Internet of Things (IoT), Artificial Intelligence (IA), there’s an App for that, … – we now have available an abundance of data, and even some information. It’s what we do about it that matters.
Although IoT is a buzzword now, it has been in the works for years – decades even. Dr. Jay Lee was one of the first to introduce me to the concept – if not the term. The reality is we do have a lot of access to data, and we have machines turning some of that data into information – but – so what?!?
If we as people don’t get involved and make decisions for, about and with that data, then we have succeed at nothing. I have seen companies working feverishly to capture the latest information on their machinery, only to ignore the actual data and let the machines run to failure. We need to step back regularly and look at the whole operation to determine what do we really need to know and why. Also, there needs to be a plan to act on what is learned. Too often we do not act on what we already know, waiting to see if there is more information around the corner. IoT will not change behavior. The process to act on data/information must be in place to utilize IoT successfully.
Automated vehicles are in the news, specifically for the failures of people to act on the data – and even information, they were given. Factory data rarely has fatal results if ignored, but the failure to act still has significant consequences. What is the point of knowing your equipment health, if the planning and scheduling system is not allowed to fix equipment before it goes into catastrophic failure? It is a common theme in after-action reviews to be able to pinpoint warnings, even multiple warnings that were ignored before the failure. I have seen leadership teams brainstorm how they can get better warning systems, rather than figure out how to act on the warnings they do have. It is always easier to push the responsibility down the road and wish for perfect information, than to accept the responsibility we have in utilizing the imperfect information already available.
I love data and information. I am an analyst at heart, but I fear that the growing IoT available will lead us to more catastrophic, and possibly even fatal events. I worry that folks in charge of making decisions will delay acting on the first sign of potential failure (P on our I-P-F curve), hoping to be ‘heroes’ by maximizing that P-F time and waiting for the really big warning from IoT to tell them time is up.
To avoid this propensity to put off making decisions on known failures, we need to reward managers who maximize the I-P portion of the curve and punish those who do not make decisions as soon as a problem is identified. That does not mean drop everything and fix problem equipment the second the defect is identified. But it does mean putting a mitigation strategy in place as soon as the defect is identified. Don’t wait for further indications of the down hill slope.
How can we make heroes of those that don’t delay? By measuring what doesn’t happen. How about measuring days since last production record? Promoting the equipment health score as metric for everyone to be proud of? Measuring time from defect identified to fixed? Bonuses for everyone in an organization that doesn’t have a catastrophic failure? There are always ways to game any metric, but focusing on positive metrics, rather than negative ones (downtime, production lost, et’c) puts the focus on performance, rather than non-performance.
Does anyone have an organization that rewards on avoidance of failures?
Before implementing any data gathering program
- Determine what you want know (production numbers, equipment health, quality statistics)
- Determine why you want to know that (product cost information, maintain equipment health, meet quality standards)
- Determine how to capture the data
- Determine how to analyze the data
- Formulas
- Frequency
- Who is accountable for the analysis – audit of the analysis
- Determine accountability for acting on the data
- Specific title or name – one person needs to be accountable
- Frequency of checking data and acting
- Parameters for acting (think of an over-damped system if the reaction is too severe)
- Determine how metrics will be published and used to drive team members to desired behavior regarding the What/Why you wanted to know
IoT is only as good as the management team that is operating it.
Engineer XX 