What Data Tells You - or Doesn't (Part 2)

Last week, we talked about the importance of data to an organization. Every organization, but especially manufacturing ones. You simply cannot operate without knowing how you’re doing against target. Without that knowledge, you are just surviving. And it probably feels like that on some days.

Data analytics is a new-er field in that they can use computers to track it all these days, but back in the 50s and 60s, plenty of people brought baseball statistics sheets out to ball parks and filled them out religiously. There are plenty of examples out there but I really liked the cleanliness of this one. A stats sheet gives you a look at everything that went on in the game, minus capturing people thumping garbage can lids to communicate, but I digress. We’re not talking about cheaters today.

I’m focused on how your manufacturing team might be losing valuable information – data that can help you directly make more money – because you don’t have a way to access it, collect it and keep it. And we’re here to help that. Looking back at the baseball stats sheet, you can see where you know how each team member is doing by at-bat and when they’re on the field. Can you say that about your team? Do you know how they’re performing on standard work and things that are longer-term as well? You can compare player to player (team member to member) and even batters to pitchers (department to department).

HOW TO FIND THE KEY IN DATA

Using technology, it is easy to look both backwards at results and forwards on current trend lines to see the projected direction (unless you implement a change). Digital data gives you both information and the ability to make better decisions. It should provide the information you need when you need it so you can make the right choices for you.

For data analytics fans, we can dive in a bit and see the different types to follow. We just talked about predictive analytics and, often, I would think that sports statisticians are looking for predictive measures of how their players may or may not perform based on their current run rate. It’s cited as the most common form, and I can see that across the board.

Prescriptive analytics dives in a bit deeper and asks and answers questions like “if I made this change, what is likely to happen?” Articles often use AI and Big Data to reference how to get there. In manufacturing, this can be a very expensive process. MES companies installing systems over the course of many months or years and charging 6 or 7 figures for systems believe they have the tools for this. I ask myself at what cost? Is there a way to get some of that same information for a price that doesn’t directly impact a large portion (or all) of your production profit?

Diagnostic analytics examines the past and drills down into what happened. You may here the term “data mining” here. People using this technique may use alerts to signal that something that has happened in the past is about to affect you again.

Descriptive data analytics is often associated with the reporting you see coming out of analysis. These may be preset and reused or something you design on the fly. I often caution people that descriptive analytics can suffer from confirmation bias. If you allow yourself to build analytic reports “on the fly”, you may end up finding what you thought you were looking for. This does not mean that it is causal – just that you can find the data you think you need to support your hypothesis. The entire reason we have a data analytics field is to keep data analysis neutral, so be aware as you start to ask questions.

The key for you in data is going to be different than almost everyone else out there. I know that doesn’t help to hear, but it’s true. Your inventory issues might be supply-chain driven; for someone else, it may be more location-based or poor planning. You cannot point to an issue easily in manufacturing and know why it’s occurring. That’s what makes manufacturing so rewarding and maddening at the same time.

THE BASIC DATA YOU NEED IN MANUFACTURING

Going back to that baseball statistics sheet, where they have simplified the stats down to the basics, we can learn something about the basic data that you need in manufacturing to really make a difference. We are relying on 26 years of experience helping manufacturers like yourself to answer this question, rather than further research. Why? We often here the underlying issue behind the why from people we talk to. We hear the connections and the frustrations. We see the gaps in the processes and can point backwards to where it starts, and it is never the same place twice.

First and foremost in our mind is the accuracy of your production schedule. We’ll talk more about that next week as we dive further into our series on data.