Why OEE, MES and Machine Monitoring Don't Matter for Many Manufacturers

This is a question I am often asked: “why are manufacturers so slow to adopt technology to improve their operations?”

I’ve heard the same two or 3 answers frequently parroted by experts the last 5 and even 10 years. The common answers suggest manufacturers aren’t tech savvy enough, or they don’t want to spend the money, or they need ‘expert assistance’ by consultants to help them. But these are not the reasons I hear when I ask many of the supposedly slow technology adopting companies that question directly. They give me a different answer.

I am in the process of helping a client finish deploying a machine monitoring system and the experience has helped me see clearly why MES and other Machine Monitoring solutions

fail to get adopted, or fail to get or sustain meaningful traction after they are installed.

Back to the question, but this time, more narrowly focused on manufacturers of high mix, low volume, or custom manufacturers: "why don't companies use machine monitoring systems?".

In my estimation, this is a large segment in the Canadian manufacturing industry, and it's not that they're slow or hesitant adopters. It's because machine efficiency is not their primary concern or problem.

Their primary concern is production efficiency, and machine efficiency plays only a very, very small part in that.

Case in point: I have a client that has 20 or so machines, 12 of them CNC lathes. On any day, I see between only 4 and 8 of the lathes running production. This is not because they're 'slow', (they’re actually doing very well); it's because they assign operators to the most suitable or practical machines that day to produce the parts needed. In their case, ‘efficiency' or ‘productivity’ can't be calculated as a function of machine running time, because the running time for every part or work order is different.

Their productivity or efficiency metric is simple; it's the number of parts produced against the expected number of parts that could have and should have been produced.  (If they produced the same part on the same machines or lines everyday, OEE would be helpful. But most small manufacturers don’t fit that profile).

To achieve a meaningful productivity or efficiency metric, manufacturers (like my client) need to know what part is being produced on the machine or at the station, what the actual running time (per part) is, and what a reasonable load/unload and operator handling factor is.

The machine running time is a variable that should come directly from the machine, because once a set up is done, all variables such as spindle speed, feed rate and tooling have been adjusted and are in line with what the set-up person has deemed optimal. Using the average of the last five cycles allows for the target to automatically recalculate when something in the process has to be adjusted for whatever reason.  

For the load/unload times, or for operator handling and inspection, each part number can have a value assigned for it, (as an example, 45 seconds, or 60 seconds or 90 seconds, depending on size), or perhaps a factor for degree of difficulty (for example, 1.05 for easy parts, 1.10 for awkward parts, or 1.15 if an operator is running multiple machines). This should all be relatively easy to set up in the company’s current business systems.

The machine tracks the number of parts produced and the operator performs a test or inspection after unloading a part, accepting or rejecting parts, giving the client an additional automatic metrics indicative of part quality and process quality. In this case, it's “first time through”, which is their indication of set up and process quality. The total parts produced can be used to keep track of raw inventory usage, and the total good parts produced (which is total minus rejects) lets the customer keep track of finished goods inventory.

It's all pretty 'simple' from these manufacturing companies’ points of view, and it's very meaningful and very relevant to their operation. Up time, down time, OEE etc have little or no direct correlation to their productivity or efficiency. Their constraint is not machine uptime - its other things, such as operators available, orders in hand, and material availability.

From my experience (as a manufacturer and as a technology developer for manufacturing), I estimate that at least 1/3 of small manufacturing companies in Ontario (perhaps Canada) fall into the exact same condition. Their measurement or productivity and efficiency has little, if anything, to do with machine efficiency. In their minds, (and mine), machine capacity limitations are easily corrected by buying more machines or faster machines, and that's what most of them do. The other problems are a tougher nut for them to crack - first, having the orders in hand, and then the availability of operators and raw material, leveraging tools and systems which should easily and seamlessly do the tracking of in-process productivity variables. It shouldn't be overly difficult or complicated. From their view (and mine), it should be that simple.

"MES", "MOMS", "Industry 4.0", "IIoT" – those are all interesting and intriguing terms to these manufacturers; they’ve heard about them, saw the demos, went to tradeshows to look at them, and even investigated financial incentive programs to implement them, but many haven’t pulled the trigger and got started. Or perhaps they did, and hundreds of thousands of dollars and years later, they let the project stall.

Why?

Its because the ‘solutions’ don’t solve the problems they are experiencing. Few of them would say they have a machine capacity issue or machine efficiency issue – even though that’s what many outside experts imply their ‘real’ problems are. Productivity is their actual problem - getting the expected production out on time, the first time, every time. The way technology has been pitched to them the last 10 or so years has been a distraction.

I hope these thoughts and insights help. To share or hear any additional insights on this topic, reach out to paul@tpi-3.ca or leave your comments below.