Digital agency · monthly close
A monthly financial workbook and a partner-facing deck of roughly thirty-one slides. It used to be three to four hours of production work, every month, forever. It took six hours to rebuild as something that produces itself — which means it paid for itself before the second month closed.
6 hours to build
Against three to four hours of manual production every month, indefinitely.
The problem
The package was spread across five separate spreadsheets, each maintained by hand with lookup formulas and a certain amount of typing. Some figures were linked into the slide deck; others were pasted in as screenshots, which meant any late correction had to be re-pasted rather than refreshed. The commentary and the flux analysis were written from scratch each month.
None of it was difficult. All of it was load-bearing, exacting, and due at the same time — and the person producing it was the same person who was supposed to be interpreting it.
What it does now
The closed month's actuals come through the profit-and-loss report, with a transaction-detail export supplying customer-level revenue.
Every dependent tab updates: monthly P&L, sales by customer with a new month-specific detail sheet, cash and AR, the summary P&L that feeds the deck, the overhead allocation headcount rows, the by-year and by-department tabs, and the forecast tabs where the closed month flips from forecast to actual.
Verified against the payroll system, with marketing rolling up analytics, paid media and SEO. Contractors don't appear in payroll, so any delta comes back as a question rather than an assumption.
A client alias map keeps sales-by-customer matching intact across sub-brands, parenthetical legal names, and clients split across more than one business unit — the thing that quietly breaks this kind of report.
Roughly thirty-one slides, combining native slides, generated commentary, rendered visuals, and an inserted executive summary and outlook. A PDF preview renders so the slides can be eyeballed before anything is shared.
Four partner personas pressure-test the deck before it ever reaches a real partner's inbox — the questions get asked while there's still time to answer them.
Where it stops for a person
Deliberately, in several places. Unmapped accounts are flagged before anything is written. Working days, utilization, average bill rate, cash, AR and DSO are supplied rather than inferred. Any headcount delta is confirmed.
Then a hard stop: the deck isn't built until the workbook is confirmed final. Commentary gets edited on the deck after the build, not before — because the wording is the part a person should own.
What actually changed
The job changed, not just the hours.
I'm no longer buried in data production. I'm reviewing the output to make sure it tells the right story — which is the work I actually want to be doing.
From the post about this month's close.
The time saving is the least interesting part. Three or four hours a month is real, but it's not what makes this worth writing down. What matters is that the reporting stopped being something to survive and became something to think about — and that the person who understands the numbers best is now spending their month-end reading them instead of typing them.
Written up on LinkedIn — 2,147 impressions, 32 reactions. This is one of six skills built for the agency.