Maximo Data Quality Best Practice

Why Maximo Data Quality Degrades Over Time

IBM Maximo is one of the most powerful enterprise asset management (EAM) systems available, but even the best implementation degrades over time without deliberate data quality management. The pattern is predictable: a system is implemented with clean, well-structured data. Over months and years, small inconsistencies accumulate. New assets are added without following the original naming conventions. Failure codes drift as different technicians interpret coding structures differently. Preventive maintenance programs are modified ad hoc without maintaining alignment to the original reliability strategy. Eventually, the organisation finds itself with a system full of data that cannot be meaningfully analysed, reported on, or used to drive decisions.

This degradation is not a technology problem — it is a governance problem. Maximo will faithfully store whatever data users enter. The challenge is ensuring that what users enter is consistent, complete, and aligned with how the organisation needs to use the data.

Hierarchy Best Practices

The asset hierarchy is the backbone of Maximo. A well-designed hierarchy enables clear reporting, accurate cost allocation, and logical navigation. A poorly designed hierarchy makes everything harder.

Parent-Child Consistency

Every asset in Maximo exists within a parent-child hierarchy. Best practice requires consistent rules about what constitutes a parent and what constitutes a child. Common principles include:

• Systems contain assets: A pump system (parent) contains the pump, motor, coupling, and associated instrumentation (children). The system represents the functional unit; the children represent maintainable components.
• Consistent depth: Similar asset types should sit at the same hierarchical depth. All pump systems across the organisation should have the same number of levels, with the same types of components at each level.
• Meaningful parent-child relationships: Every parent-child relationship should reflect a genuine physical or functional containment. Avoid creating hierarchy levels purely for organisational convenience — use Maximo's classification and attribute structures for categorisation instead.

Location vs Asset Hierarchy

Maximo distinguishes between locations (where things are) and assets (what things are). This is a powerful feature when used correctly, but a source of confusion when the distinction is blurred:

Locations represent fixed positions in the facility hierarchy: Site > Building > Floor > Room, or Site > Process Area > Equipment Position. Locations do not move.

Assets represent physical equipment that can potentially be moved between locations: a pump, motor, valve, or instrument. Assets have serial numbers, purchase histories, and condition records that travel with them.

Best practice is to maintain both hierarchies and use the location hierarchy for spatial navigation and cost centre reporting, while using the asset hierarchy for equipment history and reliability analysis. When an asset is removed for overhaul and replaced, the work history stays with the asset, not the location — enabling meaningful reliability analysis of individual equipment items.

Classification Structures

Maximo's classification structure enables consistent attribution of assets by type. A well-designed classification scheme uses a taxonomy that aligns with industry standards (such as ISO 14224 for the petroleum and natural gas industries, or the NEPM equipment taxonomy for the water sector). Classifications should be:

• Mutually exclusive: Each asset belongs to exactly one classification.
• Collectively exhaustive: Every asset type in the organisation has an appropriate classification.
• Consistently applied: The same type of equipment is always classified the same way, regardless of which site or which team created the asset record.

Failure Code Standards

Failure codes in Maximo capture what went wrong when an asset fails. This data is the foundation of reliability analysis, FMECA (Failure Mode, Effects, and Criticality Analysis), and continuous improvement of maintenance strategies. Poor failure coding renders this analysis impossible.

FMECA-Aligned Coding

Best practice aligns Maximo's three-level failure coding structure with FMECA terminology:

• Problem code (Failure Mode): What functional failure occurred? Examples: "fails to start", "reduced output", "excessive vibration", "external leak".
• Cause code (Failure Cause): Why did the failure occur? Examples: "bearing wear", "seal deterioration", "electrical insulation breakdown", "foreign object damage".
• Remedy code (Corrective Action): What was done to fix it? Examples: "replaced bearing", "replaced seal", "rewound motor", "removed foreign object".

This three-level structure, when consistently applied, enables powerful analysis: Which failure modes are most frequent? Which causes drive the most downtime? Which remedies are most effective? This data directly informs maintenance strategy optimisation.

Avoiding Free Text

One of the most common data quality failures is allowing technicians to describe failures in free text rather than using structured failure codes. Free text entries like "pump broken", "fixed leak", or "replaced parts" are essentially useless for analysis. They cannot be aggregated, trended, or compared.

To minimise free text reliance:

• Design comprehensive failure code libraries: Cover the realistic range of failure modes for each asset class. If technicians consistently cannot find an appropriate code, the code library needs expanding.
• Make failure codes mandatory: Configure Maximo workflow to require failure code entry before a work order can be closed.
• Provide a structured "Other" option: Include an "Other — see comments" code for genuinely unusual failures, but monitor its usage. If "Other" exceeds 10–15% of entries for any asset class, the code library is inadequate.
• Train technicians: Ensure every technician understands why failure codes matter and how to select the correct codes. Contextual help within Maximo (descriptions and examples for each code) significantly improves coding accuracy.

PM Program Alignment

Preventive maintenance programs in Maximo should reflect a deliberate reliability strategy, not historical habit. Too often, PM tasks accumulate over time without review, leading to over-maintenance of some assets and under-maintenance of others.

Linking Tasks to Failure Modes

Every PM task should be traceable to a specific failure mode it is designed to prevent or detect. If a PM task cannot be linked to a documented failure mode, it should be challenged. This traceability ensures that:

• No significant failure mode lacks a corresponding maintenance task.
• No maintenance task exists without a clear reliability justification.
• Changes to maintenance strategy are deliberate and documented.

RCM-Based Intervals

PM intervals should be based on reliability analysis (RCM, FMECA, or equivalent), not manufacturer recommendations or historical practice alone. Reliability-centred maintenance analysis considers the failure mode, its consequences, and the most effective intervention strategy:

• Condition-based tasks: For failure modes with detectable degradation, specify the inspection technique, acceptance criteria, and inspection interval based on the P-F interval (the time between a detectable defect and functional failure).
• Time-based tasks: For failure modes with age-related degradation patterns, specify the replacement or overhaul interval based on failure distribution data.
• Run-to-failure: For failure modes with low consequence and no cost-effective preventive task, document the deliberate decision to run to failure.

Maximo's PM module should be configured to reflect these distinctions, with job plans that include clear task descriptions, acceptance criteria, and links to relevant procedures.

Work Order Data Hygiene

Work orders are the transactional heart of Maximo. Every work order represents a maintenance event, and the quality of work order data determines the quality of maintenance analytics.

Mandatory Fields

Configure Maximo to enforce completion of critical fields before work order closure:

• Actual start and finish dates/times: Essential for labour utilisation analysis and scheduling optimisation.
• Failure codes (for corrective work): Problem, cause, and remedy codes as described above.
• Actual labour hours: Required for cost analysis and future job estimation.
• Materials used: Enables spare parts demand forecasting and inventory optimisation.
• Work type: Distinguishes corrective, preventive, predictive, and improvement work for strategy effectiveness analysis.

Technician Training

Data quality is ultimately a human problem. Technicians who understand why data quality matters and how their entries are used are far more likely to provide good data than those who see data entry as pointless administration. Effective training covers:

• How failure code data drives maintenance strategy improvements that directly affect their work.
• How accurate time recording helps the planning team schedule work more effectively.
• How materials recording prevents stockouts of critical spares.
• Practical demonstrations of reports and dashboards that use the data they enter.

Data Governance Framework

Sustainable data quality requires ongoing governance, not one-off clean-up projects. A Maximo data governance framework should include:

Roles and Responsibilities

• Data owner: A senior manager accountable for Maximo data quality across the organisation. Typically the asset management or maintenance manager.
• Data stewards: Subject matter experts responsible for maintaining data quality within specific domains (hierarchy, failure codes, PM programs, spare parts). They review data quality reports, investigate issues, and implement corrections.
• Data custodian: The Maximo system administrator responsible for configuration, access control, and technical data management.

Review Cycles

• Weekly: Automated data quality reports highlighting incomplete work orders, missing failure codes, and overdue PMs. Reviewed by maintenance supervisors.
• Monthly: Data steward review of trending issues, failure code usage patterns, and hierarchy change requests.
• Quarterly: Data governance committee meeting to review KPIs, address systemic issues, and approve changes to coding structures or hierarchy standards.
• Annual: Comprehensive data quality audit and benchmarking against the organisation's data quality targets.

KPIs for Data Quality

Measure what matters:

• Work order completion rate: Percentage of closed work orders with all mandatory fields completed (target: >95%).
• Failure code usage rate: Percentage of corrective work orders with valid failure codes (target: >90%).
• "Other" code usage: Percentage of failure code entries using "Other" (target: <10%).
• PM compliance: Percentage of scheduled PM tasks completed on time (target: >90%).
• Hierarchy compliance: Percentage of new assets created following hierarchy standards (target: 100%).

Preparing Maximo Data for AI and Analytics

Organisations increasingly want to apply artificial intelligence and advanced analytics to their Maximo data — for predictive maintenance, reliability analysis, or strategic planning. The quality of analytical outcomes is directly determined by the quality of the underlying data.

Key preparation steps include:

• Standardise failure codes: Inconsistent failure coding makes it impossible for ML models to learn meaningful patterns. Consolidate duplicate codes and ensure consistent usage across sites.
• Clean asset hierarchies: Ensure assets are correctly classified and positioned in the hierarchy. Misclassified assets will distort any analysis performed at the asset class level.
• Fill data gaps: Identify critical gaps in historical data (periods with no failure records, assets with no work history) and document them so analytical models can account for missing data rather than treating it as evidence of no failures.
• Link operational data: Where possible, link Maximo maintenance data with operational data (production records, sensor data, SCADA historians) to enable models that correlate maintenance events with operating conditions.

For more on preparing asset data for analytics, see our data analytics services.

Quick Wins vs Long-Term Improvement

Organisations with significant Maximo data quality issues should pursue both quick wins and long-term systemic improvement:

Quick Wins (1–3 Months)

• Enable mandatory fields on work order closure to prevent new bad data from entering the system.
• Run automated reports to identify and close orphaned work orders, duplicate asset records, and obviously incorrect data.
• Conduct a focused training session for all Maximo users on failure code selection and work order completion.
• Appoint data stewards with clear accountability for data quality in their domain.

Long-Term Improvement (3–12 Months)

• Redesign failure code libraries based on FMECA analysis for each critical asset class.
• Review and align PM programs with RCM-based maintenance strategies.
• Implement a formal data governance framework with regular review cycles and KPIs.
• Standardise asset hierarchy structures across all sites and business units.
• Build automated data quality dashboards that provide real-time visibility of data health.

SAS-AM's Maximo consulting services help organisations at every stage of this journey, from initial data quality assessment through to ongoing governance support. For organisations looking to leverage their Maximo data for advanced analytics, our data analytics team can help bridge the gap between raw CMMS data and actionable insights. You may also find our Maximo data quality guide a useful starting reference.