Getty Images
5 steps for implementing predictive maintenance
Launching a targeted pilot program is one of the steps that C-suite leaders should follow when implementing predictive maintenance at their organization. Learn more.
Implementing predictive maintenance requires the proper planning, collaboration and governance, so leaders such as CSCOs, CFOs and CIOs must make sure their organization has prepared.
When executed properly, predictive maintenance can bring benefits such as reducing unplanned downtime and maintenance costs. However, poor execution can lead to programs not scaling effectively.
The following steps can serve as a framework for C-suite leaders planning a predictive maintenance implementation.
1. Assess the organization’s readiness
An assessment is an important first step. CSCOs, CFOs, and CIOs should treat the assessment as a collaborative process that brings together IT, operations and finance teams.
Begin by thoroughly assessing the organization’s current maintenance practices and bottlenecks. Identify the assets that are most likely to benefit from predictive monitoring, then develop a simple financial model to assess costs before and after implementation. For example, if a machine’s downtime was reduced by 30 percent, how would that affect throughput, on-time delivery and expediting costs?
Also assess the maturity of the company’s existing data collection capabilities and the quality of historical data. The preliminary assessment should also include identifying potential points of integration with existing EAM and ERP systems and factoring in potential savings through automation.
A well-executed assessment typically takes six to 10 weeks.
2. Build a strong data foundation
Predictive models are only as good as the data that feeds them. Organizations must plan enough time for inventorying and cataloging data, auditing its quality and completeness and identifying gaps. Automated data cleansing tools can improve quality. In addition, information collected from IoT sensors and from EAM or ERP platforms could enrich existing data. A unified data pipeline will ensure that the information feeding the predictive maintenance algorithms is consistent, accurate and timely.
Carrying out this step may increase early implementation costs, but a lack of clean, integrated data could lead to false alerts, which will harm the credibility of the predictive maintenance initiative.
3. Select and integrate the right technology stack
Assembling the right technology stack is essential. The stack includes the various sensors for collecting data as well as the core analytics components, connectivity and integration with existing EAM and ERP systems. That integration enables the automation of work order creation, inventory management and financial forecasts based on predictive insights.
Vendor evaluations should be a joint effort led by the CSCO and CIO and should include an assessment of integration capabilities. Ask for proof-of-concept demonstrations that show real-time data flow from sensors into the EAM/ERP environment and look for availability of open APIs. Integration capabilities can result in higher costs, but it can lead to long-term agility and scalability.
4. Launch a targeted pilot program
A pilot that’s focused on a small number of critical assets should precede an enterprise-wide deployment. This will enable project teams to confirm that the technology stack fulfills expectations, refine predictive models if needed and monitor ROI based on real-world scenarios. A pilot will also give project leaders a chance to identify and address any organizational resistance or process gaps before the predictive maintenance program is scaled up.
Prior to pilot launch, CSCOs should identify the most important KPIs, such as downtime reduction, cost savings and predictive accuracy. The performance against those targets will hopefully help sell the overall business case for predictive maintenance. A pilot is also a good time to test and fine-tune EAM and ERP integration.
A well-executed pilot typically lasts between three and six months. It helps reduce overall project risk and can help increase confidence among line-of-business stakeholders that the full-scale predictive maintenance initiative will yield positive results.
5. Measure, monitor and scale
After a successful pilot, the predictive maintenance program will be expanded to a broader collection of assets.
Organizational change management is an important consideration at this point, as is training for employees whose jobs will be affected. Project success requires that maintenance personnel and other workers possess the right skills and knowledge and that employees adopt the new system.
Optimize automated workflows involving integrated EAM and/or ERP systems and fine-tune executive dashboards so they provide real-time visibility into program performance. The costs required for full rollout should lead to lower overall costs over time as well as reduced downtime.
These steps for implementing predictive maintenance can serve as a roadmap for minimizing risks and maximizing ROI.
James Kofalt spent 16 years at SAP working with SME business applications and was a product manager for integration technology at Microsoft's Business Solutions division. He is currently the president of DX4 Research, a technology advisory practice specializing in ERP and digital transformation.
