Is this possible with the existing toolkit?
This requires more precise planning than ever before and easy and flexible modification of plans.
Step zero is to be able to tell you at any moment how my production is going, what’s happening on the shop floor:
- What materials can I design with?
- What is the finished product stock flow?
- What is the stage of completion of my work in progress (WIP)?
- What resources do I currently have? How many machines do I have running? Which ones are not running, why not? What maintenance operations are scheduled? What unscheduled maintenance operations are currently taking place?
- How many operators will I be able to start the next shift with? This becomes a particularly exciting question as the good weather approaches, during the school holidays…
If I can confidently answer most (or more) of the above questions, and those answers are backed up by accurate data, then we are one of the lucky few manufacturing companies that already have a descriptive analytical data layer.
Microsoft, a leading manufacturer of data analytics tools, has divided analytics maturity into four levels:
- Level: Descriptive analytics. Typically seeks to answer the question “What happened?“. It gives me almost instant information on the state of my production. How are the facts compared to what was planned? Under what quality parameters have I achieved the desired number of units? How often does a particular breakdown occur on a particular production machine?
With almost instant answers to these questions, not only is decision-making time significantly reduced, but production plans are also based on real-time data. This saves planners considerable time and effort in trying to match plans to the facts.
-
Level: Diagnostic chanlanalysis. From the previous question, “What happened?” we come to the next question, “Why did this happen?”
Further analysis of the descriptive elements brings us to the roots. By recognising these, we can also focus on their prevention.
Here, not only a sudden machine failure can be evaluated on a data basis, but also, for example, the differences between planned and actual data can be evaluated, or a sudden increase in the reject rate can be answered.
-
Level: Predictive analytics. Using and analysing various historical data, we can predict future events. Here we can predict future events based on hindsight and past experience. “What will happen?“
What may first come to mind are predictive maintenance tools: machine failures can be predicted on the basis of objective data before they occur, causing an unplanned shutdown and the huge losses and production downtime that would result.
In every factory there is an invaluable veteran “specialist” who can tell from the vibration and sound of the machine that it is bearing-ridden and in need of servicing. However, this requires years or even decades of experience with the machine – which in the long run can be replaced by analysis of the data collected.
A more common example is when, based on historical data, we can predict that there will not be enough resources on the Friday afternoon shift, and therefore we will not be able to meet the weekly plan.
-
Level: Prescriptive analytics. “What to do?“
Prescriptive analytics helps to determine the immediate, direct impact of improving a process.
It also helps to optimise production plans or even improve logistics processes and improve stock levels.
A typical use case for this is which operator should work on which machine, on which order and when? How much overtime is required to meet the schedules in a given week?
In prescriptive analytics, machine learning and artificial intelligence tools support process optimisation by efficiently evaluating up to millions of variables.
The implementation of data-driven decision processes, including BI solutions, is still a real challenge for many companies. According to a 2019 US survey, only 46 percent of companies had a clearly defined and communicated data and analytics strategy to rely on (2021 IDG Data & Analytics Study ).
Making the best use of data not only enables real-time informed decisions, but also brings real business value and therefore competitive advantage to any company.
Send us your contact details and we will contact you as soon as possible, or contact us at
+36 1 204-7730!
