Why We Didn't Build an MES: Tracking Production Data With a Glorified Spreadsheet

If you've ever tried to get a straight answer out of a factory floor — "how much raw material did we actually use today, and how many units came off the line" — you already know the problem. Paper checklists say everything's fine. Reality says otherwise. And the gap between the two is exactly where downtime, defects, and budget overruns hide. That's the gap platforms like logsheet.ai are built to close.

We ran into this on a discrete manufacturing line — multiple departments, multiple handoffs, no single source of truth for material balances, defect counts, or output. The instinct is to reach for a full Manufacturing Execution System (MES). We didn't, and here's the reasoning, plus what we built instead.

Why bother tracking this at all

Before getting into tooling, it's worth being explicit about what production record-keeping actually buys you:

• A live picture of the shop floor without needing to physically walk it.
• Source data for planning — real output, consumption, and time norms only mean something if they come from real records, not estimates.
• No gap between what's actually happening and what gets reported upward.
• The ability to trace a defect back to the exact stage it occurred, and assign the cost to the right place instead of burying it in general opex.
• Downtime visibility at every stage, not just the stages someone remembered to mention in the standup.
• An audit trail for ISO and customer compliance.
• Enough historical data, eventually, to build a digital twin of the process or move toward predictive maintenance.

None of that requires a six-figure platform. It requires structured, timestamped data that people actually fill in.

The MES option, and why we passed on it

The obvious heavyweight answer is a full MES: ZIIoT (Tsifra), Golfstream (Ascon), 1C:MES Manufacturing Execution — systems that plug into ERP, distribute orders across shifts and machines, and pull real-time data off the equipment itself.

They're the right tool if you're operating at serious scale. But they come with a specific cost profile:

• Licensing and implementation running into the hundreds of millions of rubles.
• Rollout timelines measured in months at best, years in the worst case.
• A break-even point that only really makes sense north of ~10,000 employees, with high production volume and enough process variability to justify it.
• A downstream requirement to rebuild your reporting and analytics methodology around the new data model.

For a mid-size line, or for tracking a secondary process that doesn't need full automation, that's a lot of infrastructure to buy just to answer "how much material did we use."

What we built instead: log sheets, not a platform

The lightweight version is closer to a well-structured spreadsheet than a platform: a system of log sheets. Setting one up is genuinely a five-step exercise:

1. Create a flexible table.
2. Define the process it covers (quality control, inter-shop equipment transfer, whatever's relevant).
3. Set the metrics to record — deadlines, quantities, units of measurement.
4. Assign who's responsible for keeping it up to date.
5. Set how often it needs to be filled in.

From there the system prompts people for what's due and when, and entries happen from a phone or tablet right on the floor — no walking back to a terminal. Reports roll up per shift, per period, or in real time, and a single dashboard can pull from multiple logs at once. You build the report by picking the relevant logs/checklists and a reporting window — no custom query writing.

A concrete example: a plastic container line

Here's the shape of it on an actual line — a plastic bottle manufacturer that needed live logs for raw materials, finished goods, and inspection results.

Raw material log — receipts and issues, balance recalculated automatically on every entry:

Date/Time	Material	Supplier	Received, kg	Issued, kg	Balance, kg	Responsible
Jul 1, 2026 09:15	HDPE (pellets)	PolymerTrade LLC	1000	0	1000	I. Ivanov
Jul 1, 2026 10:30	HDPE (pellets)	PolymerTrade LLC	0	200	800	P. Petrov
Jul 2, 2026 11:45	Colorant (blue)	ChemReaktiv JSC	50	0	50	S. Sidorov

When the warehouse worker receives a batch, they log the quantity and the balance recalculates on its own. When material goes out to the floor, same thing in reverse. The log is always current — no end-of-week reconciliation.

Finished goods log — plan vs. actual, by line and shift:

Date & Shift	Line No.	Product	Plan / shift, pcs	Actual, pcs	Runtime, h	Operator
Jul 1, 2026 / Shift 1	1	Bottle 0.5 L	5000	4850	7.5	A. Smirnov
Jul 1, 2026 / Shift 1	2	Bottle 1.0 L	4000	4100	8.0	V. Kuznetsov
Jul 2, 2026 / Shift 2	1	Bottle 0.5 L	5000	5020	7.8	M. Orlova

This is the log you cross-reference against the raw material log to see whether output lines up with what went in.

Quality control checklist — per batch:

Inspection Step	Answer Options	Note
1. Visual appearance	Conforms / Does not conform	Photo of defect
2. Dimensions (height, diameter)	Within spec / Deviation	Enter value
3. Seal integrity	Intact / Defective
4. Unit weight, g	(enter value)
5. Color (vs. reference)	Conforms / Does not conform
Overall verdict	Batch accepted / Rejected / Partial reject

Tie all three together and the report writes itself: how much raw material went into how many finished units, what the defect rate was, and how actuals diverge from plan — auto-calculated, not hand-tallied at the end of the month.

Where else this pattern applies

The same log-sheet-plus-checklist pattern covers a lot more than this one example: general construction logs, operations logs, equipment logs, maintenance logs, temperature logs, safety logs, incident logs, and so on. It's the same five setup steps every time.

The trade-off, honestly

This isn't a replacement for an MES if you're at real scale — it doesn't do real-time machine data capture, and it won't replace deep ERP integration. What it does do: get you from zero to structured, timestamped, queryable production data in about the time it takes to define your columns, without months of implementation or a seven-figure line item. For a lot of teams, that's the actual bottleneck — not the lack of a platform, but the lack of any structured record at all.