Capstone: Turn FP&A into a Decision Product (Ship Insight, Not Files)
Introduction — “Stop Delivering Spreadsheets. Start Delivering Choices.”
After nine posts of frameworks, rituals, and politely opinionated tables, here’s the capstone: how to run FP&A as a product. Not a department, not a service desk — a product that users love because it consistently answers the real questions with just enough detail and a clear next move. This post stitches everything together: the decision questions that matter, the operating rhythm that keeps momentum, a sprint plan to reach escape velocity, and a meeting template that turns “nice deck” into “we chose option B.”
Expect practical checklists, some dry jokes (hydration is key), and a simple way to wire it all into a platform like CCH Tagetik so your new normal doesn’t crumble the first time FX sneezes.
Why FP&A Must Act Like a Product Team
Products have users, roadmaps, and feedback loops. Departments have inboxes. When FP&A behaves like a product team, your “customers” (execs, BU leaders, Sales, Ops) know exactly what they’ll get, when they’ll get it, and how to ask for improvements without a reply-all opera. You prioritise problems worth solving, measure adoption, and sunset features no one uses. Translation: fewer slides, more impact (and fewer midnight adventures with CSVs).
The shift is philosophical but practical: define the decision questions you serve, standardise a pack that answers them, lock a cadence, and run improvements in sprints. Tools don’t create discipline — they amplify it — but a platform like CCH Tagetik ensures the plumbing doesn’t leak when you start moving fast.
Five Decision Questions Your Product Should Always Answer
- What changed? Show a clean PVM for revenue and a margin walk for COGS (one page each, prose allowed, drama optional).
- Why did it change? Separate external (market, FX, freight index) from internal (discounting, mix, yield, capacity).
- What happens next if we do nothing? The Base scenario, not a morality tale.
- What are our options? Two or three levers with owners, timers, and rough ROI. No “mysterious option C.”
- What’s the cash impact? DSO/DPO/inventory effects alongside EBIT, because the board always asks (and they’re right).
If your pack answers those five crisply, you’re a product. Everything else is appendix (perfectly fine, respectfully quiet).
The Decision OS: Blueprint for Repeatable Choices
The Decision OS is three layers working together: Inputs (drivers, actuals, assumptions), Logic (scenarios, allocations, rules), and Outputs (KPIs, dashboards, actions). Wire them once, then change the world safely from the assumptions panel instead of spelunking through spreadsheets (bring a headlamp anyway, for tradition).
- Inputs: Price, volume, mix, FX, freight, headcount, capex, working capital parameters.
- Logic: Base/Best/Downside inheritance, driver-based calculations, P&L→BS→Cash linkage.
- Outputs: A 7-slide pack (snapshot, PVM, margin walk, cash, outlook, actions, appendix) and a short action log.
Governance sits around the whole thing: glossary, RACI, version locks, and a change window. It’s not glamorous, but then again neither is seatbelt design — and look how useful that is.
Twelve-Week Sprint to “New Normal”
A focused path from “we have files” to “we ship choices” — without boiling any oceans (a polite simmer will do):
| Weeks | Focus | Key Moves | Outcome |
|---|---|---|---|
| 1–2 | Definition & scope | Pick 5 decision questions; draft KPI glossary v1; lock RACI; choose pilot BU | Shared language; clear owner map |
| 3–4 | Data & drivers | Load actuals; set driver forms (price, volume, FX, freight); publish assumptions | One model; fewer mysteries |
| 5–6 | Logic & scenarios | Base/Best/Downside inheritance; P&L→BS→Cash ripple; variance thresholds | Fewer spreadsheets; faster “what if” |
| 7–8 | Pack & narrative | Build the 7-slide pack; add action log; link glossary to tiles | Consistent story; less drift |
| 9–10 | Pilot & refine | Run one full cycle; capture feedback; adjust drivers & visuals | Proof it works; fewer objections |
| 11–12 | Scale & lock | Extend to second BU; publish SLAs; lock versions pre-exec review | New normal; on-rails cadence |
Keep each step small and visible. Celebrate the first “we chose B” meeting like a tiny product launch (confetti optional, recommended).
Meeting Template: From Variance to “We Choose…”
Decision meetings are a format, not a surprise party. Use the same agenda every time so people arrive ready and leave with owners.
| Time | Topic | Output |
|---|---|---|
| 0–5 | Snapshot (5 KPIs) | “Green/Amber/Red — here’s the headline.” |
| 5–20 | What changed? (PVM & margin walk) | Top 3 drivers of variance (no archaeology). |
| 20–30 | Why? (external vs internal) | Two-sentence root cause per driver. |
| 30–40 | What next if we do nothing? | Base scenario through cash. |
| 40–55 | Options (2–3) with ROI | Pick one. Assign owner, timer, checkpoints. |
| 55–60 | Confirm & close | Decision log updated, comms ready. |
If you run out of time before choosing, it wasn’t a decision meeting — it was a documentary. Lovely cinematography; low ROI.
Service Levels for FP&A (Yes, Really)
Products come with service levels; so should FP&A. Publish a short SLA so everyone knows what “good” looks like and how to ask for more without summoning chaos. This isn’t bureaucracy; it’s empathy — the opposite of “surprise, we changed the template again.”
- Cadence: Weekly driver check (30 min); monthly decision review (60 min); quarterly deep dive (2 hours).
- Turnaround: New scenario in < 60 minutes if drivers exist; in < 3 days if new mappings are needed.
- Lock times: Pack locks 24 hours before exec review (prevent the “just one last number” phenomenon).
- Support: Office hours weekly; response to flagged data issues in 1 business day.
- Change window: Glossary & formula changes approved monthly; urgent exceptions documented.
SLAs feel formal until you’ve used them once. Then they feel like a calendar that respects humans (rare, precious).
Real-World Moments You’ll Recognise
- The Urgent Discount: Sales proposes a 10% promo. You clone Base, adjust price/volume, and show GM% and cash side-by-side. Decision in five minutes; everyone still likes each other.
- The Freight Spike: Ops flags a lane up 6%. You push a downside with higher COGS, then present three mitigations with payback windows. Suddenly “do nothing” looks expensive.
- The KPI Debate: “Which margin?” The slide links to the glossary; definition pops up; debate ends; meeting proceeds (peace breaks out).
- The Late Surprise: A number changes after lock. The versioned pack reveals when/why/by whom. Trust dips for five seconds, then recovers (audit: nature’s soothing balm).
Scoreboard: Metrics That Say “It’s Working”
- Decision conversion: ≥ 80% of agenda items end with owner & date.
- Scenario turnaround: Base/Best/Downside with cash flow-through in < 60 minutes.
- Cycle time: Close-to-pack publication in ≤ 5 business days.
- Accuracy uplift: 10–20% improvement in revenue/GM MAPE over two quarters.
- Adoption: > 90% submissions via platform; < 10% off-system edits.
- Meeting efficiency: Average decision review ≤ 60 minutes, and ends with updated action log (no cliffhangers).
Where CCH Tagetik keeps the rails on: one governed data model; central driver forms with BU overrides and audit notes; scenario inheritance from Base; automatic P&L→BS→Cash ripple; workflow reminders and version locks; a KPI glossary tied to tiles and reports; exception dashboards for variance thresholds. Translation: you run the conversation, the platform runs the mechanics (like a very calm stage manager).