MES Software: Vendors, Features & Costs Compared 2026
MES software compared: vendors, functions per VDI 5600, costs (cloud vs. on-premise) and implementation. Honest market overview 2026.
By Christian Fieg · Last updated: April 2026
What is Repairs Per Thousand (RPT)?
Repairs Per Thousand (RPT) measures the number of repair events per 1,000 units produced, sold, or in the field. The formula: RPT = (Number of repairs ÷ Number of units) × 1,000. If 12,000 vehicles are produced in a month and 84 require rework or warranty repair, RPT = (84 ÷ 12,000) × 1,000 = 7.0 RPT. The metric originated in the automotive industry — specifically in vehicle assembly and warranty management — where it serves the same purpose as PPM does for components: expressing defect rates as whole numbers at volumes where percentages become inconveniently small. RPT is used in two fundamentally different contexts: internal RPT (rework events inside the plant before the product ships) and field RPT (warranty repairs after the product reaches the customer).
Both are valid uses, but they measure different things and have different cost implications. An MES directly captures internal RPT — every rework event logged at a repair station, linked to the unit's serial number, the defect type, and the original production station — and provides the data foundation for reducing both internal and field RPT through root cause analysis.
How does RPT differ from PPM, FPY, and scrap rate?
Manufacturing quality has multiple metrics, each measuring a different aspect. Confusing them leads to wrong conclusions:
| Metric | What it counts | Scale | Key difference from RPT | MES data source |
|---|---|---|---|---|
| RPT | Repair events per 1,000 units. Counts units that were defective but repaired — not scrapped. One unit repaired twice = 2 repair events (some definitions) or 1 unit repaired (other definitions). The counting convention must be defined. | Per 1,000 | — | MES rework station logs: unit scanned → defect classified → repair performed → retest result logged. |
| PPM | Defective parts per 1,000,000. A part is either good or defective — binary. Does not distinguish between scrapped and repaired. | Per 1,000,000 | PPM counts defective units regardless of whether they were repaired or scrapped. RPT counts only the repaired ones. A unit counted in RPT may or may not be counted in PPM, depending on whether it passed after repair. | MES good/reject counts per station. |
| First Pass Yield (FPY) | Percentage of units that pass inspection on the first attempt without rework. FPY = Good first-time ÷ Total produced. | Percentage | FPY captures the full picture: scrap + rework. RPT captures only rework. A plant can have 95 % FPY with low RPT (because most failures are scrapped, not repaired) or 95 % FPY with high RPT (because most failures are repairable). | MES: first-pass good count vs. total count at each station. |
| Scrap rate | Units discarded as irrecoverable, as a percentage of total production. | Percentage | Scrap rate counts units that cannot be repaired. RPT counts units that can be and are repaired. Together, scrap + rework = total non-conforming output. | MES reject counts classified by disposition: scrap vs. rework. |
The critical relationship: FPY = 100 % − (Scrap rate + Rework rate). RPT is the rework component expressed per 1,000. A plant that reduces RPT by pushing failures from rework to scrap has not improved quality — it has only shifted the cost. Conversely, a plant that reduces RPT by fixing the root cause of the defect has genuinely improved. The MES makes the distinction visible: it tracks not just how many units were repaired, but why they needed repair, at which station the defect originated, and whether the root cause was addressed.
What are the two types of RPT?
| RPT type | What it measures | Cost per event | Data source |
|---|---|---|---|
| Internal RPT (plant RPT) | Rework events inside the plant — units that failed in-process or end-of-line inspection but were repaired before shipment. The customer never sees these defects. | Low to medium: rework labour + rework material + lost throughput at the repair station. Typically €5–€50 per event for component repairs, €50–€500 for complex assembly rework. | MES captures this directly: rework station logs with unit ID, defect type, originating station, repair action, repair time, retest result. At Neoperl, quality data analysis (which includes rework tracking) contributed to 15 % scrap reduction. |
| Field RPT (warranty RPT) | Warranty repair events after the product has been delivered to the customer. The defect escaped all internal checks. This is the RPT that automotive OEMs track on their warranty scorecards. | High to very high: dealer labour + replacement part + logistics + goodwill + potential recall if systematic. In automotive, a single warranty repair event costs €100–€2,000+. A recall can cost millions. | Warranty management system (not MES). But the MES provides traceability: when a field failure is reported, the MES can identify which machine, shift, and process parameters produced that specific unit — enabling targeted root cause analysis. |
Worked example: An automotive electronics assembly plant produces 50,000 control units per month. Internal inspection catches 350 defective units; 280 are reworked and pass retest, 70 are scrapped. Internal RPT = (280 ÷ 50,000) × 1,000 = 5.6 RPT. Of the 49,930 shipped units, the OEM reports 8 warranty claims in the first 12 months. Field RPT = (8 ÷ 49,930) × 1,000 = 0.16 RPT. The plant focuses on reducing internal RPT from 5.6 to 3.0 (cost saving: 130 fewer rework events × €30 = €3,900/month). The OEM focuses on reducing field RPT from 0.16 to 0.05 (cost saving: 5 fewer warranty events × €800 = €4,000/month). Same RPT metric, completely different scale, cost, and management context.
How does an MES reduce RPT?
| MES capability | How it reduces RPT | Without MES |
|---|---|---|
| Defect origin tracking | MES links every rework event to the originating station, machine, shift, and order. The repair station logs which station produced the defect — not just that the defect exists. This enables targeted root cause analysis: "80 % of solder defects originate from wave solder machine #3 during night shift." | Rework is logged (if at all) without traceability to the originating station. The repair technician fixes the unit but nobody investigates where the defect came from. The same defect repeats tomorrow. |
| Rework Pareto | MES generates a Pareto of rework reasons automatically: defect type × frequency × originating station. The top 3 defect types typically account for 60–80 % of all rework. The CI team attacks them in order. At Neoperl, structured quality data analysis identified exactly these patterns. | Rework reasons are hand-written on paper tags. Nobody aggregates them. The repair station is busy, not analytical. The Pareto is never generated. |
| Process parameter correlation | MES process data module correlates rework events with process parameters at the originating station. "Units reworked for insufficient solder were produced when wave solder temperature was below 248 °C." Root cause found — corrective action: tighten temperature control. | Process parameters and rework data live in separate worlds. The correlation exists in the physics — but nobody can see it because the data is not connected. |
| Digital Poka Yoke | MES prevents the defect from being created in the first place: routing enforcement (no station skipped), process parameter monitoring (alarm if out of tolerance), order-to-machine validation (correct program loaded). Prevention RPT = 0 for that failure mode. | Defects are detected at end-of-line inspection, repaired, and shipped. The plant has a busy repair station and a high RPT — but it "works." Until the OEM starts counting. |
| Warranty traceability | When a field RPT event occurs, the MES provides the production history of that specific unit: which machine, which shift, which process parameters, which material batch. At Meleghy, bidirectional SAP integration enables this order-to-machine traceability across 6 plants. | The OEM reports a warranty claim with a serial number. The plant cannot reconstruct how that unit was produced. The 8D report says "operator error" because there is no data to prove anything else. |
FAQ
Is RPT the same as "things gone wrong" (TGW)?
In automotive, "Things Gone Wrong" (TGW) is a closely related metric used in customer satisfaction surveys (e.g., J.D. Power Initial Quality Study). TGW counts the number of problems reported by customers per 100 vehicles — a broader measure than RPT because it includes non-repair complaints (e.g., "infotainment system is confusing" is a TGW but not a repair). RPT is a strict subset: it counts only events that required a physical repair. In manufacturing, RPT is the more actionable metric because every RPT event has a traceable defect with a root cause that can be eliminated.
What is a good RPT target?
It depends entirely on the product complexity and the definition (internal vs. field, repair events vs. repaired units). For automotive final assembly, internal RPT targets of 10–30 RPT are common (meaning 10–30 rework interventions per 1,000 vehicles assembled). For electronic component assembly, internal RPT below 5.0 is typical for a well-run line. Field RPT in automotive is measured differently — J.D. Power IQS reports industry averages around 150–200 TGW per 100 vehicles (which includes non-repair items). The only meaningful benchmark is the trend: is your RPT going down month over month? If yes, the continuous improvement process is working.
Should RPT count repair events or repaired units?
This is a definition question that must be agreed upon before using the metric. If one unit is repaired twice (e.g., first solder repair, then a second solder repair because the first attempt failed), does that count as RPT = 1 or RPT = 2? The automotive convention is typically to count repaired units (RPT = 1), because the customer experiences one defective product regardless of how many repair attempts were needed. But for internal process improvement, counting repair events (RPT = 2) is more useful — it captures the rework effort and exposes first-time-fix-rate problems at the repair station. The MES can track both: it logs each repair event individually and links them to the unit's serial number.
How does RPT relate to OEE?
RPT feeds the Quality component of OEE indirectly. OEE Quality = Good parts ÷ Total parts. A reworked part that passes retest is counted as a good part in OEE — it was produced, it shipped. But the rework consumed time at the repair station, labour, and potentially slowed the main line (if repair is inline). This means a plant can have 99.5 % OEE Quality and still have a high RPT. OEE does not capture rework cost — RPT does. That is why both metrics are needed: OEE for throughput efficiency, RPT for rework burden. The MES calculates both from the same data: cycle counts for OEE, repair station logs for RPT.
Related: PPM · NTF Rate · Quality Control · Poka Yoke · Root Cause Analysis · Six Sigma · OEE Explained · SYMESTIC Production Metrics · SYMESTIC Process Data · MES: Definition & Functions
About the author
Christian Fieg
Head of Sales at SYMESTIC. Six Sigma Black Belt. At Johnson Controls, managed global MES and traceability systems where every rework event was logged, traced to its originating station, and fed into the weekly RPT review — because the customer (the OEM) counted every repair, and the contract had RPT penalty clauses. Author of OEE: Eine Zahl, viele Lügen. · LinkedIn
Start working with SYMESTIC today to boost your productivity, efficiency, and quality!
