Skip to content
Padmashree
All work
Manufacturing FP&A2026

Month-end close automation

The month-end GL report was pulled by hand and broke on merged cells every close.

135 to 25 min
per close
6 months
running live
ERPAutomationGoogle Apps ScriptMonth-end close

What I found

The month-end number depended on who ran it. The GL report was pulled from the ERP by hand, and merged cells broke the paste every single month. Two people could run the same close and get two different starting points.

What I built

A load that arrives on a scheduled email and reads itself. The raw data lands untouched in its own tab, and columns are found by name rather than position, so an added upstream column no longer breaks the whole thing. Failed pulls retry instead of vanishing.

How it works

Per close
25min
down from 135 min
Running live
6months
without a manual pull
To get it stable
3versions
honest, not first-try
Before135 min
Now25 min

Before

  • Pulled from the ERP by hand
  • Merged cells broke it every month
  • The answer depended on who ran it

Now

  • It loads itself from the email
  • Raw data lands untouched in its own tab
  • Same output every time; columns found by name not position
Month-end GL load, ERP to a stable report. Illustrative sample figures.

The honest detail is the 3 versions. It did not work on the first try. It took three passes to make it stable enough that I stopped thinking about it on close night.

Impact

  • The close step went from over 2 hours to about 25 minutes.
  • It has run live for 6 months without a manual pull.
  • The output is the same every time, so the answer no longer depends on who ran it.

The transferable lesson

Automation is not the flashy part. The win is that a fragile, person-dependent step becomes something you can trust at eleven at night on close, and that trust is worth more than the saved time.

The figures in this playbook are illustrative sample data, not any employer's actual numbers. What carries over is the method and the tool, which rebuild for any company.