Outsourcing Condition Monitoring: SLA Design for Condition Monitoring Service Indonesia

Direct answer (AEO): Most outsourced condition monitoring contracts fail for a reason that has nothing to do with technical capability: the service level agreement specifies activity rather than outcome. A contract that promises a monthly vibration survey at specified points on specified machines will be delivered faithfully and will produce no reliability improvement, because completing a survey is not the same as finding and acting on a defect. A well-designed agreement specifies different things — data coverage and continuity, a defined response time to critical findings, analyst competence level, reporting that states recommendations rather than observations, and a mechanism for the plant to confirm that findings were actioned. In Indonesian practice, the difference between a contract that quietly becomes a monthly report nobody reads and one that genuinely reduces unplanned failures is almost entirely in these clauses.

Reviewing a condition monitoring contract? Contact Tiaravib — email info@tiaravib.com or WhatsApp +62 821-1378-3727.

Why Activity-Based Contracts Fail Quietly

An activity-based agreement is comfortable for both parties. The supplier knows exactly what must be delivered and can price it precisely. The plant knows what it will receive and can verify it. Performance is measurable in a trivially objective way: was the survey done, were the reports delivered, was the schedule met.

The problem is that all of these can be satisfied while the plant’s reliability does not improve at all. A supplier can visit every month, measure every point on the route, and issue a report that lists readings against thresholds — without ever identifying a developing defect, because identifying a developing defect requires looking at trends, correlating techniques, and knowing the machine’s history. The contract was fulfilled. The benefit was not delivered. And because the contract has no measure that would detect the difference, the arrangement continues indefinitely.

This failure mode is common enough that it should be designed against explicitly. The practical test of a monitoring contract is not whether the surveys happened but whether the plant can point to defects detected early. If nobody can name a defect this service found before it became a failure, the service is producing reports rather than reliability, and it will continue to do so for as long as the contract measures activity.

Specifying the Things That Matter

A defensible agreement specifies the following, and each clause exists because omitting it creates a specific known failure.

Coverage and interval, stated precisely. Which assets, which measurement points, which techniques, which interval. Precision matters because the ambiguity in “monthly vibration monitoring of critical rotating equipment” is where scope quietly shrinks. An asset list in an annex, with measurement points referenced to drawings or tagging, removes the ambiguity. The service selection process should produce this annex before pricing is discussed.

Data continuity requirements. Coverage is not the same as continuity. A supplier that misses surveys during a busy period, or changes transducer locations, has broken the trend record even though the count of surveys might still be met. The contract should require that every scheduled measurement is completed or explicitly recorded as inaccessible with a reason, that measurement locations and mounting methods are documented and unchanged, and that the data history is retained and available.

Competence level. Specify the analyst qualification required to interpret and report. ISO 18436-2 provides a recognised framework for vibration analyst categories, and referencing a category in the contract is more meaningful than a general statement about experience. A report signed by a Category II analyst carries different weight from one produced by an unqualified technician running software.

Response time to critical findings. A critical finding discovered on a survey that is reported three weeks later in the monthly report has wasted most of its value. The contract must define a critical finding, a notification deadline (hours, not weeks), the communication channel, and the named recipient at the plant. This is the single clause that most distinguishes a functioning service from a reporting service.

Reporting that carries recommendations. A report that states “bearing defect frequencies present” without saying what to do is an observation, not a finding. The agreement should require that every significant finding carries a recommendation: inspect, replace, reduce load, re-sample at shorter interval, plan intervention within a stated timeframe. This is also what makes the report actionable for a plant team without deep analytical expertise.

The Escalation Clause Nobody Writes

The most valuable clause in a monitoring contract is the one that almost never appears: what happens when the supplier identifies a critical defect and the plant does not act on it.

This situation is common and it is where most of the potential value of condition monitoring is lost. The analysis is correct, the finding is valid, the notification is timely, and the plant is too busy, or the spare is not available, or the production pressure is too high, or the recommendation conflicts with a plan someone else owns. The machine fails. In the post-incident review, the supplier is asked why nothing was done, and the honest answer is that the plant was informed — twice.

A well-designed contract anticipates this. It requires the supplier to escalate un-actioned critical findings in writing to a named level of plant management after a defined period. This protects the supplier, but more importantly it creates a mechanism that makes inaction visible while it can still be corrected. It also changes the internal conversation: a maintenance manager who knows a critical finding will escalate to their director in five days has an entirely different incentive to resolve it on day one.

The clause is unpopular with both sides during negotiation — the plant does not want a supplier escalating over its head, and a supplier does not want the conflict. But its absence is precisely why outsourced monitoring so often underdelivers, and the plants that get real value from these services almost always have some version of it.

Pricing Models and What They Reward

The commercial structure drives behaviour more reliably than the service description. Three models are common in the Indonesian market and they differ substantially in what they incentivise.

Pay per survey or per point. Simple, transparent, and it rewards completing surveys. It does not reward finding defects, and under it a supplier is economically indifferent between a report that flags a developing problem and one that lists readings within thresholds. It is appropriate for genuinely routine coverage of a large, low-consequence asset population where the output really is data collection.

Fixed scope with response commitment. A defined annual fee for a defined scope, with the response-time and notification clauses described above. This is the most common structure for critical machine coverage and it is the one that most closely aligns incentives: the supplier’s cost of delivering the baseline scope is fixed, so the marginal effort of paying attention to a trending machine is a cost to the supplier with no additional revenue — but the response commitment and escalation clauses make ignoring it commercially risky. Combined with periodic performance review, this structure generally produces better outcomes than per-survey pricing.

Outcome-linked. A portion of the fee tied to measurable reliability outcomes, such as unplanned downtime reduction or the count of defects detected early. These are attractive in principle and difficult in practice, because isolating the service’s contribution from everything else the plant does is contested. Where they are used, they usually work best on a narrow, well-defined measure — for example, the number of findings communicated before functional failure, verified by the plant.

Whichever model is chosen, the contract should make the data and history the plant’s property, with export in a defined format on request and at contract end. A supplier that holds the only copy of the trend history has a lock-in that no exit clause can remedy, because the value is in the history rather than in the current report. This matters directly when comparing contract, periodic, and remote service options, because the data ownership position differs between them and should be a selection criterion rather than an afterthought.

How to Tell Whether It Is Working

Contract performance should be reviewed against a small number of measures that a plant can actually verify, and the review should happen quarterly rather than annually.

The most telling measure is defects detected before functional failure, verified by the plant. Not findings reported — findings that were subsequently confirmed by inspection or by the failure that was avoided. If this number is zero or near zero over twelve months, the service is not detecting anything, and either the scope, the interval, or the analyst competence is inadequate. The second measure is notification timeliness, measured from the survey date to the notification date for critical findings, which tests the clause that matters most. The third is continuity: the proportion of scheduled measurements actually completed, with inaccessibility recorded and justified. The fourth is action closure — of the recommendations issued, how many resulted in work being performed. A low closure rate is sometimes the plant’s failure, which is precisely what the escalation clause exists to surface.

These four measures are simple, verifiable from records the plant controls, and collectively they reveal whether the arrangement is producing reliability or producing paperwork. Any plant that tracks them for two quarters will know which it has.

Related practice for Indonesian operators is set out in condition monitoring data quality and criticality and FMECA.

Key Takeaways

• Activity-based contracts are fulfilled perfectly while delivering no reliability benefit, because completing a survey is not the same as finding a defect.
• Specify coverage precisely, require data continuity, name the analyst competence level, and demand recommendations rather than observations.
• Critical findings must be notified within hours to a named recipient, not included in a monthly report.
• The escalation clause for un-actioned critical findings is the most valuable and most frequently omitted term.
• Per-survey pricing rewards activity; fixed scope with response commitments aligns incentives far better for critical machinery.
• Verify performance on defects detected before functional failure, notification timeliness, continuity, and action closure — not on reports delivered.

Frequently Asked Questions

What should a condition monitoring SLA actually specify?

Coverage with an asset list and measurement points, data continuity requirements, analyst competence by ISO 18436-2 category, a defined response time to critical findings with a named recipient, reporting that carries recommendations, data ownership and export terms, and an escalation path for un-actioned findings.

Why is per-survey pricing often a bad model?

Because it rewards completing surveys rather than finding defects. A supplier under that model is economically indifferent between a report flagging a developing problem and one listing readings within thresholds. For critical machinery, fixed scope with response commitments aligns incentives much better.

What is the most important clause most contracts omit?

The escalation path for critical findings that the plant does not act upon. Most lost value in outsourced monitoring comes from valid findings that nobody actioned, and a written escalation to plant management after a defined period is what makes inaction visible while it can still be corrected.

How do we know if the service is actually working?

Count defects detected before functional failure and verified by the plant. If that number is near zero over twelve months, the service is not detecting anything regardless of how many reports were delivered. Supplement with notification timeliness, measurement continuity, and the proportion of recommendations that resulted in work.

Who should own the condition monitoring data?

The plant. A supplier holding the only copy of the trend history creates a lock-in that no exit clause can fix, because the value of condition monitoring is in the history rather than in the latest report. Require full export in a defined format on request and at contract end.

Is an outcome-linked fee structure worth pursuing?

It can be, if the measure is narrow and verifiable, such as the count of findings communicated before functional failure and confirmed by the plant. Broad outcome measures like downtime reduction are contested because attributing the improvement to the service rather than to everything else the plant did is difficult.

To discuss scope, SLA, or contract structure for a monitoring service, contact Tiaravib at info@tiaravib.com or WhatsApp +62 821-1378-3727.

Outsourced condition monitoring succeeds or fails on specification quality rather than on supplier capability. Tiaravib has provided condition monitoring services to Indonesian industry since 1995, and the engagements that produce measurable reliability improvement are consistently the ones where the plant defined outcomes, response times, and accountability clearly at the outset. Relevant references include ISO 18436-2 for analyst competence categories, ISO 17359 for programme structure, and SMRP material on maintenance service metrics.