Wiki:Decks/Near-Term Workforce Strategy

Near-Term Workforce Strategy is a deck rebuild kit for the short-term half of a two-horizon workforce strategy: moving a planning function from reacting to three streams of work at once to running all three through one standard. Each section carries one slide's title, subtitle, key message, footer line, exhibit and speaker notes, in the blue house style and de-identified for external use.

Slides	Exhibits	Table slides	Style	Aspect	Generated
13	10	2	Blue	16:9	2026-09-26

How to rebuild

The exhibits carry no text of their own beyond their own data labels. No title, no subtitle, no footer line — those are template slots, filled by whoever builds the deck. Each slide section below gives the exact strings to use.

For each slide: create a slide from the corporate template, paste the Title and Subtitle into the template placeholders, insert the exhibit into the content zone at full width, and type the Footer line into the footer text box where one is given. The template supplies the legal disclaimer and the logo — the exhibits deliberately contain neither.

Two slides are table slides and carry no exhibit. Build the table from the rows given in the section.

Placeholders. Text in angle brackets — ‹like this› — is deliberately absent from both the slot text and the exhibits. Those are the points where a specific program, client or supplier name belongs, and they are filled in locally rather than published here. The exhibits leave clear space at those points so a text box can be laid over the image.

On the exhibits. Every exhibit carrying data is generated deterministically from its source values rather than drawn, so it regenerates identically when the numbers change and no figure is ever retyped into a slide. Values shown are illustrative and in the shape of an observed portfolio.

The argument

The deck has one spine — reactive to proactive — and three beats. Every slide serves exactly one of them.

Beat	The question it answers	Slides
1 · Get ahead of it	Can a migration be modeled before it arrives rather than after it breaks?	5, 6, 7
2 · Know what is happening and what is at risk	Can we see weekly what is true, what is assumed, and what would change the answer?	8, 9, 10
3 · See the whole estate	Can we look forward across business as usual and find hotspots while they are still tradeable?	11
spine	Why any of this	2, 3, 4, 12

Slides

Slide 01 — Title

Title	The near-term framework and the roadmap
Subtitle	‹function› · ‹owner› · ‹date›
Key message	Two horizons, presented together: what changes now, and what the organization becomes.
Footer	none

Title slide — no exhibit.

Speaker notes

Open on the two horizons. The near-term work is not a stopgap that runs until the reorganization lands; it is the target operating model stood up early, and the reorganization ratifies what is already working.

Slide 02 — The portfolio is concurrent

Title	Nine things are in flight at once and none of them arrive in a queue
Subtitle	Migrations, programs and business as usual across the next 26 weeks
Key message	The problem is not any individual piece of work; it is that between six and eight items are live simultaneously against a planning capacity that absorbs about four, with no mechanism to sequence them.
Footer	Illustrative shape of a live portfolio. Every bar above the dashed line is work that gets attention only when it escalates.

The portfolio, and what one team absorbs.

Speaker notes

This replaces an assertion with a magnitude. The lower strip is the point: concurrency, not complexity, is what makes the function reactive. Say the capacity line is a judgment and invite correction — it is more useful challenged than accepted.

Slide 03 — Every transition has failed the same four ways

Title	Every transition has failed the same four ways
Subtitle	Four defects, independently evidenced in every case
Key message	The recurring failures are not analytical: no point-of-departure truth, no template, no intake gate, and no owner of the number.
Footer	Four defects. Every case. Not a skills problem.

Table slide — no exhibit. Four rows, three columns: Defect · Where it has been seen · What it cost.

Defect	What it means
No point-of-departure truth	Contact rates taken from a platform norm or a whole-account average rather than the departing operation's own history for the markets being moved
No template	No incoming plan laid against the outgoing plan before cutover
No intake gate	Requests arriving through several routes at once, producing duplicate and discarded work
No owner of the number	Figures reaching an audience before any of them came out of a model

Speaker notes

The load-bearing slide. Each defect must be evidenced independently in at least two cases or it comes off the page. Keep the third column factual and dated; the moment it reads as blame the room stops evaluating the proposal.

Slide 04 — Three chunks, nine cells

Title	Three chunks. Each needs a way in, a way to model, a way to watch.
Subtitle	Nine cells. One is filled, two are partial, six are empty.
Key message	Migrations, programs and business as usual each require an admission step, a modeling method and a reporting form; naming the nine cells shows that the gap is structural rather than a matter of effort.
Footer	The columns are the argument. Six of the nine cells do not exist today, and the two that are partial are held by one person each.

The short-term strategy on one page.

Speaker notes

This is the slide the rest of the deck hangs off. Every subsequent slide fills one cell. Note that the two partial cells are each held by one person, which is the honest reading of "partial". Say plainly that nothing here asks for more people before it asks for a method.

Slide 05 — Readiness before modeling

Title	Before we model, twenty-six questions get answered or get an owner
Subtitle	Run at kickoff, reviewed at every phase gate
Key message	A fixed instrument run before the model turns undiscovered assumptions into stated, owned and sized register rows; an unanswered item is not a blocker but a quantified exposure.
Footer	Sections are mandatory by archetype — the applicable count is 18 for A, 20 for B, 24 for C and 12 for D. 15 answered, 11 open. The open items are what panel 2 of the weekly pack reports until they close.

Six sections, and the staffing effect of what is still open.

Speaker notes

The full instrument is at Migration Modeling Readiness. The design point worth making out loud: this does not stop anyone modeling. It makes the unknowns visible and sized, which is the difference between a caveat and a decision input. The right-hand bars are what directs analytical effort.

Slide 06 — What changes decides the risk

Title	What changes decides the risk, not what moves
Subtitle	Two questions of every migration: what does the customer meet, what does the agent meet
Key message	Classifying by what actually changes, rather than by the systems involved, determines which assumptions are mandatory before the model is built.
Footer	A new channel planned on the old channel's contact rate is wrong by a multiple, not a margin.

Migrations classified by what changes on each side of the contact.

Speaker notes

Full treatment at Migration Archetypes. The archetype is declared before the model; applied afterwards it explains a failure rather than preventing one.

Slide 07 — The baseline rule, quantified

Title	Online adoption did everything asked of it. The contact rate still did not fall.
Subtitle	Contacts per transaction, pooled across migrated markets
Key message	Walking from the departing operation's own pre-migration baseline through adoption and the experience change lands close to the observed arrival rate, and far from the rate the original plan assumed.
Footer	Offline share fell from about 53% to about 10% over the same period, so the residual sits on the channel, not on adoption.

Departure to arrival, with the planned rate beside it.

Speaker notes

Two ways to get the first bar wrong, both common: borrowing the arriving platform's norm imports a different population, and averaging a whole account when only part of it moves imports the markets that are staying. State the rule as a control, not as a lesson someone should have learned.

Slide 08 — What lands on Monday

Title	Here is what lands in your inbox on Monday
Subtitle	Six panels, fixed order, every migration, every week
Key message	The standard is not a promise of a process; it is a page with a fixed shape, so that any migration can be read without re-orientation and a defect found once becomes a field every later migration must complete.
Footer	Four of the six panels already exist. The one that does not is the one that would have caught the defect.

The weekly pack, as it will actually arrive.

Speaker notes

Walk the panels in order and stop on panel 3's over-under strip — blue, then gold, then two berry cells is the shape of the year in half an inch. Full specification at The Migration Health Pack; the field-level data contract is held separately.

Slide 09 — Requirement and supply on one axis

Title	Drawn apart, the reinforcement looks on time
Subtitle	A mobilization plan and a requirement outlook on separate axes hide their own gap
Key message	Plotting supply arrival on the same time axis as requirement makes a reinforcement wave landing after its peak visible; drawn separately the gap is arithmetically present and perceptually absent.
Footer	Read the service projection, not the net position. The response near break-even is not linear.

Requirement and supply on one axis, with arrival marked.

Speaker notes

Make the second point as strongly as the first: a small deficit produces a disproportionate service loss and a small surplus buys very little, so a table read on the net column treats a marginal deficit as a rounding error.

Slide 10 — The intake log

Title	Every ask, logged — six weeks of it is the business case for a gate
Subtitle	Nothing is refused. Everything is recorded, sequenced and counted.
Key message	An advisory log of every modeling request — who asked, what for, what it duplicated, whether it changed a decision — costs nothing and is the only evidence that would justify a mandatory intake later.
Footer	Six weeks of this is the evidence for a mandatory intake. Today nobody can prove the volume, so the request has no case behind it.

One week of the log, as it would actually be kept.

Speaker notes

Showing the instrument rather than describing it is the point. The duplicate column is what makes the case; the hours column is what makes it uncomfortable. Say the sequencing rule plainly: if a request does not change a decision this week, it waits.

Slide 11 — Forward risk, sized

Title	Where the estate breaks, ranked by what it costs if it does
Subtitle	Band height is pool size · rows sorted by consequence, not by the size of the number
Key message	A forward grid weighted by pool size and client value surfaces hotspots early enough to be traded rather than escalated, and ranks them by consequence rather than by the magnitude of the deficit.
Footer	A deficit of two in a pool of thirty is not the same risk as a deficit of two in a pool of eight.

Forward risk, with pool size encoded as band height.

Speaker notes

Small pools carry a proportionally larger idle cushion and a proportionally sharper failure, so the same numeric deficit is a different risk depending on pool size. Mark explicitly the rows whose underlying population churns without visibility; those are estimates and are labeled as estimates rather than presented as coverage.

Slide 12 — What is deliberately deferred

Title	Two things we are deliberately not building yet
Subtitle	Each with the condition that unblocks it
Key message	A mandatory intake and published confidence ranges are both deferred with named unblocking conditions, rather than promised and quietly dropped.
Footer	Naming the limits is what makes the rest believable.

Table slide — no exhibit. Two rows, three columns: Capability · What we do instead · What unblocks it.

Deferred	What we do instead	Unblocked by
Mandatory intake	An advisory gate and a log of every request, kept as a shared list with an owner column	Six weeks of the log — the case is evidential, not technical. No environment is required
Ranges with confidence	One planning number, a named buffer, and a written sensitivity	The calculation layer stated and run — the sensitivity tables already carried are a confidence range in an accepted vocabulary

Speaker notes

The bridge to confidence ranges already exists in a form most rooms accept — a sensitivity table is a confidence range in a different vocabulary. Close the act here and hold the cadence and the ninety-day plan for the discussion rather than the page.

Slide 13 — The data layer

Title	Twelve files a week. Eight of them have no source in any estate.
Subtitle	The canonical minimum dataset against three planning estates, and which view each file feeds
Key message	The views in this deck are produced by a renderer that already exists; what is missing is the feed. Twelve small files a week, in one agreed shape, across three estates that agree on nothing — and a stated calculation that turns them into a requirement.
Footer	The renderer already exists and reads these files and nothing else. The validator enforces the contract before anything renders.

The canonical minimum dataset, and where the gaps are.

Speaker notes

This is the slide that converts the framework from a proposal into a plan. The renderer is built and the contract is enforced by a validator, so the remaining work is producing files rather than building software. Be explicit that eight of the twelve have no system source: seven are small human records needing a home and a habit, and the eighth is a pool register. Say plainly that the requirement itself is derived rather than extracted, and that stating the calculation is the expensive move. Full method at The Interim Data Layer.

Appendix — program cards

One card per program, same shape each time. A reference section, not a second deck.

Field	What it holds
Program	The name
What it is	Two sentences, no jargon
Why now	The dependency or the clock that forces the timing
What it unblocks	The specific thing downstream that cannot start without it
First ninety days	Three items, each with one named owner
What would change the plan	The observable that would show it working, or not
Grade	Established · Inferred · Asserted

Card A — the standard

What it is. A fixed six-panel pack, refreshed weekly, applied to every migration on the near horizon, preceded by a readiness instrument run before the model.

Why now. The next migrations are already scheduled, and a template written after them documents failures rather than preventing them.

What it unblocks. Comparability across the portfolio, and the handshake by which root-cause findings retire assumption rows instead of circulating as reports.

First ninety days. Build the approach panel, which does not exist · run the readiness instrument cold against the next migration · publish a definitions page held once for the portfolio.

What would change the plan. If the pack does not answer the recurring questions consistently by the fourth week, the shape is wrong, not the cadence.

Card B — the intake

What it is. An advisory log of every modeling request arriving anywhere in the function, with a sequencing rule. Not a mandatory gate.

Why now. An organization still being assembled will route around a mandate; the log costs nothing and is the only evidence that would justify one later.

What it unblocks. A defensible case for a real intake, and visibility of where one analyst has become a single point of failure.

First ninety days. Start the log · state the sequencing rule once to each requester · review at six weeks against duplicated and discarded work.

What would change the plan. If six weeks shows low duplication and little discarded work, the intake is not the constraint and the case should be dropped.

Card C — the forward grid

What it is. A forward view of surplus and deficit by supply pool and week, weighted by client value and pool size, out to twenty-six weeks.

Why now. Seasonal peaks are known in advance and are currently met on arrival rather than in the plan.

What it unblocks. Hotspots early enough to be traded between owners rather than escalated, and the evidence base for pool consolidation.

First ninety days. Pilot on one book against the next known peak · mark the blind region rather than filling it · decide the row key before building any register.

What would change the plan. If the first pilot surfaces nothing already known, the granularity is wrong — most likely the row key.

Source

Built 2026-09-26 from a near-term workforce strategy in progress. Slot text and exhibits are de-identified; the named version is held outside the wiki. Method pages: Migration Archetypes · Migration Modeling Readiness · The Assumption Register · The Migration Health Pack · Migrating a Book of Business. Deployable project setup: Wiki:Packs/Migration Portfolio Review.