If you could see the future, what would you do? Well, first off you would probably buy a bunch of winning lottery tickets—but you might also attempt to optimize your day to a certain extent. Instead of being taken off guard and having to scramble to make arrangements when you get an unexpected call from school that your kid is sick, for instance, you’re already on the road, having made arrangements to work from home for the day so you can tend to him or her. On the way home, you know that your child’s going to want their favorite comfort food, so you’ve already called in a pizza order.
Luckily, you know in advance that there aren’t going to be any major work emergencies, so you can be pretty lackadaisical with your timeline—any work that piles up you can just make up tomorrow when your kid is feeling better. I’m betting this is a lifestyle you could get used to. Maybe you’re even wishing that your production chain could work this way as well. Well, the good news is: it can (sort of). Obviously nothing can truly predict the future, but modern advanced analytics come awfully close. As a result, they can have a huge impact on production scheduling and the entire value chain for organizations that adopt them.
Predictive vs. Prescriptive Analytics
Here you might be wondering exactly what we mean by advanced analytics. Advanced analytics flows typically fall into two categories: predictive analytics and prescriptive analytics. The former is reminiscent of the example above, i.e. it takes in data from your past and current customer orders, market conditions, supply chain information, production data, inventory levels, etc. in order to spit out predictions. These can range from demand forecasts to predicted effects for proposed changes to factory floor arrangements, shipping routes, and any other variable that impacts your value chain. Because these forecasts are informed by massive amounts of data—more than a human planner could realistically grapple with—they tend to be more accurate and more sophisticated than old-school pen and paper predictions.
Prescriptive analytics, by contrast, are used to pinpoint potential areas of improvement at various touchpoints on your existing value chain. These can be used to power digital twins, i.e. digital representations of your production facilities on which you can test proposed changes. These are, in a sense, also predictive, insofar as they offer a forecast of which types of changes will be most effective at maximizing throughput, capacity, efficiency, ROI, etc. in the long term. Taken together, these two types of advanced analytics represent a new, smarter path forward for production planning. How? We’re glad you asked…
Scheduling Around Constraints
To take an extreme example, let’s say your production flows revolve around a job shop in which no fixed machine order exists. Scheduling efficient production plans in this environment is hard enough as it is, but when you add in the need to optimize production ratios for different products or product customizations while maintaining the flexibility to slot in new orders as needed, you’re facing a task that can feel difficult or impossible, due largely in part to the complex relationships between a large number of different constraints. Because these types of production environments can be so complicated, reaching an optimal schedule by hand is somewhere between a rarity and a virtual impossibility. For an analytics workflows with access to live production and order data, on the other hand, reaching a schedule that’s at once relatively optimal and relatively flexible is all in a day’s work. Where a human planner might see a mass of difficult to decipher data, your advanced analytics solution homes in on fruitful interconnections between different value chain elements.
The result here is that your visibility is boosted and your production schedules are better aligned with the reality of your demand levels and existing orders—even in the face of extreme complexity. Not only that, but analytics integration has a cascade effect throughout the rest of the value chain. To wit, with analytically-derived demand forecasts, you can schedule production further in advance with a higher degree of certainty. This means that your sourcing can be done under more optimal conditions (especially if your sourcing workflows are also integrated with your advanced analytics in order to help you estimate likely price fluctuations for raw materials). Disruptions and unexpected events will still undoubtedly arise, but they’ll do so with less frequency, potentially improving your throughput and therefore your bottom line.
Okay, so we’ve seen the ways in which analytics processes can turn a complex set of constraints into dynamic, efficient production schedules. If you’re working with products that require a lot of customization (like modern automobiles, for instance), there is real value to be gained here. Because you can both predict likely order ratios and adjust your schedules on the fly with a high degree of certainty as to the results, you can customize more cars per day and decrease order lead times in the process. In this way, you can essentially boost your capacity from its existing levels by cutting out areas of downtime that would have inevitably resulted from attempts to schedule these customization flows by hand.
At the end of the day, this increase in efficiency and agility (powered by the decrease in complexity and uncertainty) is where a lot of the value of advanced analytics lies for production planners. We’re not saying you can quite see the future, but you can certainly better align your expectations with likely realities across various touchpoints on the value chain. The fewer surprises there are when it comes to parts availability, demand levels, order constraints, fuel pricing, raw goods pricing, and machine downtime, the better off you’ll be. Of course, in order to make these improvements possible, you’ll need to prioritize the kind of data that these analytics flows thrive on. That means systematically breaking down silos and boosting connectivity throughout your entire operation. The more sources of accessible, high quality data you can create, the more effectively you can let analytics power your production scheduling.