How I Turned a 30-Minute Daily Task into a 5-Minute Automation (Saved 105 Hours a Year!)

From Excel Hell to Python Heaven

Flow diagram 2021 x 2025:

It was 2018.

If you knew how to make a pivot table in Excel, you were a prodigy. If you could automate something with VBA, you were a God.

Back then, I was responsible for a daily sales report that included 12 pages of KPIs, comparisons and charts. The report was built in a 20MB Excel file, packed with 30+ pivot tables. Every morning, I had to go through this 7-step dance:

Old Workflow (Manual)

1. Download 5 ERP reports in .xlsx format
   


2. Open the Excel template and clean the old data
   


3. Paste in the new data
   


4. Refresh 30+ pivot tables manually
   


5. Cross-check 12 sheets for consistency
   


6. Select multiple sheets with CTRL + Click
   


7. Export as PDF and send via email

📉 Performance:

**Daily Time Spent:** ~30 minutes  

**Annual Time Spent:** 126 hours (~7590 minutes)  

**Key Risk:** 100% dependent on 1 person

It was brittle. It was bloated. And if I was out sick or late, chaos would follow.

The 2025 Automation Flow

After 3 months mapping out the steps and testing multiple automation options, I released a fully automated version by the end of 2021.

Here’s what the process looks like now:

New Workflow (Automated)

# Colab Script (simplified)

import gspread

import pandas as pd

import numpy as np



# Authenticate and connect to Google Sheets

# Process data from uploaded spreadsheets

# Update Sheets with cleaned + calculated metrics

1. Drop the .xlsx files into a Google Drive folder
   


2. Hit "Run" on a Google Colab notebook
   


3. Data is parsed and uploaded to 2 Google Sheets
   


4. 
   Google Slides automatically pull the charts and tables
   


5. Click “Update Linked Objects” → Export as PDF → DONE!

📉 Performance:

**Daily Time Now:** ~5 minutes  

**Annual Time Now:** 21 hours (~1265 minutes)

Net Savings

• 25 minutes saved daily
  


• 105 hours saved per year
  


• Zero manual steps
  


• No more Excel crashes
  


• No more single-person dependency

This wasn't about throwing AI at the problem. It was pure backend automation and smart tooling. I used:

• 
  Google Colab for the Python backend
  


• 
  gspread to connect with Google Sheets
  


• 
  numpy/pandas for all the math
  


• 
  Google Slides for a clean visual frontend

Lessons Learned

• Optimization isn't always about new tech: it's about smarter flows
  


• Mapping the steps is 80% of the solution


• Public tools (Colab, Sheets, Slides) can scale more than we think


• Always measure the actual time cost of recurring tasks

Have You Done Something Similar?

I’d love to hear from anyone who’s pulled off a process optimization like this. Especially if you did it with:

• Python


• Excel automation


• Google Workspace tools


• Airtable / Notion


• APIs and scraping

Feel free to drop a comment or share your wins!