Blog: Maintenance operations of 2020s rely of efficient data collection, processing, and utilization – but most companies are still stuck in the past
Ever since I started working at Fatman back in 2018, I’ve been amazed by the fact that maintenance operations of basically all major companies are still heavily based on human-generated triggers. It doesn’t matter whether you’re talking to large industrial companies, machine builders, commercial real estate companies, or maintenance service providers, a vast majority of actions around preventive and reactive maintenance operations seem to be completely isolated from the available data.
Let’s take a simple example: A person enters an office during a warm summer day and realizes immediately that air conditioning is not working properly. She reports her findings to the property manager who creates a work order to the maintenance service company. Maintenance service company sends their employee to check out the situation. Serviceman notices that a specific component is broken and orders a spare part from the supplier of the device. Supplier delivers the spare part, and the serviceman finally fixes the air conditioning system.
How long does this process take? Couple of days? A week? Longer?
During this time, the people working in the office are frustrated by the subpar working conditions and the seemingly long response time from the maintenance provider. Property manager gets negative feedback from the tenant and potentially bad publicity for the whole building. And even the maintenance company, who has done everything to solve the problem as fast as possible, might see its brand image take a hit due to factors it has no control over.
Preventive maintenance tasks often seen as easiest way to limit risk of unexpected failures
How to avoid this situation? Traditionally many companies have tackled this problem by planning preventive maintenance tasks that limit the risk of key components breaking down. Changing a component e.g. once a year “just to be safe” is really typical approach that does in fact lower the risks but also often creates extra work for the maintenance company and unnecessary costs for the client.
In this simple example, the harm caused by the unexpected problems is still relatively small. However, think e.g. what the impact of a critical machinery being out of order could be for an industrial company. Production line being down for several days might have a significant effect on the output and major financial implications for the company. Because of this, many large corporations spend a lot of money to proactively maintain the key components of their machines and devices even if there would not be a real need for this work.
Data-driven property management tackles these problems without generating unnecessary costs
We at Fatman feel that the correct solution for this scenario can be divided into four steps:
- Step 1: Systematic and efficient data collection
- Step 2: Data structuring and analysis
- Step 3: Forecasting upcoming needs
- Step 4: Optimizing use of resources
Our approach to data-driven property management takes into consideration the needs of all the above-mentioned segments (industrial companies, machine builders, commercial real estate companies, maintenance service providers). Regardless of the use case, some or all of these steps can be implemented efficiently to tackle the problems described before.
Going back to the simple example. If the company would be using our services to collect real-time data from the air conditioning system, they could 1) react fast to the immediate problem, and 2) analyze what has happened and why. Through systematic data collection, the serviceman would also know beforehand which component is broken and take the right spare part with him when visiting the site. This kind of monitoring of critical components also enables the service company to optimize their stock levels of the most important spare parts.
These steps, often labeled as descriptive analytics, form the basis for the next step where we use this data in our forecasting algorithms to define what might happen in the future. This predictive analytics work is what we do prevent a situation where the machine breaks down unexpectedly and reactive maintenance with potentially long response times is initiated.
In addition, we can take this process even further by providing prescriptive analytics services that generate actionable items (such as work orders) that tell who should do what, where and when. These solutions can be used to optimize e.g. the use of resources, task management or spare part supply. In our example this could mean that we calculate which serviceman should go to the office to replace the component that is expected to break in couple of weeks. Then we would send him an automatic work order that tells where he needs to go, when he needs to go, and what he needs to do. Interesting information on this topic can be found at https://www.gurobi.com/resources/prescriptive-analytics/.
If you feel your company is still doing things the old-fashioned way and would like to modernize your maintenance operations, please contact us at firstname.lastname@example.org to learn more about what can be done with a ready-made set of solutions. We are happy to discuss more about your actual needs and concrete use cases to find the way to make your maintenance operations more efficient without increasing the risks involved.