06-reference

cfo secrets building fpa series iii sculpting the team

2026-07-18·reference·source: CFO Secrets·by The Secret CFO (anon)
fp&aorg-designfinance-teamoperating-modelbusiness-partnerhub-and-spokematurity-framework

"The trouble with 'business partners'" — @The Secret CFO

Why this is in the vault

Part III of the Building FP&A series: a five-step framework for designing an FP&A operating model from first principles — from activity inventory through maturity assessment, placement decisions, model selection, and role specs.

⚠️ Sponsorship

Sponsored by Summation (AI Analyst product). The sponsor block frames the pitch as: recurring finance work auto-completed → analysis that "never had time to happen" finally does → leadership gets confident, defensible answers on schedule. Direct-to-CFO positioning; tracked-link URL confirms per-issue sponsorship (not a standing relationship). Not a vendor RDCO uses.

The core argument

The "Finance Business Partner" title is a consulting-industry import from HR circa 1999 — "the meat curry of finance job titles." Its vagueness creates two failure modes in practice: (1) too distant, the FBP becomes a glorified admin and information servant; (2) too close, they "go native" and forget their primary loyalty is to financial performance of the business, not to the exec they support.

The deeper structural point: the org chart is the output of function design, not the starting point. Most CFOs do the reverse — inherit boxes, add roles, copy titles. The five-step model corrects this:

  1. Start with the work — use the 35-activity inventory from Part II; design for current + near-term needs, not aspirational future state
  2. Set maturity levels — L1 (informal/ad hoc) to L5 (fully integrated into how the business runs); the goal is NOT L5 everywhere — overbuilding is as harmful as underbuilding; assess current vs. required maturity per activity bucket
  3. Decide placement — four options per activity: senior hands-on (CFO-proximate), centralized, embedded (close to the business unit), or automated/self-serve
  4. Choose the operating model — CFO-only, one-person show, central team, hub-and-spoke, or embedded; bigger/more embedded is NOT automatically better
  5. Design actual roles — convert placement decisions to work packages → bundle into roles by shared characteristics (business proximity, work type, cadence, seniority) → write role specs with explicit handoffs, escalation points, and exclusions

Forcing-function test: impose an artificial headcount constraint (one fewer person). Map exactly which work disappears and what the business impact would be. If losing that work doesn't feel "extremely painful," the function isn't yet value-dense.

Teaser: Part IV (next week) covers AI in FP&A — explicitly flagged as the last piece of the series.

Mapping against Ray Data Co

The hub-and-spoke model in Step 4 is the exact frame phData's DIE structure already uses — see [[2026-07-06-die-fabric-hub-spoke-map-and-roadmap-implications]] — and the maturity ladder (L1-L5) maps cleanly onto the CAF PM question of where the current Fabric capability sits vs. where the organization needs it. The "org chart as output not input" principle is a direct diagnostic for how not to approach the CAF restructure: if boxes come before activity-mapping, the structure will misfit the actual work. The artificial-constraint test is a useful heuristic for Ray's DSA role: when scoping FP&A-adjacent client engagements at phData, the test surfaces whether the proposed FP&A capability lift is genuinely load-bearing or a nice-to-have.

Related