Not all maintenance data is created equal
Data: It’s the backbone of any maintenance program. It’s what you use to measure success. It tells you what assets need more attention and how that will impact your schedule. It’s what helps you survive maintenance audits unscathed. In short, data is the language that helps you tell the story of your maintenance team.
But not all data is created equal. And it could be that yours is failing to say what it needs to. Jason Afara, a Senior Solutions Engineer at Fiix, experienced this when he was a maintenance manager.
“We had more technicians than we did CMMS software licenses, so we had people logging in after they had already completed a work order, just trying to fill in all the details they could remember,” he says. “We were always trying to catch up, and that impacted our credibility.”
The cost of bad maintenance data
That’s just it—when your data is off, it’s harder to go to bat for your team. It’s not as easy to justify buying a new piece of equipment, trade production time for maintenance, or make a new hire if the data isn’t there to support that request.
It can impact your team on a day-to-day basis as well. For example, a technician might wait until the end of the day to log completed work. This gap in time could lead them to misremember how long it took them to do a job. Maybe they round down. No big deal, right? Except it is.
That one mistake could cause a domino effect. The next time you go to schedule that job, you plan less time for it. Now the technician is rushing to complete the work, increasing risk for both them and the machine. You’ll also lowball the cost of labor hours in your budget, putting you in a tricky situation with your finances.
Let’s dive into where your data can go wrong, and how you can audit it to start steering things in the right direction.
Where bad maintenance data begins
Bad data is often born from the best intentions. That makes it hard to spot. But there will always be a silver lining to go along with these issues—you have a data-driven culture. You know the numbers are key and the insight you get from them is even more valuable. That’s the most important ingredient for finding and eliminating bad data.
Here are two aspects of maintenance programs that most often contribute to bad or incomplete data.
Trying to boil the ocean
A lot of maintenance teams try to do too much, too soon with their data. Having the ability to track things is great, but if you don’t have a well-thought-out plan in place for what you’re going to measure—and why—you’ll run into problems.
It’s an easy trap to fall into. The advent of IIoT technology, like sensors that track every second of an asset’s behaviour, has introduced seemingly infinite ways to capture data. The trouble for maintenance managers doesn’t come from having too much data, but from not knowing how to pull out the data that matters.
Brandon De Melo, a Customer Success Manager at Fiix, puts it this way, “Let’s say you have a sensor that’s pulling machine data. That’s great, but you can’t stop there. You have to consider all the things that factor into that data, like downtime or other external factors that could affect it.”
Not thinking critically about metrics
Every maintenance team is held to certain KPIs—but are they the right ones? As Stuart Fergusson, Fiix’s Director of Solutions Engineering, points out, it can be easy to get caught in a cycle of tracking a number like labour hours simply because it’s the metric that comes from your boss (or their boss).
It’s important to take a critical lens to maintenance metrics and really think about whether they should be measured.
“At the end of the day, you need to be measuring the metrics that support your department,” says Fergusson. “Not enough people understand why they’re measuring what they’re measuring.”
Where bad maintenance data lives
We know what contributes to bad data, but where does it show up? Bad data is really good at blending in with clean data, so it’s not always obvious. But knowing the telltale signs of inaccurate information will help you spot it without pouring over dozens of reports. Here are the most common places where you can find bad maintenance data.
In your storeroom
Bad data can lurk alongside bearings and motors on the shelves of your storeroom. There are a few ways this can happen.
Firstly, it’s easy to have an out-of-date inventory count if you have obsolete parts sitting on shelves. If you don’t check in on your inventory to make sure it matches up with what’s actually available, you’ll run into problems when you have to pay for a part you weren’t expecting.
And then there’s the danger of fudging the numbers to make the bottom line look better.
“Let’s say it’s near the end of the month and you have to replace a $3,000 part,” says Afara.
“Some maintenance managers will say, ‘You know what? Let’s just wait for that repair so it actually hits our books next month.’ It turns into a bit of a game.” This hesitation can negatively impact the whole business if what’s in the books is valued over what’s actually needed to improve production.”
In your preventive maintenance schedule
Every maintenance team has their regular PMs—but how many of them are actually necessary?
“Maintenance can get really emotional really quickly,” says Afara. “You’ll have what’s called an emotional PM, where the team is doing a regular check just because there was a failure six plant managers ago and no one’s changed it.”
When maintenance teams inherit PMs, it’s easy not to question it, but it’s easy to see how things can snowball and tell an inaccurate story of which work actually needs to be done.
In your work order and asset histories
It doesn’t take much for data to go haywire when documenting work. Attention tends to go to the wrong places when a plant’s priorities are out of sorts.
“What commonly happens is, there’s such a focus on technician time,” says Afara. “A message comes from the top that every minute needs to be accounted for, and the result is that technicians are just making up time on work orders to show that they’ve done the eight hours they’ve been asked to.”
As we touched on earlier, the root problem here is a lack of specific planning. You’re worrying about the metric at the expense of strategy, which results in data that doesn’t tell the truth and can’t be used to drive real change.
In your reports
Every data set has its spikes and dips. The important part is how you’re making sense of the fluctuations that show up in your maintenance reports.
“Do you actually have anything in place to explain why, for example, a drop can happen in September and then happen again in January?” says De Melo.
Without critical analysis or an understanding of what contributed to an anomaly in the data, tracking those fluctuations is useless. You need to understand what happened before you can begin to understand what you could have done differently.
How to audit maintenance data
Now that we have a clearer picture of where maintenance data can go wrong, how can you start fixing it?
The answer will be different for each team, but the right place to start is wherever you’re having a problem with no way to explain why you’re having it.
“Let’s say you can’t figure out why you have so much unplanned downtime, and looking at the data isn’t helping you at all,” says De Melo.
“In this scenario, you’d want to talk to the production manager and start asking questions like, ‘How is this being tracked? Is there a system in place?’ There will always be a process of tracking down the right information, but you can’t just sit there and just twiddle your thumbs, hoping that the answer is going to come to you.”
In terms of creating a data audit checklist, again, your best bet is to approach it from a strategic perspective.
“Sit with some key stakeholders, like plant managers and technicians, and do some brainstorming around what you want to improve and understand better,” says De Melo.
“Once you know what you’re looking for, you can build a checklist that makes sense.”
The best maintenance data is data with a purpose
Taking a critical and thoughtful approach to auditing your maintenance data ensures that everything you’re tracking and analyzing is being examined for a reason. This helps you understand how each piece of data is connected. Then you can make actual improvements to your maintenance program instead of making smaller, less impactful changes around the margins.
“If you really understand your maintenance activity, everything else is just going to flow in behind it,” says Fergusson.
“Your plant leadership may not understand maintenance backlog or OT, but when you tell them that delaying a maintenance window is going to cost another $250,000 in our plant maintenance budget because of X, Y, Z, and you have the right data to back it up, they’ll listen.”
When all is said and done, the data is the easy part.
“If you have the culture and the metrics and the right people and processes in place to track everything, and you just don’t have the actual data, no problem. You can get that up and running in a week,” says Fergusson.
“More often, though, it’s the opposite. You have all the data, it’s all flowing somewhere, and everybody’s looking at different pieces of it, but none of it’s building to a true story.”