Field notes from rebuilding FP&A.
Short notes and longer reads on what we're learning as we build a complete FP&A stack around real teams - written by the people doing the work, not a marketing team.
Why FP&A tools never fit (and what fits looks like)
Every FP&A rollout misfits in the same three places, for the same business-model reason. A field essay on what actually fits, and why fitting requires a specific kind of team.
Why we ship in 14 days, not quarters
The reason FP&A rollouts take six to twelve months isn't software complexity. It's the handoff tax between three or four companies. Here's what we cut, and what made it possible.
AI is only as good as the data layer underneath it
Most 'AI for finance' demos work because the data behind them was prepared for the demo. A note on what actually has to be true under the chat box, and what we built differently.
Build vs buy FP&A: what teams underestimate before they start
Building it in-house looks like the right move when you have engineering capacity. The hidden costs are predictable and almost always missed. A field note on the 18-month arc.
Connecting LLMs to FP&A
A finance lead asks Claude for Q1 ARR by entity and gets her real numbers back. A note on what it took to make that safe, useful, and consistent with the rest of the Novaplan data layer.
FP&A software for mid-market: the two tiers, and the gap between them
The FP&A market split into Excel-adjacent tools and planning platforms. Each solves one problem and creates another, and mid-market teams live in the gap. A map of the landscape, and what actually closes it.
Anaplan alternatives for mid-market finance teams
Anaplan is rarely the problem. Buying enterprise modeling gravity you can't staff is. An honest look at when Anaplan fits, when it's overkill, and the real alternatives for a mid-market team.
Headcount planning: the model most finance teams get wrong
Headcount is 60 to 70 percent of opex and usually modeled as FTEs times an average salary. That shortcut is wrong on both the amount and the timing. What a real headcount model tracks instead.
Driver-based planning, explained (and how to actually build one)
Driver-based planning means forecasting the operational drivers that move your numbers and letting the numbers follow. What it is, why line-item budgets go stale, and how to build one that doesn't.
What a monthly close should really take, and where the days go
A healthy mid-market close runs about five business days. Many teams take ten or twelve, and the gap is coordination, not accounting. Where the days actually go, and how teams compress them.
Pigment alternatives for mid-market finance teams
Pigment is genuinely well-designed, so an honest alternatives piece starts there. When Pigment fits, where mid-market teams start looking around, and the options grouped by who has to own the model.
How to build a board deck finance actually trusts
A board deck is an argument that you know your own numbers, and one figure that does not reconcile turns the meeting into an audit. What a trustworthy deck requires, and why the assembly method is the whole game.
Rolling forecast vs annual budget: when to switch, and how
The annual budget stops being true by February. A rolling forecast re-anchors on actuals every month across a fixed horizon. When the switch is worth it, and the failure mode that quietly kills it.
Curious whether any of this fits your team?
Tell me about your FP&A