This manufacturing digitization case study starts the way most of them do: with a spreadsheet nobody but one person fully understood. A family-owned manufacturer producing household goods for both retail and B2B clients had grown from a workshop into a real operation, forty-plus people, three production lines, dozens of SKUs. The systems running it, though, were still forty-something linked spreadsheets, each one feeding formulas into the next, maintained by a production manager who had built the whole chain over eight years and was, by his own admission, the only person who could safely touch it.
That is not a stable foundation for a business doing meaningful revenue. It is a single point of failure wearing a spreadsheet costume, and the owners knew it, which is why they finally called.
What the spreadsheets were actually doing
Before touching anything, we mapped the existing flow rather than assuming what needed to change, the same discipline covered in mapping the process before you automate it. What we found was a chain that had grown organically and never been redesigned:
- A raw material intake sheet, updated by hand from supplier delivery notes.
- A production log sheet, one tab per line, tracking output and material consumption per batch.
- A costing sheet that pulled from both of the above via cross-file formulas to estimate cost per unit.
- A separate stock sheet for finished goods, updated manually and often a day or two behind actual warehouse counts.
- A sales tracking sheet that had no live connection to any of the above, so margin figures were recalculated by hand at month end.
Each handoff between sheets was a manual copy-paste or a formula that broke silently whenever someone inserted a row in the wrong place. The production manager caught most of these breaks because he had memorized where they tended to happen. That is not a system, that is institutional memory standing in for one, and institutional memory does not scale and does not take vacation.
The digitization plan
We did not propose a full ERP replacement in one go. A forty-person manufacturer does not need, and cannot absorb, a six-month enterprise rollout with a steep training curve and a six-figure license. The plan was a lightweight, purpose-built system covering exactly the chain that was breaking:
- Single intake point for raw materials, replacing the manual delivery-note entry with a simple form tied directly to a stock database, no more separate sheet to reconcile.
- Production logging by line, still fast for line supervisors to fill in during a shift, but writing to one shared database instead of one tab per line.
- Automatic costing, calculated per batch from actual material consumption and labor hours logged, not from a formula chain that could silently point at the wrong cell.
- Live finished goods stock, updated the moment a batch closes, visible to sales without anyone needing to ask the warehouse.
- A margin view per SKU, pulling real cost and real sales price into one place for the first time.
None of this required exotic technology. It required someone to sit with the production floor for two weeks, watch how the sheets were actually used versus how they were designed to be used, and build the smaller, connected system underneath. The interface stayed close to what the team already knew, forms and tables, because retraining forty people on a completely unfamiliar tool would have cost more in disruption than the project was worth.
What the real numbers revealed
The uncomfortable finding came in month two, once the costing engine was running on real batch data instead of formulas that had drifted from reality over years of edits. Two of the company's product lines, both steady sellers, had been priced below their true production cost for at least eighteen months. The original costing sheet had a stale labor-rate assumption baked into a formula from an earlier year, and nobody had revisited it because the sheet "worked," in the sense that it produced a number and nobody questioned the number.
Once real costing was visible per SKU, the picture was stark:
| Product line | Assumed margin (old sheet) | Actual margin (new system) |
|---|---|---|
| Line A (core seller) | +12% | +9% |
| Line B (secondary) | +8% | -3% |
| Line C (secondary) | +6% | -2% |
The two underwater lines were not small experiments, they were meaningful volume. The company had been effectively subsidizing customers on those SKUs with margin earned elsewhere, and nobody had chosen that trade-off deliberately. It had happened by drift.
What changed after
With real numbers in hand, the owners made three decisions inside a month: they repriced Line B for new orders, renegotiated a key input cost on Line C with a supplier, and set a quarterly review of the margin dashboard as a standing item, so a stale assumption cannot silently compound for eighteen months again. That last habit matters more than any single repricing decision, and it is worth pairing with a broader quarterly tech budget review so the system itself does not quietly rot the way the old spreadsheets did.
The production manager, freed from being the sole keeper of a fragile formula chain, moved into a role focused on actual production planning rather than sheet maintenance. That is usually the real return on this kind of project, not just correct numbers, but a key person no longer trapped maintaining infrastructure instead of doing the job they were hired for.
The lesson for any owner still running on linked sheets
If your business runs on a spreadsheet chain that one person understands, you do not have a system, you have a dependency. The fix is rarely a giant ERP purchase. It is usually a smaller, purpose-built layer that replaces the fragile handoffs while keeping the parts of the workflow your team already knows. And the finding that pays for the project is almost never the efficiency gain, it is the margin truth that was hiding in a formula nobody had checked in years.