Placing your orders into your truck seems simple: there are 90 products of each 1m³, and the truck has 90m³, done! This looks beautiful in 3D graphs and it can even have the picture of your product on the object. However, what if these 90 products have to go to 30 different customers, and what if they’re not equal SKUs?
Suddenly it matters which products go in first and last. What if certain things are fragile and can’t bear the weight of other products on top or can’t be loaded on the same truck/pallet because it’s dangerous or will contaminate? These are only the basic questions resolved by load optimization. This article discusses critical concepts for consumer packaged goods: The Load Optimizer and Perfect Shipment.
Back in the 80s, the first algorithms were mainly used to optimize the packages for production, answering questions like how many products fit in which combination and into what case. This being sorted then and becoming a standard process in product creation made it slowly die down. This process created the next revolution: how to put these cases on the pallet. However, very quickly it was seen that this is part of the same process, and production pallets became the norm. Now how to put these pallets into the truck or container and especially if you have customers that are requesting different SKUs and different locations?
To illustrate this argument, let’s look at three different scenarios: 30 stops in a bay truck, 5 stops in a refrigerated truck and a ‘normal’ Full Truck Load.
30 stops in a bay truck
A bay truck has various bays, one for each pallet and a door for each bay. You can load them warehouse-focused, where every bay has a pallet of one product type and layers starting with the one with the largest number of a single SKU at the bottom and the smallest number of a single SKU at the top. You can load it route-focused, so you have one pallet per customer, and the pallet of the customer is on the side of the truck that is closest to the side of the road of the customer’s address. Or you can load hybrid in 2 ways: half per product group, half per customer, or everything per product group but with the SKUs for the first delivery on top in every single bay, so the driver doesn’t have to offload one tonne of products to get to the lowest SKU on the pallet for his first delivery and then loading that one tonne back in.
5 stops in a refrigerated truck
a refrigerated truck usually only opens at the back to keep the cold inside. This means it has to be loaded front to back. This also means that, in general, you want to have your first customer at the door and your last one next to the driver. However, what happens if you have heavy products that can’t be loaded on top of fragile products? Because, conceptually, starting at the driver, you will put heavy products first on the pallet going lighter and lighter until my lightest product, but then you’ll have to start customer 2 potentially on top of fragile products of customer 1. Should you start with your heaviest products again or with your lightest ones? Again do you prioritize your warehouse and ignore this restriction or do you prioritize your delivery and take the extra time to always have heavy products at the bottom of each pallet, meaning you’ll have to have more people working in your warehouse to separate the same SKU three or more times for the same customer.
Full truck load
And last but not least, the full truck load. This is an interesting one because everybody thinks that the full truck load is perfect, there is nothing to be improved, you’ve put it full of production pallets and can’t fit another pallet. Great, that’s the idea, but is it full in terms of weight and volume as well? Did you check the axle weights? Are you sure your pallets are in an optimal position? Are you adhering to your customers’ restrictions? Are you sure the order will actually fit? Enter: Perfect Shipment. Before you accept the order, you will double check all of the above. If you have less than 90% weight and/or Volume occupation with full footprint/max number of pallets in the truck, you need to occupy the space above the pallets. If you’re exceeding axle weight, you will need to move the load around. If your production pallet is 1.8m and your customer only accepts 1.4m, you’ll need to delaye. If he accepts pallets of 1.9m but not double stacking, you’ll need to add products to your production pallet to save more. If he orders 70% or 120% of your truck you’ll need to call him to either have him order more or to reduce his order, otherwise your loads will be very inefficient.
The future here is really in collaboration, data sharing and forecasting. Everyone in the supply chain seeing what is happening in that truck, sharing their sales data to enable forecasting of orders and allow delivery as VMI before you hit safety stock, receiving only exactly what you need at the times that you can process it. This would enable building the perfect pallet using production. For all three cases it would be very different, but also very similar potentially If your warehouse is as automated as your production. You could circulate pallets back to production to add only one layer of product and in a further future to add even individual SKU’s eliminating picking completely.
Load optimization has as many solutions as companies. It varies greatly with profile, costs and vision. The future calls integrated forecasting for all customers up to a point where the warehouse and customer ‘orders’/demands are a viable part of your production process. This is already the concept of S&OP, but the future will make this a reality. The real question is when will this level of technology become cheap enough to apply everywhere? For this to happen your company’s profile, your industry’s costs and your leaders’ vision have to be aligned.