The Two-Arc Roadmap
Part of the Planning Week chain · previous: Work Placement Landscape · next: The Function Service Pack
The two-arc roadmap is a one-page format for a workforce function's plan: an arc band above eight quarterly columns, seven lanes as side tabs, a bar per lane for steady-state work, one star per lane for the completion event, and value markers on a companion timeline that shares the axis. It matters because it is the strategy on a page in a second view, so a program owner and a leader read the same content, and because its dependency order stops the climb being taken in the order that is easiest. The page produces the roadmap lane table (T8) and the first-quarter front-load list, and is worked in a session with Wiki:Packs/Two-Arc Roadmap (CP-WFM-015); decisions D-15 and D-16 are taken on it on Day 4 afternoon of a planning week.
The format
The chrome is the standard integrated-roadmap format: side-tab lanes, bars named on the bar rather than in a legend, a dotted bar for work approved in direction but not yet dated, and a companion benefits view on the identical axis so a reader can flip between "what is done when" and "what value lands when" without reorienting. The extensions:
| Element | Rule | Why |
|---|---|---|
| Arc band | Two arcs above the quarters: Arc 1 consolidate and standardize (Level 2 to 3 by hand), Arc 2 automate (Level 3 to 4 by engine); four quarters each | The page reads at the strategic-arc level and the execution-quarter level at once |
| Seven lanes | Definitions and comparability · process and capacity cycle · governance and decision rights · partner contracts and supply-seat portfolio · capability ontology · simulation and placement engine · organization design and role evolution | The capabilities the estate builds; a lane is a what, a program is a how |
| One star per lane | The lane's completion event for the arc, written as an observable sentence identical to the outcome cell on Strategy on a Page | The star is the test that the two views have not drifted |
| Bar fill | Shaded by maturity level; switches from dotted to solid where the engine takes over | The hand-to-engine transition is visible on every lane |
| Value markers | On the companion timeline: a plain circle for a one-time benefit, a circle with a tail for one that accrues | Value-unit milestones, not only savings; no figure travels without its source |
| A cross-cutting row | The agent program is a row beneath the lanes, not a lane: it delivers into lanes 1, 2 and 6 and owns none of them | A layer that cuts across lanes is drawn as one, so its owner is not asked to own three lanes |
A room may rename a lane and should resist adding one. Where "level" appears on this page it means the maturity model's five levels, except in lane 2's documentation milestones, which use the L0–L3 documentation levels of Process Decomposition (L0–L3).
The dependency order
The lanes are drawn in an order, and the order is a dependency, not a start sequence. Every lane starts in the first quarter, because the standard is built concurrently with the climb and never serialized behind it (principle 6 on Operating Principles for Work Placement); what the order fixes is that a later step's milestone is never dated before the milestone it depends on. The generic order is:
- definitions → comparability → governance → contracts → automation → engine
The chain's order is stated on The Case for Operating Model Planning: data definitions before comparability, comparability before governance, governance before contracts, contracts before automation, and all of them before an engine can run. Vendor Ecosystem Restructuring Agenda carries the reasoning beneath it, tier by tier: pricing depends on measurement, measurement on a shared definition of the work, the definition on consolidation having made one instrument affordable, and everything on an explicit statement of what each supplier — and on this chain, each node — is for. This page cites both and restates neither. One divergence is worth naming: the Restructuring Agenda builds the placement engine in parallel with the decisions rather than after them, and this chain agrees, which is why lane 6 runs from the first quarter. What the chain dates into Arc 2 is not the engine's build but its running at the annual horizon, which waits on a year of hand-run gate output pages (The Placement Gate). What this page adds is the mapping from steps to lanes and the test a room applies to any proposed date.
| Step | Lane(s) whose star it is | The test on any date |
|---|---|---|
| Definitions | 1 definitions and comparability; 5 capability ontology (the supply-side definitions) | No comparability milestone before the definitions register's first version |
| Comparability | 1 (one instrument across nodes; the first matched-tenure comparison) | No contract milestone that cites the instrument before the instrument is live at one node |
| Governance | 3 governance and decision rights; 2 process (the standards committee is a governance body) | No automation milestone before the committee and the acceptance gate exist |
| Contracts | 4 partner contracts and supply-seat portfolio | No automation clause before the comparability clause; no renewal signed before the shared instrument |
| Automation | 2 (the first automated components); the agent row | No engine milestone at the annual horizon before the first catch rate is published |
| Engine | 6 simulation and placement engine | The engine's build runs from the first quarter; its running at the annual horizon waits on a year of hand-run gate output pages (The Placement Gate) |
Lane 7, organization design and role evolution, sits beside the order rather than in it: its star is dated against governance and automation, and its content is line work rather than a program. That the order in which a system's rules and information flows are changed matters more than the size of any one change[1] is the whole argument for drawing the order on the page rather than trusting it to be remembered.
Lanes and programs, reconciled
A lane is delivered by one or more programs (Function and Program Charters); lanes 4 and 7 are line work with no program wrapper, and lane 3's content is line work that the program office administers. A lane with no program under it will not move; a program with no lane above it has no star to aim at.
| Lane | Delivered by | Owner (a seat, never a name) | Dependency step |
|---|---|---|---|
| 1 Definitions and comparability | the definitions-register program; the migration program as forcing function; the agent program (definitions in use) | the definitions owner | definitions → comparability |
| 2 Process and capacity cycle | the program office (the standards committee); the agent program (first components); the platform-instance program | the placement seat as steward; the portfolio planning seats for the cycle | governance → automation |
| 3 Governance and decision rights | line work; the program office administers the protocol and the forums | the function's leader; the placement seat for the change-control board | governance |
| 4 Partner contracts and supply-seat portfolio | line work on the renewal calendar; the definitions program supplies the comparability clause | the placement seat; the contracts seat | comparability → contracts |
| 5 Capability ontology | the capability-ontology program | the placement seat; the definitions owner in the paper phase | definitions → engine |
| 6 Simulation and placement engine | the long-term capacity engine extended into simulation; the ontology program for the supply side; the hand-run gate | the placement seat; the portfolio planning seat for the platform | contracts → engine |
| 7 Organization design and role evolution | line work with the people function; the platform-instance program supplies the sunset condition | the function's leader; the people partner | governance → automation |
A seat is a post on the org chart, a director-graded or manager-graded box with its scope and its sunset condition; a supply seat is a partner capacity block, and the two are always qualified. Owners are seats because the roadmap outlives any one holder. The agent row's own sequence, the short-term forecasting loop first with planners approving, the question register second, integration adapters third, the monthly and annual plan fourth, scheduling and real time last, is on A Roadmap Pattern for Agent Teams in WFM, with the one measurement to make before planning on it, what a planner-hour covers today; the roadmap carries the row and cites the pattern. One distinction belongs here and is made once for the whole chain, because the two objects are easily crossed: the agent ladder governs the function's own planning work and sits in Arc 1, while the landscape's automated band governs the bodies of work the function plans for and is an Arc 2 transition (Work Placement Landscape). Arc 1's "by hand" qualifies the placement gate, not automation generally. Linking a strategy to its operational plan through one shared set of objectives, so that the plan is reviewed against the strategy rather than beside it, is the closed-loop management the lane table is a workforce instance of.[2]
The first-quarter front-load
The first quarter after the week is front-loaded so that the foundations every later quarter depends on are in place before the second quarter opens. The front-load is a list, not a bar: each item with an owner (a seat), a date inside the quarter, and what it depends on. Six items are foundations and every later star depends on them; the rest are what the estate can staff. Decision D-15 is the room's ratification of the list: strike what cannot be staffed and say which quarter absorbs it. The dependency order is ratified with it, not voted.
| # | Foundation | Owner (seat) | Depends on |
|---|---|---|---|
| 1 | The definitions register's first version, the six core definitions with ramp, signed jointly | the definitions owner | D-12 |
| 2 | The standards committee chartered and sitting, with its queue | the placement seat (chair) | D-06 |
| 3 | The intake door and the executive register live; the weekly report derived | the methodology steward; the program office | D-09, D-14 |
| 4 | The capability ontology on paper; the cardinality test run on returned pool data | the placement seat; the definitions owner | pool data returned |
| 5 | The agent team on one book, planner gate on | the portfolio planning seat that owns the book | D-13 |
| 6 | The platform-instance program chartered; its first book chosen | the standardization lead | D-13, D-14 |
Two external clocks are read before the list is dated: the first partner renewal (lane 4) and the commercial function's pricing milestone (lane 1's value marker); both are actions for the program office, not decisions for the room. D-16 closes the block: the validation circuit's names and its date, after which nothing on the page is settled until it has run. Sequencing dependent work across owners and holding the critical path is the discipline the front-load borrows;[3] Program and Portfolio Management carries it in full.
A director's view is a filter
No director's roadmap is separately authored. Each is the master table filtered to that director's rows (the lanes they own, the programs they carry, the front-load items with their seat) and re-titled, because the moment the roadmap forks into per-director versions, the versions drift. The filtered view is section 13 of The Function Service Pack, "your roadmap lane," inherited from the master and filled with the function's own bars beneath the enterprise star.
Worked example
The chain's worked example function ratified its roadmap on Day 4 afternoon, Thu 23 Apr 2026. Its eight quarters run Q2 2026 through Q1 2028: Arc 1 from Q2 2026 (the front-load quarter, ending Tue 30 Jun 2026) to Q1 2027 (star quarter, ending Wed 31 Mar 2027); Arc 2 from Q2 2027 to Q1 2028 (star quarter, ending Fri 31 Mar 2028). The seven lanes L-001 to L-007 carry the seven programs PG-001 to PG-007 as reconciled above — two on L-001, none of their own on L-004 and L-007 — with PG-007, the agent program, on the cross-cutting row. D-15 was taken with twelve front-load items, the six foundations non-negotiable; eleven were dated inside Q2 2026 and one, title harmonization across the three heritages, was struck as unstaffable and absorbed by Q3 2026. D-16 set the validation circuit to close Fri 8 May 2026. The value timeline carried three markers, none with a figure: a one-time marker at Q3 2026 for the first recorded decline with its price, a tailed marker from Q1 2027 for consolidation of the partner blocks by body of work, and a tailed marker from Q2 2027 for the first serviced body of work reaching the automated band with a published catch rate, which is the landscape's transition and not the ladder's. The agent row was drawn against the state the function was already in: the planning-loop team at pilot, the scheduling and real-time teams at beta, no team at production. What the room decided was the next rung and the clone onto a second book, both inside Arc 1. The Standard's expiry, Fri 30 Apr 2027, was drawn as a dated marker on lane 2.
The artifact this page produces
The roadmap lane table (T8), one row per lane, eight quarter cells, a star for the completion event, plus the front-load list. One filled example row:
| ID | Lane | Owner (seat) | Programs | Q2 2026 | Q3 2026 | Q4 2026 | Q1 2027 ★ | Q2 2027 | Q3 2027 | Q4 2027 | Q1 2028 ★ | Dependency step | Value marker |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| L-001 | Definitions and comparability | the definitions owner | PG-001; PG-004 as forcing function; PG-007 (definitions in use) | register v1: the six core definitions with ramp; signed jointly; read by the first report | one instrument (quality score and CX index paired) live at one node | one data core, one plan of record; the headcount perimeter agreed with the finance function | ★ one question, one answer: the instrument across all nodes; the first matched-tenure comparison run | governed metric records with lineage in every report | explainability records | comparative scoring at matched tenure | ★ every output carries its evidence | definitions → comparability | Q1 2027, one-time: a comparison across nodes that both sides accept |
Produced in a working session with Wiki:Packs/Two-Arc Roadmap (CP-WFM-015); the filled set is part of blueprint v0.1.
What would change this
The page's claim is that a dependency order drawn on the page keeps the climb from being taken in the easiest order. The observation that would overturn it is a roadmap, reviewed at the Arc 1 star, on which a lane's star was met ahead of the step it depends on and the result held: an automation clause signed before a comparability clause with no dispute in a year, or an engine run on definitions that were never registered with no cannot-run outcomes. One such case would be an exception; two in different lanes would mean the order is a preference and the page should present it as one.
How this connects
- Previous in the chain: Work Placement Landscape — the transitions the lanes carry
- Next in the chain: The Function Service Pack — where a director's filtered lane goes
- Same content, first view: Strategy on a Page — the stars are its outcome cells
- Defers to: Vendor Ecosystem Restructuring Agenda (the dependency order and its reasons) · A Roadmap Pattern for Agent Teams in WFM (the agent row's own sequence and the planner-hour measurement) · Program and Portfolio Management (integrated planning and the critical path) · Technology Journey from Level 2 to Level 5 (what arrives at each level; the bar fill's meaning)
- Feeds: Function and Program Charters (the programs under the lanes) · Technology Migration Plan for a Workforce Function (the build order is the technology lane's dated form) · AI Agent Program for a Resource Optimization Center (the cross-cutting row) · Planning Week for a Workforce Function (the readout section)
Maturity Model Position
Arc 1 is the climb from Level 2 to Level 3 on the WFM Labs Maturity Model™ (Level 2: Structured Workforce Management to Level 3: The Automation Layer) and Arc 2 the climb from Level 3 to Level 4 (Level 4: Planning in Distributions); the bar fill's shade ramp is the five levels and its dotted-to-solid switch is the Level 3 to 4 boundary. What lies beyond the eighth quarter is Level 5 Pioneering and is drawn as a direction, not a commitment. Four scales on this wiki use the word "level"; Planning Week for a Workforce Function states which is which. This page uses the maturity Levels 1–5, except in lane 2's documentation milestones.
See Also
- Planning Week for a Workforce Function
- Build the Rule, Not the Destination — why the roadmap is a sort order and not the strategy
- Standardize Before You Automate — the six standardizations Arc 1 builds before anything is handed over
- The Agentic Journey Map — the automation arc from the servicing side
- Executive Issue Register — where the front-load's dates are tracked after the week
References
- ↑ Meadows, D. H. (2008). Thinking in Systems: A Primer. White River Junction, VT: Chelsea Green. Chapter 6, "Leverage Points: Places to Intervene in a System". ISBN 978-1-60358-055-7.
- ↑ Kaplan, R. S., & Norton, D. P. (2008). "Mastering the Management System". Harvard Business Review 86(1), 62–77.
- ↑ Project Management Institute (2017). A Guide to the Project Management Body of Knowledge (PMBOK Guide) (6th ed.). Newtown Square, PA: PMI. Section 6.5, "Develop Schedule" (critical path method). ISBN 978-1-62825-184-5.
