The Spreadsheet Is Not the Problem
Every operations technology discussion eventually arrives at the same scapegoat: the spreadsheet. If we could just get off Excel, the argument goes, everything would be better.
This misidentifies the problem.
The spreadsheet is not the problem. The spreadsheet is the symptom. It tells you that the systems that should be providing operational data — ERP, MES, WMS — are either not capturing the right data, not surfacing it at the right time, or not making it accessible enough for frontline decision-makers to use.
When supervisors build their own shift tracking spreadsheets, they're not being resistant to technology. They're solving a real information gap with the tool available to them. The question is not "how do we eliminate the spreadsheet?" but "what information do people actually need, and how do we get it to them better?"
The Three-Layer Operational Technology Stack
A modern operations technology stack has three distinct layers, and most facilities are strong in one or two but weak in a third.
Layer 1: Systems of Record ERP systems (SAP, Oracle, NetSuite) and WMS platforms are systems of record. They capture transactions — purchase orders, inventory movements, production orders, shipments. They are excellent at what they do. What they are not designed for is real-time operational intelligence at the line level. They were built for financial reporting and supply chain management, not for a shift supervisor who needs to know right now whether Line 3 is behind pace.
Layer 2: Data Capture PLCs (programmable logic controllers), barcode scanners, RFID readers, conveyor sensors, and vision systems are data capture infrastructure. Many facilities have this technology in place — but the data it generates stays in the equipment or in proprietary software that doesn't talk to anything else. The PLC on your stamping press knows exactly when it cycled and whether the part was ejected cleanly. That data rarely reaches a dashboard visible to an operations manager.
Layer 3: Operational Intelligence This is the gap. The operational intelligence layer sits between data capture and the decision-makers who need to act on it. It aggregates data from sensors, ERP transactions, and manual inputs. It calculates real-time KPIs. It surfaces anomalies and alerts. It enables comparisons across shifts, lines, and time periods. It turns data into decisions.
Most facilities have Layer 1 (ERP) and some Layer 2 (sensors), but no Layer 3. The gap between the data and the decision-maker is filled by spreadsheets, white boards, and verbal shift handoffs.
Why the Multi-Year IT Project Approach Fails
The traditional approach to solving the operational intelligence gap is an enterprise software implementation: select a vendor, run a 12-month requirements process, implement a Manufacturing Execution System (MES) or WMS upgrade, go live 18 months later with a system that the floor team didn't have input on and doesn't know how to use.
This approach fails for three predictable reasons.
First, the requirements process is backward-looking. By the time a system is implemented, the operational problems it was designed to solve have often evolved into different problems — or been solved by workarounds that the new system disrupts.
Second, enterprise MES systems are typically configured for process manufacturing — chemical plants, pharmaceutical production, food and beverage — not for the high-mix, variable-volume environments common in automotive supply and third-party logistics. The configuration overhead to make them fit is enormous.
Third, the implementation timeline means that the people who championed the project are often no longer in the same roles by go-live. Organizational knowledge of why decisions were made evaporates. The system gets deployed in a compromised version that satisfies neither the original requirements nor the current operational reality.
What the Modern Stack Actually Looks Like
The operations that have successfully modernized their data infrastructure have done it incrementally, starting with the highest-value information gap and expanding from there.
Start with real-time throughput visibility. Before anything else, get accurate, continuous data on output rate vs. target for each production line. This single capability — knowing whether you're on pace right now, not at end of shift — is the foundation everything else builds on.
Add downtime capture. When a line stops, the system should know immediately and start a clock. Downtime reasons should be logged at the point of occurrence, not reconstructed at shift end. This doesn't require complex MES — it requires a way to log events in real time.
Connect to existing data sources. Most ERP systems have APIs or export capabilities that haven't been used. Production orders, labor assignments, and inventory levels from the ERP, combined with real-time production data from sensors, create an operational picture that neither system could provide alone.
Surface it where decisions are made. A dashboard in the plant manager's office is useful for weekly reviews. A display at the line is useful for supervisors making minute-to-minute decisions. The same data, surfaced at different levels of the organization in different formats, drives different but equally important decisions.
The Implementation Path That Works
The operations teams that succeed with this transition share one characteristic: they start small and prove value before expanding.
Pick one line. Implement real-time throughput tracking. Run it for 60 days. Measure the improvement in throughput visibility, response time to slowdowns, and end-of-shift accuracy. If the numbers support expansion — and they almost always do — extend to additional lines.
This approach avoids the IT project failure modes because it produces visible value quickly, builds internal champions at the floor level, and lets the system evolve based on how it's actually used rather than how it was theoretically designed.
See how OpsOS tracks this in real time → [Book a Demo](https://opsos.pro/#contact)
Related: [Why Headcount Optimization Starts With Better Data, Not More Cuts](/blog/headcount-optimization-data) | [The Warehouse KPIs That Actually Predict Problems Before They Happen](/blog/warehouse-kpis-predict-problems)
Frequently Asked Questions
QWhat is the operational intelligence layer in a manufacturing technology stack?
The operational intelligence layer sits between data capture infrastructure (sensors, PLCs, scanners) and decision-makers. It aggregates real-time data, calculates live KPIs, surfaces anomalies, and enables cross-shift comparisons. Most facilities have ERP and some sensor data but lack this layer — which is why spreadsheets fill the gap.
QWhy do large MES implementation projects often fail?
Enterprise MES implementations typically fail for three reasons: requirements are backward-looking and outdated by go-live, the systems are configured for process manufacturing rather than high-mix variable-volume environments, and the long implementation timeline means organizational knowledge of the original requirements evaporates before deployment.
QWhat is the right starting point for modernizing operational data infrastructure?
Start with real-time throughput visibility on a single line. This single capability — knowing whether you are on pace right now, not at shift end — provides immediate value and builds the foundation for additional capabilities. Prove value in 60 days before expanding to additional lines.