In TPM, Isn’t it needlessly costly to replace parts that are working fine? What do you think?

Dear Gemba Coach,

I have a question about Total Productive Maintenance. My management has hired a TPM consultant who makes us systematically replace certain parts in our equipment even though they’re working fine. This seems needlessly costly. What do you think?

Ahh… I have to confess I’m not a TPM expert. My very first gemba workshop was a consultant-driven TPM workshop and it remains a vivid memory: picture a young theory of knowledge PhD. Doctoral student cleaning parts in the gut of a huge cold-stamping press. This was the start of my long love affair with the gemba, but, I have to say, not with TPM. I have come across several such programs since with varying degrees of results and so, again, I’m weary of general statements. In some plants I’ve come across, TPM has been hugely positive to get the guys to realize they have to take care of the plant and not just run machines (or people for that matter) into the ground and then replace them. Muri remains the first, and to my mind, most bitter waste.

Over the years, I’ve also witnessed how Toyota engineers helped their suppliers with machine uptime and machine upkeep. It turned out to be quite different, as you can expect, and, if you bear with me as I take a detour, I’ll try to describe the gap.

As a young social sciences researcher I was studying how Toyota taught TPS to one of its supplier. One Toyota engineer convinced me then that, in order to really understand what I was seeing, I needed to get my hands dirty and do some kaizen myself, and so I started helping out a shop floor consulting firm as an extra pair of hands to run their workshops – these were the crazy days of five-days full immersion events and moving machinery around on Wednesday night and hoping the plant would restart on Thursday because top brass was coming on Friday. Looking back, it was completely nuts (a big hello to those of you who still do that), but tremendous fun (for the consultants, that is).

Lean vs. Lean

This is when my research work became very, very interesting because I slowly realized that the “lean” (it wasn’t called lean yet) the consultants were doing and the TPS the Toyota engineers were practicing looked the same but were qualitatively different. As a sociologist, I had landed in one of these rather unique events: a paradigm shift. Paradigm is a big word to describe a set of thoughts and practices in a certain field at a certain time – for instance the Newtonian paradigm describes the world as clockwork driven by gravity, whereas the Quantum paradigm describes the world as, well, a weird incomprehensible place driven by quantum equations our monkey-brains can’t quite grasp. A paradigm shift is when the dominant worldview shifts from one to the other, and as worldviews tend to be “sticky” that doesn’t happen very often. I came to accept that I should not try to understand TPS through the eyes of other engineers, but seek out what was different, not similar.

Back to TPM. All TPM programs have the same goals: maximize equipment effectiveness through a system of preventative maintenance throughout the equipment’s lifetime. Basically, it’s about reducing down time, speed losses and defects. We all agree.

The way to do this in the traditional paradigm is through a maintenance schedule. You use the data you have on equipment breakdown to analyze the most sensitive parts and how long they last, and, much as with your car’s service, you plan the maintenance schedule of every piece of kit. In practice:

1. List all the equipment requiring maintenance work
2. Assign priorities (ha!)
3. Estimate costs
4. Create a schedule
5. Make sure all the parts needed are ordered on time
6. Implement the calendar and review failure data.

Sensible? Of course this is far from simple because equipment is rarely available when scheduled, maintenance is often organized by functional specialties (electricians won’t touch mechanics, etc.) and the maintenance guys are rarely free when scheduled and finally, well, breakdowns to occur and every one will be pulled in getting the machine back to work and there goes your plan.

I suspect that this is what the consultant you mentioned set up for your plant – a rigorous (and probably smart) schedule of maintenance and parts replacement that should, in theory, pay for itself in machine availability and OEE.

Autonomous Team Members

What I saw Toyota engineers teaching their suppliers was rather different. They weren’t looking to establish a schedule upfront (though that did happen on some things). They were trying to solve three specific problems:

1. How do we improve team member’s autonomy with the machines they handle?
2. How do we treat each equipment as an individual?
3. How do we free the maintenance department from as many things as we can so they can focus on rapid intervention and complex tasks?

To grasp how different this is, think of kanban. In the old paradigm, kanban is about replacing MRP with cards – essentially replacing one production scheduling system with another. People are obsessed with calculating the ideal number of kanbans, with seeing the full flow of kanbans in their VSMs and so on.

TPS engineers were trying to do something rather different: they wanted the operators on the cell to be autonomous as regards to their logistics. Team members had to be able to tell at any moment:

1. What they were supposed to work on now (the current production kanban)
2. What they would do next (the next kanban in the launcher)
3. Order the materials they needed for the next hour of production (the materials ordering kanban).

The system had to work to make team members autonomous as regards their logistics function. When you think about it, this is quite an endeavor: simplifying the logistics function to the point that team members can handle it without it breaking their cycle by distracting them from their work (think about what happens when one has to decipher a production instruction order). The spirit of kanban is cell autonomy, and the fact is that when kanbans fail, the cell STOPS, CALLS (through the andon) and WAITS (rather than start working on another part – overproduction).

The same spirit applies to maintenance: how can team members be more autonomous in handling their own equipment. Well, the same simplifying hard work needs to be done until regular maintenance can be done by the team members themselves without losing more than 5 minutes of production time per shift. On the gemba this appears as:

1. Clearly indicated checkpoints on the cell to show what needs to be maintained.
2. Easy maintenance activity with clear instructions (cleaning, oil lubrication, etc.- should be self-explanatory.
3. Constant training and coordination with maintenance technicians to further simplify maintenance.

With this in mind, we can now build a profile for the machine according to their age – I remember in Japan a Toyota supplier where all machines had a large sign stating their age, a few going back to before I was born – and their peculiarities. The maintenance parts ordering system can now work just-in-time with:

1. Current parts on hand in stock replenishment
2. Some specific parts with one part in stock and immediate replenishment when used
3. Parts that need to be ordered when needed, but with a clear identified supplier and a taxi system to fly them in quickly.

In this way, the plant learns to do what your consultant is trying to get you to do without carrying the extra cost that you have – rightly – noticed.

Machines Are Your Patients

In the end, maintenance will increasingly be about big data. Toyota veterans tell me that they keep historical information on all their important machines and they can compare performance across the group! Clever IT is definitely needed, but not so much on the scheduling front, but on understanding each machine as one would a patient and drawing the right conclusions for engineering to focus on: what do they need to strengthen in the process to avoid repetitive problems.

The deeper point here is about changing one’s perspective. The insight of lean is not to try to maximize the efficiency of the system at system level (no matter how consultants project this onto lean) but about developing the autonomy of team members and equipment by knowing them as individuals and training them one by one. In this, the lean ideal is indeed that of craftsmanship – the pilot in love with his 1930s biplane, the old car enthusiast – within the context of mass production efficiencies. We need to seek that space in all we do, including maintenance.

I recently had the opportunity to visit Toyota’s first historical plant out of Japan in Sao Paulo in Brazil, where they showed me a 1937 press still making parts for the Corolla, in mint condition. Now, that is TPM!