Where close days actually go — Baxter Mountain Group

Short answer: a month-end close is as long as its longest chain of dependent steps, the critical path, not as long as its busiest step. To close faster, find the steps that sit on that chain and shorten those. Speeding up work that isn’t on the critical path saves effort but doesn’t move the close date.

Why doesn’t working harder shorten the close?

Because most close work runs in parallel, and only one chain of it sets the finish line. Bank reconciliations, accruals and intercompany often run at the same time. If intercompany has to finish before consolidation can start, and consolidation before review, that chain decides the day you close. Every other task has slack, room to slip without changing the date.

This is why teams can put in long nights and still close on the same day as last quarter. The extra effort went into steps with slack.

Critical path: the longest sequence of dependent steps from the start of the close to the final sign-off. Shorten a step on it and the close gets shorter. Shorten a step off it and the close date stays the same.

A worked example: closing on day 7 against a day-5 target

Fictional numbers, chosen to show the method. A company runs six close phases. Intercompany depends on the sub-ledgers. Consolidation waits for intercompany and accruals. Review waits for consolidation.

The close lands on business day 7, two days past a day-5 target.
The critical path is sub-ledgers, then intercompany, then consolidation, then review and reporting.
Intercompany is the bottleneck at 2.5 days, and it sits on that path.
Bank reconciliations have slack. Cutting a day from them changes nothing about the close date.
Effort is 280 hours, about 68% of it manual.

Now test the obvious fix. Cut intercompany by one day and the close moves to day 6. Cut it by a day and a half and the close is still day 6, because the critical path has moved: accruals now sets the date. A good close redesign checks this every time, because the bottleneck moves as you fix it.

Where do close days usually go?

In our experience, a handful of steps sit on the critical path again and again.

Intercompany. Confirming balances between entities means waiting for other teams, often in other time zones. One late entity holds up everyone.
Reconciliations with high exception volume. Matching is fast. Investigating the exceptions is slow.
Accruals that depend on late information, such as unreceived invoices or usage data from operations.
Review cycles where a reviewer finds an issue and sends work back, adding a loop to the end of the chain.

Which step should you redesign first?

Rank each phase by two numbers: how many days it adds to the critical path, and how much of it is manual. A phase that is on the path and mostly manual is your first candidate. A phase that is heavily manual but off the path is worth automating for effort, but it won’t move the date.

Phase type	Where an AI agent helps	Where a person stays
Bank reconciliations	Matching transactions and proposing explanations for exceptions	Approving write-offs and unusual items
Intercompany	Chasing confirmations, comparing balances, drafting eliminations	Resolving disputes between entities
Accruals	Drafting entries from open purchase orders and usage data	Approving estimates and judgement calls
Review and reporting	Drafting flux commentary and flagging unexplained variances	Signing off the numbers

We recommend keeping a person on every approval and judgement call, and capping automation at around 80% for most close tasks.

Find the chain that sets your close date, shorten the steps on it, then check where the chain moved.

How do you move toward a continuous close?

Move work out of the crunch. Reconciliations, accrual prep and intercompany confirmations can run during the month instead of all landing after period end. Every hour moved earlier comes off the critical path, and the close becomes a review of work already done.

Key takeaways

The close date is set by the critical path, not the busiest step.
Speeding up steps with slack saves effort but doesn’t move the date.
Rank phases by critical-path days times manual share, and fix the top one first.
Recheck after every change, because the bottleneck moves.
Keep people on approvals, and let agents take the matching, chasing and drafting.

Map your own close

The Close Cycle Benchmark lays your phases on a business-day timeline, finds your critical path, names the bottleneck and lets you test “what if this step were faster.” It’s free and runs in your browser.

Frequently asked questions

What is the critical path in a month-end close?

It is the longest chain of dependent close steps from start to final sign-off. The close finishes when that chain finishes, so only steps on it change the close date.

Why is intercompany so often the bottleneck?

It depends on other entities confirming balances, often across teams and time zones. One late confirmation delays consolidation and everything after it.

How do I find which close step to fix first?

Score each phase by the days it adds to the critical path and the share of manual work in it. The phase with the highest combined score is usually the best first redesign.

Can AI agents shorten the close?

Yes, when they take on work that sits on the critical path, such as matching, chasing confirmations and drafting entries and commentary. People should still approve estimates, write-offs and the final numbers.