Plan vs ActualYou may know it as FP&A
Rolling forecast refresh
Each month the system pulls closed actuals and the latest pipeline, headcount and usage drivers.
What to expect: ideas for what could change, what you need first, and how to test whether it helps. Nothing here is a measured result.
Software and SaaS, Professional services, Manufacturing, Retail and consumer
- Impact: Medium
- Ease: Medium
- Prep: Some
Indicative rating based on the workflow described. Your readiness and implementation effort may differ.
Evidence status: Editorial / indicative assessment
What it does
Most forecast refreshes start with days of copying actuals and CRM exports into spreadsheets. By the time the numbers tie, there is little time left to explain them or test what could change.
- After close, the system pulls actuals from the general ledger and drivers from CRM, billing and HR systems.
- It checks the data for gaps, such as missing bookings or unmapped accounts, and flags them before modeling.
- A model produces a baseline forecast by driver, such as new bookings, churn, headcount and hosting costs.
- It compares the new forecast with the last one and writes plain-language notes on the biggest changes.
A model builds a driver-based baseline forecast for the coming year and a half and writes a first pass at variance commentary. FP&A reviews the baseline, adjusts what the model cannot know, and publishes. The team spends its time on judgment, not spreadsheet plumbing.
Who does what
AI does the work. A person reviews the results and signs off.
What it could move, and how you would check
Of the five measures below, this use case may help two. These are expected effects to test against your own baseline, not measured results. The last column says how to check each one.
| Fact | Expected effect | Why |
|---|---|---|
| Process steps | Not directly addressed | This use case does not directly change the number of steps. |
| Cycle time | Potential direct benefit | Work may wait less between tasks, so items can finish sooner. |
| Exception loops | Possible downstream benefit | Fewer items may go back for rework. |
| First-time-right rate | Not directly addressed | This use case does not directly change how many items are right first time. |
| Cost per invoice | Not directly addressed | Cost per invoice measures invoice work, and this use case does not act on invoices directly. |
What you need first
This is an illustrative improvement approach, not a final solution design. The workflow, level of automation, integrations, and review requirements will depend on your existing tools, available capabilities, data quality, company policies, and team needs. Confirm these during discovery and a controlled pilot before implementation.
Ease of getting started: Medium. The tools exist. Some of the setup work is yours.
- A reliable monthly close with actuals final within a set number of days.
- At least 24 months of consistent history for the main drivers.
- CRM pipeline stages and close dates that sales actually maintains.
- An agreed list of forecast drivers and who owns each one.
- Each published forecast is approved by the FP&A lead or CFO.
Evidence to keep
- Reconciliation of forecast actuals to the closed general ledger each month.
- Override log showing who changed which driver and why.
- Approval record for each published version.
What can go wrong, and how a pilot would show it
- The model can miss one-off events, such as a large customer leaving, unless someone overrides it.
- Teams may accept the baseline without challenge, so review must stay real.
- Poor CRM hygiene produces confident-looking but wrong bookings forecasts.
Year one, in hours
- Forecast refreshes per year, 300-person SaaS company (assumption)assumption
- 12
- FP&A hours per refresh today (assumption)assumption
- 40
- FP&A hours per refresh with a model baseline (assumption)assumption
- 16
- Hours saved per refresh
- 24
- Hours saved per year (12 x 24)
- 288
Illustrative, for a 300-person SaaS company. Every input is an assumption; replace it with your own refresh calendar and the hours your team logged last cycle.
Your next step
Write down the five to eight drivers that explain most of your revenue and cost. Check that each driver has at least two years of clean history and a named owner.
The pilot steps are under "See how to test it" in the Ways to start panel. Paid support is optional: use Get implementation support if you want help.