How a 380-Person Manufacturer Replaced Excel Production Tracking in 6 Weeks

A 380-person precision components manufacturer in Baden-Württemberg had a familiar problem: SAP ERP held the production plan, but the production floor ran on Excel spreadsheets and a WhatsApp group with 47 members.

This is how we helped them replace Excel production tracking in manufacturing operations, and what it took to get both systems telling the same story.

The Situation: Two Parallel Truths

The company supplies machined components to automotive and industrial equipment OEMs. Three production halls, two shifts, 12 CNC machining centers, and a legacy MES that hadn't been updated since 2019.

The Head of Production, who brought us in, described it simply: "SAP knows what we planned. Excel knows what actually happened. Neither knows what's happening right now."

Their operational setup before we started:

SAP ERP for production orders, BOM management, and customer scheduling
Excel spreadsheets (14 active files) for shift tracking, downtime logging, and quality deviations
WhatsApp group for real-time floor communication, machine status, and urgent escalations
Paper batch records for ISO 9001 traceability requirements
Legacy MES connected to 4 of 12 machines (the rest required manual entry)

The Production Planning Lead spent Monday mornings reconciling SAP planned quantities against Excel actuals from the previous week. Average reconciliation time: 4 hours. Every week.

The Problem: Information Dying Between Shifts

The shift handoff was a notebook. Literally. A spiral-bound notebook passed from Shift 1 lead to Shift 2 lead, with handwritten notes about machine issues, pending quality holds, and customer priorities.

"When something goes wrong on Tuesday night shift, by Wednesday morning the context is gone," the Quality Manager told us. "We find out there was a deviation when the customer calls about a delivery delay."

The numbers told the story:

OEE calculated monthly, 3-4 weeks after the production actually happened
Root cause analysis for quality deviations averaged 18 days (mostly waiting for someone to find the right Excel file)
Downtime logging accuracy estimated at 60-70% (operators logged what they remembered, not what happened)
4.5 hours per week spent by supervisors manually entering data from paper into SAP

The WhatsApp group had become the de facto MES. When a machine went down, someone posted a photo. When a batch was complete, someone typed "Maschine 7 fertig, 847 Stück." This information existed nowhere else.

The ISO 9001 auditor had flagged this in the last surveillance audit. "Documented procedures exist, but evidence of systematic implementation is inconsistent." Translation: you have SOPs, but your actual data lives in WhatsApp messages and personal Excel files.

flowchart TD
    subgraph SAP["SAP ERP (Planned Reality)"]
        PO[Production Orders]
        BOM[Bill of Materials]
        SCHED[Customer Schedule]
    end
    
    subgraph Floor["Production Floor (Actual Reality)"]
        OP1[Operator logs downtime in Excel]
        OP2[Shift lead notes in notebook]
        OP3[WhatsApp: 'Maschine 4 steht'"]
        OP4[Paper batch record filled by hand]
    end
    
    subgraph Monday["Monday Morning"]
        REC[Production Planning Lead reconciles for 4 hours]
    end
    
    PO --> |"Planned quantities"| REC
    OP1 --> |"Actual quantities, maybe"| REC
    OP2 --> |"Lost between shifts"| REC
    OP3 --> |"Nowhere in system"| REC
    OP4 --> |"Scanned, then re-keyed"| REC
    
    REC --> |"Best guess OEE"| REPORT[Month-End Report]

This diagram was not an exaggeration. We mapped it during the first week.

What We Built: A Bridge, Not a Replacement

The Head of Production was clear: "We are not ripping out SAP. We just had a failed MES upgrade project that took 18 months and delivered nothing. I cannot go to the board with another multi-year initiative."

So we didn't propose one.

The goal was to replace Excel production tracking in manufacturing without replacing SAP, without replacing the legacy MES, and without asking operators to learn a new system. We built a coordination layer that sat between what they already had.

For operators: A tablet interface at each machine. One screen. Three inputs: status (running/down/changeover), current production order (scanned from SAP barcode), and any notes. Average interaction time: 8 seconds.

For shift leads: A dashboard showing all 12 machines in real-time. Color-coded status. Tap a machine to see history, current order, and any notes from previous shifts. The notebook became a screen, but the information finally persisted.

For the Production Planning Lead: Automatic reconciliation. SAP planned quantities pulled via RFC, actual quantities aggregated from operator inputs, variance calculated in real-time. That Monday morning 4-hour session became a 15-minute review.

For quality: Every deviation now had a timestamp, machine ID, operator ID, and shift context. Root cause analysis that took 18 days dropped to 3 days (more on the numbers below).

The technical depth: we built a parser that handled their 14 different Excel templates. Why? Because we needed to import historical data for baseline comparison, and each shift lead had created their own format over the years. 14 templates, 3 different date formats, 2 different ways of recording downtime codes. The parser normalized everything into a single schema.

We also integrated with 4 machines via OPC-UA (the ones already connected to the legacy MES) for automatic cycle counting. The other 8 machines remained manual input, but now that input went somewhere useful.

sequenceDiagram
    participant OP as Operator (Tablet)
    participant API as TIMPIA Platform
    participant SAP as SAP ERP (RFC)
    participant DASH as Shift Lead Dashboard
    participant QM as Quality Manager
    
    OP->>API: Status: Running, Order: 4711-A, Qty: 50
    API->>SAP: Query planned qty for 4711-A
    SAP-->>API: Planned: 500, Due: Friday
    API->>DASH: Machine 7: 50/500 complete, on track
    
    OP->>API: Status: Down, Reason: Tool breakage
    API->>DASH: Machine 7: DOWN (tool breakage)
    API->>QM: Alert: Unplanned downtime, review required
    
    Note over API: All events timestamped, linked to shift, persisted
    
    OP->>API: Status: Running, Qty: 75 (after repair)
    API->>DASH: Machine 7: Running, 125/500, projected completion: Thursday 14:00

Timeline: Working prototype with 2 machines in 7 days. Full rollout to all 12 machines in 6 weeks. The 6-week timeline included operator training (which took 2 hours per shift, not the "full day per user" they expected from past software projects).

The Results: Before and After

We measured for 90 days after go-live. Here's what changed:

OEE Visibility

Before: Calculated monthly, 3-4 weeks after production
After: Calculated hourly, visible on shift lead dashboard
Impact: Production Planning Lead now catches performance drops within the same shift, not the next month

Weekly Reconciliation Time

Before: 4 hours every Monday
After: 15 minutes (review only, no manual reconciliation)
Calculation: 3.75 hours saved × 52 weeks × EUR 65/hour (loaded cost) = EUR 12,675/year

Root Cause Analysis Cycle Time

Before: 18 days average
After: 3.2 days average
Calculation: Quality Manager estimated each deviation investigation consumed 6 hours of cross-functional time. With ~40 deviations/year, reducing cycle time freed capacity but the bigger win was catching issues before they propagated to customer shipments.

Downtime Logging Accuracy

Before: 60-70% estimated (based on comparing logged downtime to machine availability)
After: 94% (verified against OPC-UA data from the 4 connected machines)
Impact: Maintenance planning now based on actual failure patterns, not operator memory

Data Entry Elimination

Before: 4.5 hours/week of supervisor time entering paper records into SAP
After: 0.5 hours/week (edge cases only)
Calculation: 4 hours saved × 52 weeks × EUR 55/hour = EUR 11,440/year

ISO 9001 Audit Finding

Before: "Inconsistent evidence of systematic implementation"
After: Surveillance audit 4 months post-go-live: finding closed, auditor specifically noted "electronic records provide complete traceability"

Total estimated annual savings: EUR 24,115 (direct labor only, not counting quality cost avoidance or faster customer response)

The Head of Production summarized it: "We did not replace SAP. We did not replace the MES. We replaced the notebook and the WhatsApp group. That turned out to be the part that mattered."

For context on what a more comprehensive quality system build looks like, we documented a similar project with a 520-person automotive supplier building quality management system software. That project went deeper into IATF 16949 requirements, but the pattern of bridging existing systems rather than replacing them was the same.

Why Excel Replacement Fails (And Why This Didn't)

Most attempts to replace Excel production tracking in manufacturing fail for one of three reasons:

Scope creep into full MES replacement. The project that should take 8 weeks becomes an 18-month initiative requiring IT, consultants, machine vendors, and eventually a steering committee. By the time it launches, the original problem owner has moved on.
Operators reject it. The new system requires 47 clicks where Excel required 3. Operators route around it. Within 6 months, Excel is back.
It doesn't connect to SAP. A standalone tracking tool creates a third system of truth instead of eliminating the second one.

This project avoided all three. Scope was fixed: bridge SAP to floor reality, nothing more. Operator interface was simpler than Excel (one screen, three inputs, 8 seconds). SAP integration was built in from day one.

If you're evaluating whether to replace Excel or build something more comprehensive, we compared the tradeoffs in Production Dashboard Real-Time Manufacturing: Excel vs MES vs Custom. For this client, the custom bridge was the right answer. For others, the calculus is different.

flowchart LR
    subgraph Before["Before: 3 Systems of Truth"]
        SAP1[SAP ERP]
        EXCEL1[14 Excel Files]
        WA[WhatsApp Group]
    end
    
    subgraph After["After: 1 System of Truth"]
        SAP2[SAP ERP]
        BRIDGE[TIMPIA Bridge Layer]
        TABLET[Operator Tablets]
    end
    
    SAP1 -.-> |"Planned"| REPORT1[Month-End Reconciliation]
    EXCEL1 -.-> |"Actual, maybe"| REPORT1
    WA -.-> |"Lost"| REPORT1
    
    TABLET --> |"8 seconds"| BRIDGE
    BRIDGE <--> |"RFC, real-time"| SAP2
    BRIDGE --> |"Instant"| REPORT2[Live Dashboard]

If This Looks Familiar

Most production operations we see have some version of this problem. SAP (or another ERP) holds the plan. Excel (or paper, or WhatsApp) holds the reality. The gap costs hours every week and creates audit exposure.

If your manufacturing operations look like the "before" picture, we can show you the "after" in 7 days. The proof-of-concept costs EUR 3,500, covers 2-3 machines or work centers, and you keep the working prototype regardless of whether you proceed.

Contact us to schedule a technical scoping call. We'll map your current data flows, identify the integration points with your existing systems, and tell you honestly whether a 6-week bridge makes sense or whether you need something different.