Nobody sets out to have a three-week close. It accumulates. A workaround here, a manual step there, and eighteen months later the first half of every month belongs to the close.
In our experience the causes are almost always structural rather than effort-related. Here are the five we see most, roughly in order of how much time they consume.
If someone is matching bank lines to ledger entries by eye, that is the largest single time cost in most closes, and it scales directly with transaction volume. Double the customers, double the work.
Rules-based matching handles the routine majority and leaves a genuine exception queue. The improvement is not incremental — it changes reconciliation from a task measured in days to one measured in hours, and it stops growing with the business.
The tell: your reconciliation time this year is materially higher than last year, with the same process.
Deferred revenue calculated in a spreadsheet and posted as a manual journal creates two problems. The obvious one is the time to maintain it. The less obvious one is the time to defend it — every mid-term change requires the schedule to be rebuilt and re-reconciled.
Teams often underestimate this because the monthly journal itself is quick. The cost is in the amendments, the reconciliation, and the investigation when the balance does not tie.
Deriving the schedule from the contract record removes the class of work entirely, which is what revenue automation is for.
For groups, this is frequently the longest single step. Export each entity's trial balance, map accounts, apply FX, post eliminations manually, check the balance sheet balances, discover it does not, investigate.
The work is not just the build — it is that the build is re-done monthly, so any improvement has to be re-applied monthly too. Nothing compounds.
Native consolidation computes the group view from entity data continuously, which turns this step into a review rather than a construction. We wrote about the underlying mechanics in intercompany eliminations for Australian groups.
A surprising proportion of close time is waiting. Waiting for a manager to approve an accrual, for a department head to confirm a cost allocation, for someone to answer a question about an invoice.
Chasing is invisible in process documentation and very visible in elapsed time. When approvals are routed and tracked by the system, with visible status, the waiting does not disappear but it stops requiring a person to manage it.
The tell: your close has a lot of elapsed days but not many worked hours.
If the close exists as institutional knowledge rather than a tracked list of tasks with owners and dependencies, then every month is partly improvised. Nobody knows what is blocked on what, two people do the same reconciliation, and one task gets missed until someone notices at review.
This is the cheapest of the five to fix, because it needs no software — a shared checklist with owners and due days improves most closes immediately. Systems with built-in close task management make it durable rather than dependent on one person maintaining a spreadsheet.
If you want the largest reduction for the least disruption, in order:
Do not compress the close by skipping review. A five-day close where nobody checks the flux is worse than a ten-day close with a proper review, because the errors just move downstream to the board pack or the audit.
The goal is a close that is fast because the mechanical work is automated, leaving more time for review, not less.
We map the close before configuring anything, because the bottlenecks differ by company and the assumption is often wrong. See our month-end close checklist for the sequence we work towards, or get in touch.