Numbers arrive too late to act on
A close landing on the twentieth means every operational decision in the preceding three weeks was made on instinct. The duration matters less than the timing.
By role
The recurring frustration in this job is not that the numbers are wrong. It is that producing them consumes the team, arrives late enough to be historical, and cannot be interrogated in the meeting where it matters. That is a systems problem and it is solvable without a nine-month implementation.
The situation
A close landing on the twentieth means every operational decision in the preceding three weeks was made on instinct. The duration matters less than the timing.
Exporting two trial balances, sorting by magnitude, and opening transactions one at a time — re-derived from scratch every period.
Terms-based ageing with a blanket fifteen-day slip applied to every customer, which is three wrong assumptions wearing one number.
Two days of collecting statements, rebuilding KPI movement, and writing commentary from a blank page, every quarter.
A raise or a sale turns a working close into an urgent problem — rev rec, deferred revenue, and controls, in that order.
Hiring into finance to keep up with volume rather than to add capability, and knowing it will happen again at the next growth step.
The honest sequence is unglamorous. Month one is connection and reporting — the dimensional P&L you have been rebuilding in Excel becomes a query, and the AP agent starts drafting. No workflow changes and no risk.
Month two is where the close moves, because reconciliation runs daily rather than at period end and the close checklist starts chasing its own blockers. Most customers see close duration fall by a third in the second cycle and roughly halve by the fourth.
Month three is when the analytical work starts being worth something, because there are finally three clean periods to reason across. Variance decomposition and behaviour-weighted cash forecasting both need history to be useful.
If a raise, a sale, or a lender review is anywhere in the next two years, this is the part worth weighting heavily. Diligence teams pull three things in order: the deferred revenue rollforward, a sample of contracts traced to recognised revenue, and the treatment of modifications.
All three are answerable in minutes when schedules are driven from contract records and take a week of spreadsheet archaeology when they are not. The difference shows up as discount rather than as a finding — reviewers price uncertainty rather than arguing about it, and that pricing is usually larger than the entire cost of fixing the systems.
If your close works, you have one entity, and the specific pain is a reporting gap, you do not need a new ledger. That is a reporting engagement and it costs a fraction of a migration. Roughly a third of the assessments we run for CFOs conclude exactly that, and saying so is the reason the other two thirds are worth taking seriously.
Where to start
Read-only integration and the dimensional reporting your current system cannot produce. No workflow change, no risk, and it establishes whether the data supports what you want.
The largest block of manual work, agent-drafted and human-reviewed. This is where the labelled corpus builds and where the team feels the difference first.
Daily reconciliation and a structured close checklist. This is where the duration number starts moving and where the timing benefit arrives.
Behaviour-weighted cash and decomposed variance, once three clean periods exist to reason over. Attempting these earlier produces confident noise.
Questions
Three questions and a written answer within a business day, including the answer that says change nothing.