I spent two weeks at a food processing unit in Maharashtra last year. Sixty staff, three shift patterns, one head office in Bengaluru, two cold-store franchisees in Pune, and roughly eight hundred SKUs across packaged dry, frozen, and ready-to-cook.

By the end of the second week, I had figured out the actual ERP. It wasn't software.

It was a spreadsheet. About fourteen tabs. Maintained by Pradeep, the operations manager. The single source of truth for procurement, batch tracking, costing, and the daily cash position. Each tab inherited from the one before. Each shift change involved Pradeep updating four tabs over fifteen minutes, then PDF-ing the day's batch sheet to WhatsApp.

The factory had Tally. The factory had a shipping management module someone had sold them four years ago. Neither was the system of record. Pradeep's spreadsheet was.

What I expected to find#

I went in expecting to see a chaos that needed structure. A small business with no system, papering over gaps with manual work, primed for a proper ERP.

What I actually found#

A small business with a system. The system was personal. The system was the spreadsheet. And the spreadsheet was, on its own terms, good.

It had:

  • Batch IDs cross-referenced from procurement through to dispatch.
  • A daily cost-of-production rollup, computed live.
  • An exceptions tab that flagged anything that broke the model — a batch under-yielding, a raw material cost spiking.
  • A row for every audit-able event. Three years of history. Searchable by Ctrl-F.

What it didn't have:

  • Mobile access (Pradeep's phone struggled with the sheet).
  • Multi-user simultaneous editing without breaking.
  • A floor-friendly UX for the line workers who reported into it.

The design move#

The conventional move would be to build "a real ERP" — a system that replaces the spreadsheet. Modules, workflows, role-based access, the works.

I did the opposite. The first version of SunnyERP runs the spreadsheet. The data model mirrors Pradeep's tabs, row for row. The interface looks like a spreadsheet on web and a clean flat list on mobile. The features that the spreadsheet does well — search, paste, undo, formula columns — are preserved.

What the product adds is what the spreadsheet couldn't do: floor-friendly mobile entry, multi-user safe-write, FSSAI batch traceability, GST e-invoicing.

The point is not to replace the spreadsheet. The point is to make Pradeep's spreadsheet possible for ten thousand other Pradeeps.
— SunnyERP design note, June

What I learned#

Three things.

First: the user's existing system is rarely as broken as the new software's pitch deck claims. It survived for a reason. Find the reason. Encode it.

Second: a vertical ERP is not "Sage but in food." It's a product designed around the actual nouns and verbs of a specific industry. The verbs in food processing are receive, batch, yield, dispatch. They are not the verbs in a generic ERP.

Third: the people who run small businesses are smart, calibrated, and underestimated by enterprise software. The product they deserve is the product the lab is trying to build.

Pradeep, when I showed him the first pilot version, opened the batch screen, scrolled through three weeks of history, and said: "ok, where's the export-to-Excel button?"

We're keeping that.