What customer reliability engineering means for industrial equipment reliability
Customer reliability engineering in one sentence
Customer reliability engineering is an operating model that makes reliability a shared responsibility between the organization that builds or supports equipment and the customer that depends on it in production. The term comes from the software and cloud world, particularly Google Cloud’s 2016 description of Customer Reliability Engineering. For industrial equipment, the useful part is not the software vocabulary itself. It is the discipline behind it: define what reliable service means to the customer, measure it with agreed indicators, prepare systems before handover, and respond to failures with joint learning rather than isolated blame.
For industrial teams, customer reliability engineering should not replace established reliability engineering, maintenance planning, or asset management. Its value is in connecting those disciplines to customer-facing outcomes such as uptime, throughput, safety, quality, repair speed, and lifecycle cost.
Where the concept came from and why it matters
Customer reliability engineering, often shortened to CRE, is closely related to site reliability engineering. In the cloud context, Google described CRE as a way to create shared operational responsibility with customers running critical applications. The model used practices such as production readiness reviews, common monitoring, service level objectives, error budgets, and incident learning.
Industrial equipment is different from cloud software, but the reliability problem has a similar shape. A customer does not only buy a compressor, conveyor, CNC machine, packaging line, or pump skid. The customer is buying a production outcome: safe operation, predictable output, acceptable product quality, and controlled maintenance cost. When equipment fails, the effect is rarely limited to the maintenance department. It can affect delivery commitments, energy use, scrap, labor planning, regulatory obligations, and customer satisfaction further downstream.
Traditional reliability engineering already addresses many of these issues through failure analysis, design margins, maintainability, preventive maintenance, condition monitoring, and lifecycle planning. Customer reliability engineering adds a relationship layer. It asks whether the supplier, integrator, service provider, and asset owner have the same view of what matters, the same data to detect degradation, and the same process for deciding when reliability risk is no longer acceptable.
How CRE differs from traditional support and maintenance
Traditional customer support is often reactive. A fault occurs, the customer opens a ticket, a service team diagnoses the issue, parts are ordered, and the machine returns to service. That model is necessary, but it can miss chronic problems that sit between design, operation, and maintenance. A recurring seal failure, nuisance trip, software interlock, poor spare-part strategy, or operator workaround may never be solved if every event is treated as a separate repair.
Customer reliability engineering changes the question from: how fast did we fix the last breakdown? to: are we operating inside the reliability envelope that the customer’s process requires? That shift matters because many industrial failures are influenced by shared conditions, including installation quality, process loads, lubrication practices, control settings, operator behavior, maintenance intervals, spares availability, and environmental exposure.
It also changes how teams use data. A service dashboard that counts open tickets is not the same as a reliability dashboard. CRE needs indicators that connect asset behavior to customer outcomes. These may include unplanned downtime, mean time between failures, mean time to repair, repeat failures, first-pass quality, start-up losses, alarm frequency, energy deviation, critical spares availability, and unresolved corrective actions from root cause analysis.
Translating SRE ideas into industrial equipment terms
The most useful CRE practices are not copied directly from software. They need to be translated into industrial language, safety requirements, and asset lifecycle realities. The table below shows practical equivalents.
| CRE or SRE concept | Industrial equipment equivalent | Why it matters |
|---|---|---|
| Service level indicator | A measurable signal such as availability, downtime minutes, critical alarm rate, product quality loss, or repair duration | It turns reliability from a general expectation into something teams can observe. |
| Service level objective | An agreed target for the equipment or line, such as maximum unplanned downtime in a period or minimum availability under defined operating conditions | It aligns the supplier and customer on what is good enough for the process. |
| Error budget | The tolerated amount of lost availability, failed starts, rejects, or reliability-related production loss before corrective action takes priority | It helps balance production pressure against reliability risk. |
| Production readiness review | A structured review before commissioning, handover, major modification, or warranty start | It checks whether design, spares, training, procedures, telemetry, and support paths are ready. |
| Post-incident review | Root cause analysis with corrective actions assigned across design, operation, maintenance, and support | It prevents the same failure from being repaired repeatedly without being solved. |
| Shared monitoring | Common telemetry from PLCs, sensors, CMMS, historian, service records, and inspection results | It reduces arguments about facts and speeds diagnosis. |
For example, a 99.5% monthly availability objective allows about 3.6 hours of downtime in a 30-day month. That may be acceptable for one auxiliary asset and unacceptable for a bottleneck machine on a high-volume line. The goal is not to chase the highest possible percentage everywhere. The goal is to set a target that reflects process criticality, safety, redundancy, maintenance access, and economic consequence.
Building a customer reliability engineering model
Start with the customer-critical function
The first step is to define the function that must remain reliable. For a pump, the customer-critical function may be flow at a required pressure. For a packaging machine, it may be stable output at a target speed with acceptable rejects. For a furnace, it may be temperature uniformity, safe operation, and recoverability after a trip. CRE loses value when teams measure what is easy instead of what affects the customer’s operation.
Create a shared reliability baseline
A useful baseline includes asset hierarchy, operating context, criticality, known failure modes, maintenance history, alarm history, spares status, and actual operating load. Standards and guidance such as ISO 55000:2024 emphasize value from assets and the balance of cost, risk, and performance. IEC dependability management guidance also reinforces that reliability, availability, and maintainability need to be managed as part of a system, not as isolated statistics.
Run readiness reviews before the customer carries the risk
A readiness review should take place before a new asset is placed into sustained production, before a major retrofit goes live, or before a service contract depends on performance outcomes. The review should check whether the equipment has defined reliability objectives, verified safeguards, maintainable access, spare parts, operating limits, calibration plans, escalation contacts, and monitoring points. If these basics are missing, a warranty or service agreement may only move the cost of failure from one party to another rather than reducing the failure risk itself.
Use incident reviews for learning, not blame
After a significant failure, the question should not stop at which component broke. CRE asks why the failure reached the customer, why it was not detected earlier, why the response took as long as it did, and which party can remove the condition that allowed recurrence. Some corrective actions may belong to the customer’s maintenance process. Others may belong to the supplier’s design, documentation, firmware, recommended spares list, or field service training.
Metrics that support CRE and metrics that can mislead
The Society for Maintenance and Reliability Professionals publishes maintenance and reliability metrics guidance intended to standardize common indicators. This matters because poorly defined metrics create false confidence. Mean time between failures, for example, is only meaningful when the asset population, failure definition, and operating context are clear. Mean time to repair may hide waiting time for permits, parts, access, or diagnosis. Overall equipment effectiveness can be useful, but it can also mask quality losses or chronic micro-stoppages if its components are not reviewed separately. See also: automation and controls.
A balanced CRE dashboard should include a small number of customer-facing and engineering-facing measures. Too many metrics dilute attention; too few can distort behavior.
| Metric | Useful question | Common limitation |
|---|---|---|
| Availability | Was the equipment able to perform when required? | It may ignore reduced speed, quality loss, or unsafe workarounds. |
| MTBF | Are failures becoming more or less frequent? | It can be misleading across mixed equipment or changing duty cycles. |
| MTTR | How quickly can the asset be restored after failure? | It may combine technical repair time with waiting time and logistics delays. |
| Repeat failure rate | Are teams eliminating causes or repeating repairs? | It requires consistent coding of failure modes. |
| OEE | How do availability, performance, and quality interact? | It may encourage output at the expense of long-term reliability if used alone. |
| Corrective action closure | Are root cause findings actually implemented? | Closure quality matters more than closure count. |
The best CRE metric is often not a single number. It is a decision rule. For example, if repeat failures occur on a critical asset within a defined period, the issue moves from normal support to joint reliability review. If downtime consumes a defined portion of the agreed tolerance, production changes and maintenance deferrals require management approval. Clear thresholds prevent reliability from becoming a topic only after a major outage.
Governance between supplier, service team and asset owner
Customer reliability engineering needs clear boundaries. It should not imply that a supplier controls every operating condition or that a customer can transfer all maintenance responsibility. Industrial reliability is produced by a system of decisions across design, procurement, installation, operation, maintenance, and support.
A practical governance model defines who owns each part of the reliability loop. The equipment manufacturer may own design recommendations, technical bulletins, software updates, and failure mode knowledge. The customer may own operating discipline, routine inspections, lubrication, housekeeping, and access planning. A service provider may own response execution, inspection quality, and repair documentation. Engineering leadership should own recurring risk decisions and capital changes.
This is where CRE can create practical value. It gives teams a forum to compare facts before positions harden. If a gearbox overheats repeatedly, the answer may involve load profile, alignment, oil selection, cooling, installation tolerance, or design margin. Without shared telemetry and a joint review process, each party may defend its own scope instead of solving the failure mechanism.
A practical roadmap for industrial teams
Organizations do not need a large new department to start using customer reliability engineering. A focused pilot on a critical asset or production line is usually more useful than a broad program with vague goals.
- Select a customer-critical asset. Choose equipment where reliability affects safety, output, quality, contractual performance, or major cost.
- Define the reliability outcome. State the required function, operating context, and measurable target in plain language.
- Agree on data sources. Identify which signals come from controls, historians, CMMS records, inspections, service reports, and operator logs.
- Set escalation thresholds. Decide what level of downtime, repeat failure, alarm frequency, or risk triggers a joint reliability review.
- Review readiness. Check spares, procedures, training, access, monitoring, documentation, and support contacts before the next high-risk operating period.
- Run disciplined incident reviews. Separate immediate restoration from long-term corrective action, and track whether fixes prevent recurrence.
- Revisit the targets. Adjust objectives when duty cycles, product mix, safety requirements, or production economics change.
The main limitation is organizational discipline. CRE is not a dashboard purchase or a new label for warranty support. It requires teams to agree on definitions, expose uncomfortable data, fund corrective work, and avoid using metrics as weapons. When those conditions are absent, the term becomes branding. When they are present, it can help convert reliability engineering from a back-office technical activity into a customer-visible operating system.
Frequently asked questions
Is customer reliability engineering the same as site reliability engineering?
No. Customer reliability engineering borrows ideas from site reliability engineering, especially shared reliability goals, readiness reviews, service level objectives, and incident learning. In industrial equipment, those ideas must be adapted to physical assets, safety constraints, maintenance work, spare parts, and process conditions.
Does CRE replace maintenance or asset management?
No. CRE works best as a connecting layer. Maintenance executes inspections and repairs. Reliability engineering analyzes failures and improves asset performance. Asset management balances cost, risk, and value across the lifecycle. CRE aligns those activities with the customer’s operational outcomes.
What is the first metric to define?
Start with the customer-critical function and choose a metric that shows whether that function is being delivered. For a bottleneck machine, that may be unplanned downtime or availability. For quality-sensitive equipment, it may be defect-related loss. For repairable assets, repeat failure rate and restoration time may be more useful than a single average.
Can small manufacturers use customer reliability engineering?
Yes, but the approach should stay lightweight. A small team can begin with one critical machine, one shared reliability target, one readiness checklist, and one recurring review of failures and corrective actions. The discipline matters more than the size of the program.
