The Spare Parts Puzzle: Insights with Richard Jeffers
Show notes
Welcome to the Trend Detection podcast, brought to you by Senseye Predictive Maintenance – which gives you visibility and insights into all your assets, from single machines to full plants to help you reduce downtime, increase knowledge sharing and accelerate digital transformation across your organization.
In this episode you will learn about:
The Hidden Cost of Poor Spare Parts Management: Discover how inadequately managed spare parts lead to significant capital tie-up, increased downtime, and often overlooked financial burdens in industrial settings.
Why Spare Parts Are an Underserved Area: Understand the historical reasons behind the neglect of spare parts, from complex naming conventions and lack of transparency to accounting treatments that obscure their true impact.
The Transformative Power of AI and LLMs: Learn how emerging technologies like Large Language Models (LLMs) can revolutionize spare parts management by accelerating RCM studies, standardizing inventory data, and enabling more strategic stocking decisions.
Integrating Spare Parts into Your Digital Strategy: Explore how effective spare parts management is a crucial, yet often underestimated, component of a comprehensive digital transformation and predictive maintenance strategy.
Gaining Executive Buy-in for Optimization: Get insights into how to justify investment in spare parts management by demonstrating clear return on investment through reduced working capital, minimized downtime, and enhanced operational efficiency.
You can find out more about how Senseye Predictive Maintenance can reduce unplanned downtime and contribute towards improved sustainability within your manufacturing plants, by visiting: www.siemens.com/senseye-predictive-maintenance
Transcript
So if you're starting your Greenfield factory and your journey towards digitalization, you wanna stop these legacy problems happening. um, and the, the way you do that is by really robustly applying, you know, techniques like early crip management, like um hazard operational reviews like er RCM right from the word go. Welcome to the trend detection podcast powered by Siemens, and now your host, Nia Sullivan. Hello and welcome to the latest episode of the Trend detection podcast. Um, I'm really pleased to be joined by one of our regulars. Um, we were just discussing before your 4th time now, Richard. Matt, I think so, yeah. Fantastic. Always nice to have you back. You are memorable, but it's, it's just because I've recalled so many, I forget. So just to clarify that. But, um, but no, today, um, it's actually Richard who's brought the, um, the topic to the table today, and I believe amongst all hundreds of episodes that we haven't really covered, it's probably been mentioned in a few podcasts, but whether we've covered it. In depth and it's all around spare parts and how it fits into your sort of digital transformation strategy, into your maintenance strategy, um, and maybe even predict a maintenance strategy. Why not throw that into there. But, um, but maybe Richard, I guess, because it's been a little while, maybe you could reintroduce yourself to the audience and then maybe a bit of background to why you brought this, um, topic to the table. Yeah, so, uh, good to see you again, Niall, um. Uh, so I, I'm an independent consultant in, er asset management, er, with a particular, um, slant on how to use emerging technologies to, er, improve the asset management, um, life cycle journey. Uh, I've been doing this as a, as an independent for about a year now, er, prior to that, er, and the previous conversations we had, I was within. The RS Group er running the um reliability and condition monitoring team, and, and prior to that, my career was in er operations and maintenance management, um, predominantly in the brewing sector, so I normally tell people I've built breweries, run breweries, and, er, maintained breweries. Yeah, sounds like the dream job to me, um, to be honest, but being around a brewery, but, um, but I guess, yeah, so on the, on spare parts, where, where did this sort of come into your mind? Is it something to do with conversations you're having a lot at the moment, Richard? Yeah, so, so I, I collaborate with a, with a, um, AI business called Ensemble AI who are focused in, uh, the use of AI within manufacturing. Uh, and we were looking across the asset life cycle and thinking about, you know, what are the, what are the unsolved problems that exist in the asset life cycle where, where AI and data, you know, has, has yet to, to deliver a transformational change. So, you know, we started off talking about time series data, um, and actually Sensei and Siemens is my default, er, example, er, of a good platform and someone who's nailed this, um, so, you know, let's not try and recreate that capability, but, but the use of unstructured data and, and transactional data is one that's, you know, really not been, er, grasped fully, um, yet in, in my view. Um, and you look across the asset life cycle. And there's a whole bunch of processes that are well understood, but actually are really difficult to do because they take a huge amount of bandwidth to do them. And they could be processes like RCM studies, um, or hazard and operational reviews, um, where we know the methodology, but it just takes a lot of human effort to do it. Um, and the one, you know, the specific one around spare parts is, you know, we, we all know that most factories have got, er, you know. Hundreds of thousands if not millions of pounds worth of engineering spares, um, but it, it's often badly, badly, badly referenced, so it's difficult to understand what, exactly what you've got, um, and that leads to an increase in working Capital, but also to, uh, increase in downtime because you think you don't have the spare part you need. So we decided to, To really drill into spare parts, um, as an area that's a known problem, definitely not solved, er, and, and one where we believe that er that AI can deliver a a transformational step change. And do you think, um, so, yeah, do you think spare parts, I don't want to, maybe use this term lightly, but do you think it's been neglected in some way? I think very much so, Niall. I think it's, um, you know, we, we did a bit of, uh, we did a bit of fermion maths, a bit of research, um. And we reckon there's something like 3/4 to 3/4 of a trillion dollars to $1 trillion worth of spare parts across the planet tied up in, in inventory, um, across factories, um, and other industrial assets, and, and one that. You know, stuff gets put on a shelf and then forgotten about and the, and the value of that spare part is almost degraded, er and and diluted. And there are a lot of the, the processes in place at the moment, I guess the fact you're talking about digitizing them suggests that a lot of it's very manually, you know, records are kept manually and tracking of spare parts is, is done very manually. I, I, I think it's not so much the records because the, you know, the spare part is going to be managed through the Enterprise Asset Management platform or the CMOS platform. You know, the, the, the challenge there is about item creation. So you deliver a large Capital project. The er the supplier hands you a spare parts list of all the recommended spares. Um, and, and we all, we all know that the right thing to do with that spare part list is to, um, do a, um, FICA study or an RCM study on the asset, identify the spare parts that, that should be stocked, um, to cross reference that with your existing spare parts to make sure you're not adding a duplicate, uh, and to look at your supply chain to see if you can transfer the risk of that spare part into your supply chain rather than holding it yourself. The the reality of what happens in, in almost every industry is that that list comes in through the project team late, er and there's, you know, a lot of pressure on time in order just to get the spares in ready to, to run the asset, and businesses end up er overinvesting in spare parts during the er the Capital phase, um, and then exacerbate that problem by not having a common naming convention, so that they bring in. You know, a bearing with the part number the OEM's given it, rather than normalizing everything back to a standard naming convention along with the rest of their spare parts. So over time, your factory ends up with er, you know, a variety of different naming conventions for spare parts, er er a, you know, not a risk-based er reliability driven approach to stocking decisions, er and, and then a mountain of spare parts that never get consumed. Yeah, it's interesting you mentioned about this like overinvestment in spare parts, but then often there's sort of a reluctance or a determination to, you know, and maybe not, it's not a bad thing, but, you know, to calculate return on investment or potential return on investment from the technology. Is there, there's an imbalance, it seems at the moment, maybe. Exactly, Nia, I think, um, you know, an investment in Data and, you know, data analysis as, as we both know, gets, you know, an enormous amount of, of focus uh in terms of ROI but I think with spare parts, because they go on the balance sheet, not er in, in not hitting the P&L in the first instance, um, you know, decisions around that er that investment in work Capital are made in a very different way from decisions that are, you know, about either a Capex or a revenue investment. Uh, and, you know, my experience across multiple sectors with a lot less rigor. Um, just to, just to put some stats into it, uh, you know, research would suggest that in an average, uh, industrial site, 50 to 60% of the spares on hand never move, um, and, uh, of that, maybe 10% is a genuine insurance spare. So that means that 50% of the stock held in an average industrial setting. was a poor investment when it was because it's not moving and it's not an insurance fair. Um, and if you multiply that over the average stock holding of, of, of a factory, you know, it, it rapidly becomes tens if not hundreds of millions of pounds. Yeah, especially, and I guess also the, um, the environmental impacts as well of, um, that, well, I mean, shipping. Shipping things out, not being able to track where, you know, what's, what's in stock and things like that as well, so you're having to rush parts over and that kind of thing. Uh, I mean, there's, there's, there's, there's tons of issues now, there's the, you know, there's the, there's the working Capital impact in the first place, there's the obsolescence risk, uh, because you've got something on the shelf and maybe it'll be obsolete before you need to use it. Uh, people tend to be poor at maintaining the maintainable assets within their, their engineering stores, so people aren't rotating motors to avoid flat spots or, or checking gearboxes are still, you know, in a good condition, etc. etc. So when you do need to use them, you suddenly find they're actually not fit for purpose anyway. So, um, you, you make this investment, um. You don't, you don't love them enough when they're in the stores, er, and, and then you end up having to make emergency distress purchases, er, you know, when you do need a spare part. So I guess they shift, shift the gear a little bit, but in terms of sort of digital transformation and technology and digitalization in general, so how, how is that now sort of impacting how spare parts are tracked and stored and and used, I guess, as well. So, so I think, you know, that they are, you know, in terms of how they're tracked, you know, as I said, most people will be using their, their enterprise asset management platform or their CMMS to to manage that engineering stock. As a transactional, um, you know, for the transactional element of managing it. What, what they're not doing though is using technologies in order to, er, you know, manage that stock at a strategic level, to understand what stock they should have, er why they're stocking things, er and to make er you know, you know, those tough choices about where to invest and where to, to not hold the stock, um, and, and that's where, er, you know. Well, you know, obviously the, the, the time series analysis, er, you know, through platforms like Senseii is clearly gonna tell you when you've got a pending problem and allow you to order a replacement component in with a lead time. um, but, but actually if you've already invested in the wrong stock, you know, it doesn't, it doesn't address it. And it doesn't help you to do the reliability studies that you need to do, um, to, to choose the stock you should hold on hand in the first place. So things like, you know, large language models do give you an opportunity to accelerate the, the RCM er analysis journey, er and also to analyze your existing stock on hand to really understand, you know, what it is, er and normalize the, the naming convention, so you can compare, er, you know, the same component from different um suppliers in a very transparent way. And I mean, when talking to, um, I guess people in the industry about predicting maintenance, is, is spare parts like a central part of the conversation, or is it something that's not really considered when you consider downtime and other sustainability factors or other factors? I, I, I think, you know, my experience now is when you, you know, you start talking to people about predictive maintenance and maintenance maturity, and, and, uh, you know, you'll talk about condition monitoring and stuff like that. And, and then later on in the conversation, you'll say, out of interest, do you have a problem with your, with, um, you know, codification of your engineering spares? Are you sitting on, you know, a large mountain of non-moving spares, at which point people go, oh my God, uh, you know, if only you knew how big my, my spare part mountain is. So it, it's, it's definitely something that people recognize as a big, as a big problem. But I think don't necessarily join it together with with part of an overarching maintenance strategy. Um, for me, management of engineering spares is, you know, is equally as important as your condition monitoring strategy or your, uh, work planning strategy, er, in order to ensure you've got, you know, an effective maintenance organization. And we also, well, very closely relate to this, of course, is data and data is probably, you know, if you go back through all of these podcasts I record, data is probably at the heart of most of the of the discussions, but I guess it's exactly the same for spare parts strategies, right? You need to utilize data. Again, at the recent mainstream event, we both attended, we're talking to someone who's collected 3 years' worth of data, right, from from there. For the machines, but is it a case about how to apply it and the context behind it, I guess. Yeah, exactly. I mean, that, that, you know, there's there's lots of different data associated with your, with your spare parts. Um, you know, the, the first and, and for me, most important is the, uh, is, is the naming data, you know, they, they're being really crystal clear. Uh, of, of what the component is and what the attributes of that component are, um, you then got the, the risk-based viability data about whether you should hold that stock in the first place, um, and you've got the transactional data that tells you how much you consume, uh, which you can use to set your, you know, your stocking strategy, your mint max, um, and reorder points. So there's a lot of different types of data that, that really. Uh, you know, could do with being managed for an engineering spares, uh, and, and the average CMS platform, will, will simply manage the transactional data associated with it, rather than the reliability data and the, and the, and the naming data. And, and, and why is that? I guess, just for my own knowledge, I mean, others out there probably know, know all this, but just for my own information, you know, it, it is hard because, um, you know, 11 supplier will call a component a different thing from another supplier because, you know, people don't want transparency of their supply chain, because if you've built a bearing into your machine. You don't necessarily want people knowing exactly which OEM bearing it is because you want that after sales revenue to go through you as the machine manufacturer, and you want to maintain some mystique about what that bearing is. You know, the reality is the bearings probably come from one of the mainstream bearing manufacturers, and actually as a consumer, as a customer, you'd rather get it directly from the, from the bearing REM rather than through the machine manufacturer. Um, so part of that is, is deliberate, um, obfuscation of the, of the, of the provenance of the, of the bearing. Um, the, the next thing is it's not the most sexy and exciting thing to spend your time in, in a, in a maintenance environment. You know, if you're out there on the cutting edge of condition monitoring and, you know, looking at your trends of data and going, I can predict this is gonna fail in 3 months' time. You know, it's, it's much more exciting than, uh, trawling through master data around engineering spares. Um, and, and, and I'm, and I am absolutely convinced that the accounting treatment of, uh, of engineering spares of having it on the balance sheet, you know, does push it down the, uh, you know, the radar for people because it's not, you know, if you, if you're a budget holder and you put a, put $1,000,000 worth of stock on, on, on the balance sheet. That doesn't hit your budget, you don't see that because it's seen an increase in working Capital. And, and the size of the working Capital in spares is normally substantially smaller than the working Capital tied up in raw materials and finished goods, which means that, you know, even the finance community are not necessarily banging the drum about it because it's a fraction of the cost of your rules and er you know, your raw materials and finished products. Yeah, so it goes under, under the radar a little bit. Yeah, despite being, despite being a big number, um, so, uh, you know, back in my old Heineken days we would maybe have, you know, 100 million pounds of finished goods and raw materials, uh, and, you know, 8 million pounds worth of engineering spares, so it's only, it's only 8% of the, of the working Capital, but it is 8 million pounds. Um, uh, and if that was hitting your budget as a, as a maintenance manager, I'm sure you'd pay a lot more attention to it. Yeah, I guess, and to your point, you know, talk about, you know, maybe sexier, you know, things to worry about than, um, no, people can't see the quotes, I realized on a podcast, I was quoting. But, um, yeah, so. He talked about, you know, more exciting technologies like condition monitoring, you know, predictive maintenance, but is there, is there something Yeah, how can we get people excited about spare parts? Maybe that's the question. Yeah, I think, you know, a lot of it is just, you know, talking about it and talking about the real problems, because, um, you know, people are used to talking about spares in terms of the procurement journey and the relationship with their supply chain. I, I think they're less confident talking about the spare parts they've got on hand and talking about how to manage that better, so, um, you know, for me, just. You know, raising the topic and talking about, you know, technology can be a solution to this problem. In a way that, you know, it's not been possible to do it in the past, you know, in the past it's required, you know, individual engineers to trawl through masses of data to do that, er, to do that, you know, cross referencing of spares, but actually with the advent of technology, you can use that to, to do the heavy lifting quickly, and then the engineers can bring in, you know, more interesting professional judgment into the should we stock it or not, because they're not tied up with just the fairly tedious cross referencing element of it. So I think, you know, calling out the opportunity technology gives to, to do the fun stuff with spares of, you know, my my reliability study says I should carry these spare parts er and get technology to do the heavy lifting about, um, you know, the cross referencing and naming nomenclature challenge. Yeah. No, exactly. Exactly. I think, yeah, I think everything we've said so far, it's it's sort of an underserved, um, you know, a part of the maintenance strategy. I mean, I mean, broadening it out a little bit. Are there other areas as well? I know we're talking about spare parts, but there are other areas that sort of fall under the radar. Yeah, I think knowledge management is, is definitely another underserved area where the emergence of LLMs give, gives fantastic opportunity to do things differently. Uh, you know, I was talking to, well, I think back to my old life, uh, when I was running, you know, breweries, you know, managing the maintenance of breweries, none of those were in their first flush of youth. Uh, you know, we had assets, you know, up to 30, 40, 50 years old. Um, and, you know, you know, imagine a project team comes in and delivers a new asset, um, and in their, in their folder structure is everything you want to know about those assets on the day that they are commissioned. The project that the the the asset gets commissioned and that project team disperses, and suddenly that knowledge, even if it's still on a, on a, on a shared, shared drive somewhere, suddenly becomes incredibly difficult to access. …