We all love a great video game, right? I know I do and one of my earlier experiences with video games was Tetris. For those of you born after the millennium, below is a photo of what Tetris used to look like on a Nintendo Gameboy.
Source: YouTube
It started out very rudimentary and it provided hours and hours of entertainment, provided you had sufficient batteries, because the bl##dy things didn’t last very long. But I’m getting sidetracked here.
Playing the game in the late eighties, I never thought it would come back to help me with what we do today.
A lot of our projects involve some form of sizing. Whether it is sizing of a warehouse, areas within a warehouse, automated solutions, office space, yard space. You name it, we size it. We have various in-house developed tools and methods that help us do so with a high degree of accuracy. In the end though, what we are constantly doing is playing Tetris. Granted, it’s a slightly advanced version of Tetris but it is still Tetris.
Every customer is different, each has a different product range, different requirements, different stock on hand levels, different processes, etc. However, if we take it back to the core principles, we’re still just trying to come up with the optimal way of fitting someone’s operation into the best possible footprint with the most optimal layout.
A while back we published an Fuzzy Friday about different types of pallet racking called “can’t see the forest through the uprights” and it explained the 7 most common types of pallet racking available in the Australian market.
But how do you work out how many of each type you need? So today we take it a step further and explain what tools we use to determine the optimal layout and the use of these pallet racking types.
There are a few basic requirements before we can do anything. We will need to have:
- Product dimensions
- Pallet dimensions
- Pallet heights
- Stock on hand levels per product
- Pick volume
- Order profile
Unfortunately, the first item on the list is usually the hardest to come by. For some reason, almost every single customer we deal with does not have accurate product dimensions. So, what usually happens is that our data wizard, Yohan, does his magic and scrapes and calculates whatever he can from the data that is available. Whether it is from getting product dimensions from product descriptions, applying average values across a product category if one or two products in that category do have accurate dimensions to applying various other methods to fill in the blanks.
Once we do have all the information available, we run a data analysis. What we want to achieve is the following:
- What type of pick location is most suited for a product using:
- Quantity picked
- Number of location visits
- Product dimensions
- How much reserve storage space does a product require using:
- Stock on hand levels per product minus what is already in the pick location
- Pallet dimensions
- Pallet height
- What is the overall size of the warehouse going to be based on the previous 2 analysis:
- What is the optimal location within the warehouse of:
- The pick location for each product
- The reserve stock for each product in relation to the pick location
Now we have developed a few tools to help us with this and to speed up the process. The analysis is usually done using our standard PowerBI model followed by a sizing exercise using our standard Excel warehouse calculation, a sample of which is below.
The above sample is based on single selective racking. So, if you would apply the logic from the previous article on the above calculations, you’d end up with a storage area of ~12,420 square metres in 2018 up to ~22,490 square metres in 2031.
Great, job done!
Well, not really. It’s great that we now know how much space we need to store everything but how do we determine what the best location is for a product? So, we’ve only covered the first 3 items on our list, we haven’t determined where each product “has to live”.
We now enter the realm of slotting.
This is where Tetris really becomes 3D. Imagine you have 8,000 different products, like we have in our sample. How are you going to ensure these are all assigned to the optimal pick locations taking into account:
- Walking/driving distances during picking
- Location size and types available
- Congestion in aisles
- Vicinity of reserve stock
And that is without having to worry about seasonality.
Slotting is a very difficult task, which is why most companies rely on specialised software systems to support this task. We have our own Excel based tools that allow us to run a simple slotting exercise for a customer, but it doesn’t consider vicinity of reserve stock for instance or congestion in the aisles.
For that we use Slot3D. a dedicated software tool designed to optimise slotting in a warehouse. It requires the same inputs we need for our sizing and calculation exercise, but it has the right algorithms to slot an entire warehouse within minutes.
Even better, it can continuously re-slot that same warehouse based on historical transactions and forecasted sales. So, seasonality is no longer an issue as well.
Now, how does Slot3D work?
It is based on AutoCAD, believe it or not. At first, we thought they were a bit wacky but the more you think about it, the more it makes sense. The biggest waste in any warehouse is walking/driving which is linked to distances. So, if you can 3D model your warehouse and know exactly where each product is located, both the pick location and the reserve stock location, you can run scenarios to reduce walking and driving, thus reducing cost. A simple 3D model is shown below.
Now of course this article is too short to explain exactly how Slot3D works but I did want to give you one sample of what we do with it, which is reducing replenishments.
Imagine if a product is slotted to a carton live storage location, that fits 1 layer of a full pallet of 6 layers and it sells at 2 layers per week. From a replenishment point of view, you’re replenishing that location twice a week. Worse still the replenishment is labour intensive because 2 out of 3 times, the operator must bring the reserve pallet down from the reserve location, fill the carton live storage location with 1 layer of stock and then place the reserve pallet back again in the reserve location.
If, however, this would be assigned to a full pallet pick location on the ground floor, you would be replenishing the location once every week and a half and the replenishment effort would be as simple as dropping the reserve pallet from its reserve location into its pick location.
The sample below is based on re-slotting several products to a more optimal location to reduce the replenishment effort. The good thing about Slot3D is that it actually calculates the original effort required in hours versus the proposed including the associated labour savings.
This is only one of the many slotting improvements possible with Slot3D. If you would like to know more watch this space as we will be having a Slot3D roadshow in September 2019 with live demonstrations around the country.
About the author: Bas Schilders is the principal consultant at Fuzzy LogX who are the leading warehouse, logistics, and process improvement consultants in Australia. With a career spanning 2 continents and 20 years in warehouse improvements and solution implementations, Bas is one of the most experienced consultants in the country when it comes to logistics improvements and implementing warehouse automation solutions.