The State of Maintenance Intelligence

Hello, everyone, and welcome. Thank you so much for spending part of your day with us. We know that stepping away from the workday even for a little bit is never quite as easy as it sounds, so we really truly appreciate you being here today. Some quick introductions before we dig in.

I'm Nicole Walsh. I'm head of research at KixTAN, and my team designed and fielded this year's state of maintenance study together with Limble. So I spent the last couple of months up to my elbows in this data, and I'm here today as your guide to what it says. Amanda?

Good morning or good afternoon depending on where you're joining from. My name is Amanda Myers. I have the pleasure of serving as Limble senior director of customer and product marketing. We know that this industry is changing. And in February 2026, we published our own report, but we knew there was more to the story. So we reached out to our friends at KickStand to see what else we could uncover about the maintenance industry. Maintenance pros rarely get to step away for...

To see industry wide numbers, and that's really what this number is for. Gary just talked about real clean data as the foundation.

This session is that idea with 686 data points behind it. If you have any questions, feel free to drop them in the chat, and we're going to get to as many as we can. Also, we've got something going live during this session, so you'll wanna stick around.

Yes. Definitely don't go anywhere. So Don't go anywhere. So here's the crux of what we're gonna be talking through today.

If you ask a maintenance team where they sit on the maturity ladder from reactive, preventive, predictive, most are gonna be able to answer without a lot of hesitation. So this summer, we checked to see how those answers actually hold up. We surveyed 686 maintenance and operations professionals across The US, split almost dead even between the people who plan the work and the people who perform it. And we asked where they would place themselves.

And then separately, we scored how the work actually runs. And those two pictures do not match. That gap between perception and reality is one of the main themes that we're gonna be discussing today.

So before we jump into our findings, we would love to have you in this data as well. So we're launching a poll right now. You should see it in the chat. If you had to place your maintenance operation on the ladder today, where would you say that it sits?

Reactive means you fix what breaks. Preventive means you execute planned work on a schedule that you mostly protect, or predictive means the data tells you what's going to fail before it ever does. Go with your gut. Nobody's grading you yet.

While those votes come in, let me just tell you how the study worked. Every one of our 686 respondents placed themselves on that same ladder exactly like you're doing right now, but they answered three questions about how the work actually runs as well.

One is how maintenance hours are actually spent from all reactive to substantially predictive. Two, how the team tracks and records its work from all paper to all digital. And three, how many core maintenance metrics they can estimate with confidence?

Those answers produced a score from zero to nine, completely independent of whatever label they had picked for themselves. Zero to two landed you on the reactive run. Three to six, preventive. Seven to nine, predictive.

Industry, company size, budget were not scoring inputs here. The score is pure behavior. What your team actually did, not what it can afford.

Excellent. Thanks, Nicole. And why behavior and not labels? The labels come from conference talks and vendor decks. The score comes from what the team actually did last month. Exactly the real clean data Gary opened the day with. Take a look at how this room is voting and hold on to your own answer because here's what happened when we scored 686 of your peers.

So here are those two pictures side by side. On the left, how teams describe themselves, 22% said reactive, 48% preventive, 30% said predictive. Now on the other side, the exact same team scored on behavioral alone. Reactive does not hold at 22. It drops down to eight.

Preventive does not hold at 48. It swells up to 68. And then predictive settles in at 24.

Now two big things stick out to me looking at this chart, which is roughly two thirds of the teams calling themselves reactive have built a lot more structure really than it gives themselves credit for. And nearly everybody, two out of three maintenance organizations in America is standing on that same middle rung. Amanda, you saw this breakdown before basically anybody else. What did you make of it?

Well, teams certainly grade themselves harder than the actual scoring does. When a team self identifies as reactive, it can mean we had a rough month and not necessarily we have no system. The good news, most of you are probably further along than you think. Reactive is a day planned by whatever broke overnight.

Preventive is a schedule you mostly project...

Protect, and predictive is the data flagging the bearing before it even fails. Many teams have never heard the stages defined behaviorally before, which is why exactly the self misplacement is is an issue. The gap exists because no one's ever had a shared behavioral definition of the stages. Teams borrowed the language without a practical benchmark to measure themselves against.

That benchmark can now be a yardstick. My first observations looking at the data, some teams who said they were reactive have behaviors suggesting they're more advanced than they realize. The gap is really a lack of awareness of what each stage looks like from the inside, An awareness is fixable in a way that budget isn't. Here's the catch.

That middle rug Nicole mentioned, that preventive space isn't just crowded. It's sticky. Most teams climb, get there, and stop. That's really what this session is about.

Getting from reactive to preventive is a discipline issue. Write the schedule, protect it from interruptions. It's not easy, but it's largely about forming habits. Getting from preventive to condition based and then predictive, that's another thing entirely.

It's an infrastructure problem. It takes asset history deep enough to show patterns and data consistent enough to trust. So, basically, the first climb asks a team to change their habits. The second asks those habits to have been in place long enough to leave a traceable pattern behind.

Yes. And so I think then the next question that we have to be asking is exactly what is it costing everybody to be stuck on that middle rung.

So let's try to put a price on that plateau starting with what is, in my opinion, maybe the most quotable number in this entire report. 54% is the average preventive maintenance compliance rate or the estimated percentage of scheduled upkeep that actually gets done.

Now what makes that number maybe sting a little bit is that it's not measured against an industry standard or even like some consultant stretch goal. This is an estimate of completion against the team's own schedule, which means a plan that they wrote for themselves, which by definition puts that target at 100%. So nearly half of the maintenance that teams themselves decided was necessary just doesn't happen. Amanda, I know you talk to maintenance teams for a a lot of a day. Where does that other 46% go?

Listen. Urgency beats importance every time. Reactive work is loud. A line is down. Production is waiting. Preventative work, by contrast, is silent when it slips. Nothing appears to happen when a PM slips a week until it does.

Benchmark framing, not shame framing. If you are at 54%, you're exactly average. And now you have a number to measure against. What customers actually describe is that nobody decides to skip a PM. The same technician gets pulled to a line that's down, and the day makes the decision for them. If you can't state your compliance number confidently, that's really the first fix. You can't protect a schedule you don't measure, and confidence in that compliance to your PM completion rate, confidence in the metrics was literally one of the three scoring inputs for this question.

Yes. And when that PM does slip, here is what that bill looks like once it comes to you.

Among leaders, 47% log 10 of unplanned downtime in a typical month. 24% regularly blow right past twenty hours, and twenty hours a month is roughly 240 a year, which is close to six weeks of production time for a single shift operation. In dollars, four in 10 liters put their annual downtime cost at $100,000 or more. And for the manufacturers with us today, and we know that's likely quite a bit of this room, you are apparently carrying the heaviest end of this because manufacturing respondents were 27% more likely than average to land in that 100,000 plus group. What do you make of that, Amanda?

The 100,000 figure is a big one.

It's the number that finance sees and the number that gets maintenance seat in budget conversations. Even more importantly, it costs more than just money. 61 percent of respondents say their current approach creates a moderate to very high risk across downtime, cost, and safety. That plateau has a safety bill too.

The risk gap runs office to floor as well. So twenty four percent of technicians rate their risk high or very high versus fourteen percent of leaders. The people closest to the equipment are rating the danger highest. And why that makes sense?

Planned work happens on a de energized machine at a scheduled hour with parts stage.

Unplanned work happens wherever the failure left things.

Yeah.

So so what I'm hearing here is the schedule is slipping. The bill is really hefty when it comes in, which raises the question I think I would like to spend our next stretch here on, which is where is that time actually going? Second poll should be on your screen right now. In a typical eight hour shift, how many hours does a technician on your team actually spend hands on with the equipment?

It means diagnosing, repairing, inspecting, doing scheduled work, actual wrench time. Under one hour, one to three, three to five, or five plus. If you're a leader, answer for your crew. If you're a technician, you will know this exactly.

We'll give everybody a minute to respond, and we're seeing the answers come into the chat. It's always fascinating to see what our respondents had to say versus our audience who is here today. But let's go ahead and keep moving forward, you know, to understand what this looks like in the study.

Yeah. So here's what 686 peers told us, and it's actually one of the few places in the entire survey where the office and the floor agreed almost exactly.

60% of leaders and 62 of technicians say that hands on time is three hours or less out of an eight hour shift. So both the people who plan the work and the people who do it agree that the wrench is in someone's hand for less than half the day.

Where they stop agreeing is everything else. Where do the other five hours actually go? So, Amanda, in your view, what is eating those five hours?

Alyssa, I know you loved that 54% stat early in the presentation, but this is the one that caught my attention.

Leaders work from a work order queue, and technicians work from the shift. And that shift contains a lot of activity that no work order ever captures. A work order logs when the job started, when it closed. It doesn't necessarily capture the forty minutes that was spent hunting for a part before the repair could even begin.

The approvals gap is probably the cleanest example. Right? So 48% of technicians say they lose an hour or more per shift waiting on approvals or site access. Leaders put that at 37%, and that's a difference of about a third.

In Limble's benchmark report from earlier in the year, executives said everything was fine while managers said their data was terrible, and they needed help. It was a disconnect on every level. This study adds the next level down, managers, nurses, technicians. Each layer of the organization sees a calmer picture than the one below it.

And when work isn't getting done, it's usually because of friction versus effort.

Yes. Absolutely. So we actually ask technicians to name exactly what it is that's getting in the way and causing that friction, And here was the list of the top blockers. Approvals actually led at 51%. Equipment records that are incomplete or out of date came in at 41%. Then parts they can't locate, hard to use software, and work orders that arrive with enough detail, which came in for 40%, 35%, and 28% of our respondents respectively. Now, Amanda, correct me if I'm wrong, but when I looked at this list, it seemed to me that four out of these five are really just the same problem wearing sort of different The approval of the records, the parts location, the missing detail, every piece of that info exists somewhere in the operation, but then it doesn't reach the person standing at the machine without them stopping to go find it.

Absolutely. This is a process conversation before it is ever a conversation about technology. A parts example from the data makes it very vivid for me. So single site teams were 22% more likely to name unlocatable parts as a blocker to wrench time.

It's single site. So one building usually means inventory tracking never got set up because we all know where everything is right until the person who knows where everything is is out. Maybe they're out sick or with family or on a cruise.

And what makes the next number we're going to show you so uncomfortable, it's because most teams already own the things built exactly to fix this specific issue.

They do. And here is that uncomfortable number, which is that 82% of maintenance organizations already own and operate a CMMS, the system whose whole job is moving exactly all that information that we just talked about. 39% have inconsistent use across the whole team.

This is not a Limble stat. Our survey respondents run every CMMS on the market. Across all of it, the overall picture is bought, 82, actually used, 39. Amanda, what separates those teams that close the gap from the teams that don't?

It's a great question. Every stall out reason we saw in the data is a post purchase decision. People say they were never trained. They're desktop bound, slower than the workaround, framed as optional. Training is the first line item cut when an implementation is running long, and it's the single biggest reason systems sit unused. 47% of people not using the CMMS said that they were never trained.

But stale data creates a spiral that is hard to come out of. A tech pulls up an asset record that says repaired from eight months ago, gets nothing useful and stops checking. Technicians are one hundred and sixteen sixteen percent more likely than leaders to name stale data as a reason a CMMS does not get used.

A story from the field. Limble rebuilt its mobile app from the ground up after sending people out to watch technicians actually work. Small findings like screens unreadable in direct sun that you can only learn when you're standing at an asset. It's a small thing we did to really avoid small data, but also to ensure that our system wasn't slower than a workaround. It was in frame as optional. It was actually very useful.

Adoption is won or lost at the asset, not in the office. And this also reflects what you see in the field. Right? 35% say their system isn't usable in the field. So, again, back to that example with the sunlight. If closing a work order means walking all the way back to a terminal, the workaround is gonna win every time.

And last but not least, optionality kills the system. 22% say CMMS use is framed as optional, and a system half the team uses, produces records nobody trusts, which feeds the stale data spiral even more.

We wanted to provide one more example from the field. BenTech, now Next Power, a Liberal customer who found themselves in a situation where maintenance was managed through emails and manual processes. What was most convenient for the tech was those manual processes, those workarounds, but it led to limited visibility into work and safety issues, not to mention time consuming inspections. Right? So when the team implemented and really committed to using the CMMS, using Limble, they moved from manual to mobile, and they created real time reporting across maintenance and centralized visibility.

As a result, 30% time savings for the maintenance team and real time hazard reporting that protected their facility, their team, and production. Like Nextpower, more than half of those surveyed in this study were more likely to have real time visibility when that CMMS was used consistently. And almost three out of four reported, they were more likely to have a defined data to decision on as part of that process.

That makes a lot of sense. And then you knew that we were not going to get through our entire thirty minutes here without having the AI conversation. Right? So, of course, we weren't. We asked our respondents whether AI needs good complete maintenance data to work, and 92% agreed that it does. So we can basically accept that as common fact since we rarely see that kind of consensus. Now just look at how teams actually use AI today.

65 use it for work order documentation and summarization, 56% for PM scheduling, 49 parts forecasting, and failure prediction, the thing everybody really means when they say AI and maintenance, sits dead last at 30%.

So my read as a researcher looking at these numbers is maybe that this list seems like a little bit of a sequence. Right? Because the documentation that everybody's doing first is what builds the asset history that prediction eventually runs on, but you kind of have to record yourself to the top of that list. Right? You can't really buy yourself there. Do you agree, Amanda?

Oh, I I...

Yes. I have so many thoughts on this. The honest position really is fixing how work gets recorded before you buy an AI feature. Feature.

The model matters less than the data underneath it. A closeout note dictated at the asset isn't paperwork. It's training data for everything that comes after. So the AI barriers teams names, training 33%, where to start, 32%.

Are the CMMS adoption story all over again? Same lesson, new tool, which means the path to the top of that maintenance maturity ladder starts somewhere almost disappointingly ordinary.

The note a technician writes at the asset today while details are still fresh, and that happens to bring us to what we're leaving you with.

Yes. And this is really important, which is that everything you walk through today is maybe a third of what is in this full report. The knowledge loss numbers alone could be their own session.

And as of today, it's gonna be live. The 2026 data maintenance report is going to be sent out, following today's summit. It's free. It's built to be a real benchmark so that you can put your team's numbers next to these and know exactly which rung of the lander... Ladder that you're standing on right now and what you're gonna need to get to that next one.

And if that ladder left you wondering about your team's exact next move, that's what the next session is all about. Colin from Limble is going to be partnering with our friends from Aviva to talk about the maturity road map. Then the CMMS conversation continues with John. Your CMMS is talking, but is your team listening?

He's got some great practical hands on tips you can take away today to really make the most of that data. The people pressures, hiring, knowledge, walking out the door. Get a spotlight in a worksite fireside chat, and real customers walk through all of it in a panel. Stay in the chat.

We're gonna be there answering all day, and stick around for the maintenance hero awards, one of our favorite parts and favorite days of the year. Thank you so much to all of you who joined us today and to the 686 professionals who shared their experience of how this work actually gets done.

‍