In today’s technologically advanced world, instant gratification is more often than not expected and assumed. You have a question? Google has the answer at your fingertips. You need to tell your mother you will be late for dinner? Send her a quick text message and she will know to keep your food warm for you. A new episode of your favorite show has come out? Stream it from your device of choice right away. And the list goes on.
While arguments can be made for the downside to the emphasis on immediacy and information constantly being thrown at us from all different directions, there is a place where the upside is huge and the downside is miniscule. That place is your factory.
Whether you focus on it or not, your factory relies heavily on data. The number of people working per shift. The hours of a day that the machines are running. The units of electricity per minute used by each machine. The list is endless and ranges from the broad to the tiniest detail. Within this wealth of data lies the keys to your factory’s success.
Most likely, you have a system in place for the collection and analysis of data. You have probably identified certain metrics that you track on a daily, weekly or monthly basis. That data is then analyzed and funneled into reporting templates that show the overall factory results. You may even have customizable reports that can be used to answer specific questions that come up.
But let’s think for a minute about how this data is gathered and analyzed and what problems may arise, specifically problems around timing, accuracy and bias.
A long-term analysis looks at trends and requires data collected over time. There is no question that long-term analyses are important and can provide valuable insight into the productivity of a factory and can help determine whether changes need to be made.
But, a factory floor is an extremely dynamic environment, making short-term analysis equally, if not more, valuable. Data from 5 minutes ago is already old data and possibly irrelevant, to say nothing of day-old data. Depending on how long it takes to collect the information, it could potentially be days or weeks before you realize that a particular machine is not producing as it should – by the time the problem is detected, a significant impact can already have been made on the bottom line.
Imagine if you could know instantly that a machine was going offline unnecessarily for a minute at a time…think of all the wasted minutes that could be saved immediately.
How is data currently collected at your factory? If you are relying on an employee to manually input data into a system, you must assume that there will be inaccuracies. Even the most careful employee will occasionally make a mistake. It could be a momentary distraction, a slip of a finger on the wrong key, or just a careless error, but inaccurate data can lead to poor decision making.
If significant time passes between the collection of the data and the inputting of the information, the risk of inaccuracy grows. The more people involved in the process also increases the chances that a mistake will be made. Checking and double-checking may reduce the risk of errors, but also adds on time to a process that could otherwise be more streamlined.
Imagine not having to wonder whether one wrongly-inputted number could be throwing off all of your results…think of all the time and energy you could save by not second-guessing and double-checking data.
Each person has a unique perspective on a situation and will bring that bias, consciously or subconsciously, to whatever they produce. For example, if there is a certain type of data that gets collected at the end of each day, and 3 times a week Employee A is responsible for this and the rest of the time it is Employee B’s job. A slight difference in the way each of them interprets a particular piece of information can mean skewed results.
In a more severe example, an employee could in theory alter data to try to hide a mistake or make performance look better than it actually was.
Imagine there is no need for human intervention in the collection and reporting of data…think of the human resource power you can divert elsewhere and the peace of mind in knowing that the numbers are just numbers with no personal interpretation.
All of the problems described above could be easily solved with real-time data. Imagine a world in which you could instantly see what is happening with each machine, at whatever level of detail you require. Imagine how quickly you could respond to situations and how quickly you could watch your profit grow.