accountable name per row, enforced.
| Deliverable | PM | Creative | Account | Client |
|---|---|---|---|---|
| Creative concept | C | A | R | I |
| Production schedule | R | I | C | A |
| Budget approval | C | I | A | R |
| Localization sign-off | R | C | I | A |
| Launch readiness | A | C | I | R |
R = Responsible · A = Accountable · C = Consulted · I = Informed
Project health dashboard
One screen leadership can read in thirty seconds. Status is carried by a written word and a shape, never by colour alone, so it survives a print-out, a projector and colour-blind readers.
Each row carries a shape and a word as well as a colour, so status survives greyscale printing.
Weekly status report
Written for someone who has not been in the room. Leads with decisions needed rather than activity completed, because the reader's job is to unblock, not to admire progress.
Decisions needed
- Confirm final cut length by Thursday — blocks localization start.
- Approve additional freelance day rate — blocks resourcing for week 6.
Changed since last week
- Shoot moved +2 days; launch date held.
- Media plan v3 received.
At risk
Localization window compressed to 5 days. Mitigation agreed: translation starts on locked v4 rather than final.
Risk register
Turns anxiety into a list with owners and trigger points. A risk without a named owner and a defined trigger is just a worry, and worries do not get managed.
| Risk | Impact | Owner | Trigger | State |
|---|---|---|---|---|
| Client approval delay | High | PM | No response in 48h | Watch |
| Freelance unavailable wk6 | Medium | Resource | Not confirmed by wk4 | Active |
| Weather on shoot day | Medium | Production | Forecast at T-3 days | Watch |
Change request workflow
Every change gets priced in dates and resource before anyone agrees to it. This is the single highest-leverage process in campaign delivery: it converts silent scope creep into an explicit decision.
- 1Request logged
- 2Impact priced in days + resource
- 3Trade-off presented
- 4Decision recorded
- 5Schedule reissued
Step 2 is the control. A change with no stated cost is scope creep; a change with a stated cost is a decision.
Stakeholder map
Plots influence against involvement so communication is proportionate. Prevents the two standard failures: over-reporting to people who do not need it, and surprising someone who did.
Production tracker
Follows each deliverable through its actual production states rather than a percentage. Answers the only question that matters before a handoff: can the next step start.
| Deliverable | State | Next step can start |
|---|---|---|
| Hero film 60s | Online | Yes |
| Cutdown 15s | Offline | Blocked |
| Key visual | Retouch | Yes |
| JA subtitles | Awaiting lock | Blocked |
Budget status view
Committed versus spent versus remaining, per workstream. Shows variance while it is still correctable, rather than at reconciliation when the only option is explanation.
Committed sits ahead of spent everywhere, so there are no unrecorded obligations.
Retrospective framework
Separates what happened from what to change, and forces one owned action per theme. A retrospective without an owner is a conversation, not a improvement.
What happened
- Approval sat 6 days
- Two late change requests
Why
- Approver not named
- No impact stated
Change + owner
- Name approver at kickoff — PM
- Price all changes — PM
