Short cycle time is the main advantage in many organizations. If we have a short lead time, we give value to our customers.
Most customers are highly uncertain about the number of products they need from us. Their uncertainty will decrease as we advance in time. For example, our customer produces washing machines and buys motors from us. Suppose that we are in January. The customer can hardly suggest the demand in June.
The unknown demand is both the number and the type of motors. On the other hand, the customer knows exactly the number and types of motors that he needs tomorrow. So as we move to the future from tomorrow, the uncertainty will increase about the number and type of motors the customer needs. If our lead time was one day, the customer would be very happy since he knows exactly how much he needs and he doesn’t need to hold inventory. If our lead time was one month, his uncertainty would be about a month in advance, and so on. Increasing the lead time will increase the customer’s uncertainty.
Some organizations thrive on cycle time advantage. For example, Amazon
can deliver goods in hours to some cities in the USA. Amazon fights the competition by lowering the cycle time. Although Amazon is not always the cheapest place to buy, it is always the fastest. The sale numbers speak for themselves:
Amazon is the number one company with a huge gap over the second company on the list.
Another example: let us take an ice cream company. Ice cream companies will usually have a couple of weeks’ inventory, maybe a little more. Storing all that inventory costs a lot as it requires huge warehouses with subzero temperatures.
But if someone on the beach wants to buy an ice cream bar and the store is out of stock of this bar, the customer will either buy a different bar from this company or buy this or a different bar from a competitor and the company will lose the sale.
In order for the ice cream refrigerator to fill up:
1. Someone needs to notice that we are out of stock.
2. We need to have inventory at the local distributor warehouse. This is the most difficult part since it requires the whole supply chain to hold a lot of stock in frozen temperatures. During the summer the distributor will have a problem holding more than a couple of weeks’ inventory.
3. When an order is made to the distributor and the distributor has stock of the item he needs to collect the items to a container and then send it on the right track that might not deliver every day to the point of sale that made the order.
When an order is made to the distributor and the distributor has stock of the item, he needs to collect the items to a container and send the container on the right track, which may not deliver every day to the point of sale that made the order.
During the time between lack of inventory and replenishment of stock the ice cream bar is losing customers to the competition. What if we could reduce that time by 50%?
How to reduce cycle time
First, let us define cycle time:
Cycle time = the time the customer got the product – the time of the customer’s order
*Note: Of course we will only use customer orders where the customer wants the product as soon as possible.
The big picture
We will start by looking at the general cycle time of an organization/factory and then go down to categories and subcategories. To see something with meaning we need to group the items that will have a similar cycle time.
What does it look like?
Let us look at the Amazon* example in more detail:
We will want to see the cycle time for the whole of Amazon at first. What does it mean? It means that some of Amazon’s products are supplied by vendors in China and their delivery time is very different from those who are “fulfilled by Amazon”. So this cycle time is the general cycle time. We want to focus on reducing the cycle time for only products delivered from Amazon warehouses. We filtered all these products and get the overall average time to deliver that type of customer orders.
From here we will go deeper and see the time it takes to deliver from each warehouse to see if we are getting better or not. We can use warehouses like I did, or we can filter by product type. For example, if we deliver TV and refrigerators and they have different cycle time for some reason, then we can separate them by category. I chose to go deeper by logistic warehouses in this example. Our goal is to understand whether we are getting better or not.
In the following charts we can see that although we have 23-24 hours as the average time in all warehouses, each warehouse is performing differently. Those charts allow us to compare time periods and warehouses, but it does not allow us to know where to improve.
*Disclaimer – this example is not based on real numbers. All figures are fictional.
Cycle time breakdown
In the example above we saw the general time it takes Amazon to deliver and the average time for each warehouse. Those numbers do not help us to improve. We need to break the process down to smaller operations that we can improve.
The breakdown of the cycle time depends on data points that we have in the process and the resolution we want to handle. Each element that we use has an entry time and an exit time.
In the Amazon example:
we will filter only customer orders that should be fulfilled by Amazon and have inventory at hand. We will look at each customer order for all the data points and find the average for each element. We will build a database which will include the following timestamps:
- Customer order entry time
- The time when the pickup order started
- The actual time when the pickup started
- Start of packaging time
- Time of entry to loading dock queue
- Truck leaving warehouse time
- Time of the first package delivered from the truck
- Time of packages delivered to the customer
Amazon probably has many many more data points but let us focus on the eight above.
Respond time to customer order: point 2 minus point 1. If the inventory is available in the current warehouse, then this time will be the computer response time. If we need to deliver the inventory from other warehouses, then we will have waiting time.
Time in pick up queue: point 4 minus point 3. The actual time that the pickup person works on the customer order. This is usually the most expensive cost after the cost of shipment. Every improvement at this stage will have a big impact.
Pick up time: point 4 minus point 3. The actual time that the pickup person works on the customer order. This is usually the most expensive cost after the cost of shipment. Every improvement in this stage will have a big impact.
Waiting for delivery: point 5 minus point 4. This is the time when the packed and picked up item is waiting for shipping in a truck. If this is a popular destination, we might have a truck every 2 hours, hence the waiting time will be no more than 2 hours. If we have only one truck a day to this destination, the waiting time will be a very important part of the cycle time. When we track this waiting time, we can plan the process differently. For example, let us suppose that we have one truck leaving only once a day at 17:00 to a not popular destination. If we plan right, we will give priority in picking up and packaging to the packages that need to be on that truck from 15:00 to 16:50. After that, the packages will get a low priority since they will wait for the next truck tomorrow.
Loading time: point 6 minus point 5. The time it takes us to load the truck with the packages. Reducing this time allows us to use the trucks more efficiently. To improve this time we need to use the SMED method just like in any setup improvement.
Travel time: point 7 minus point 6. The time it takes for the delivery truck to reach the delivery area. This is a very important time-consuming period of time. We can improve it using optimizations.
Time to delivery: point 8 minus point 7. The time it took to deliver the package since when the truck entered the delivery area (we assume first delivery to be at the end of the delivery area).
Customer order cycle time: point 8 minus point 1. This is the full time it took from the customer making his order until delivery.
How to produce the data
Most organization databases save the dates of building new records or changing status in an existing record. We can usually find the data in dedicated log tables or as action timestamps in each record. If you don’t find the right date, you can always define one, but it’s usually much better to use an existing timestamp since you can use the history.
Collect the data by order number or order line. Here is an example of what it should look like:
After building the table we will build a stacked bar. That way we can see the full cycle time and how each element affects the overall time. The most important information that we get from this chart is the way the data change in time. We can figure out whether we are getting better or not.
The 2 most important points we get from the chart:
1. Where the time is getting “wasted”. the chart shows that most time is wasted at loading. Loading time? If we check why the loading time that should take minutes takes so long, we will see that a truck is not allowed to exit after 15:00 as it would arrive at the customer house too late, so it stays at the loading platform until the morning. Can we plan better? Can we let more “long” delivery routes leave in the morning and “short” deliveries later in the day?
2. We can see an improvement in delivery time. The time goes down, which means that we are doing something right (for example real-time optimization of the route when we know where there are traffic problems now).
Let us zoom in and filter the “time to delivery” section:
We can see the improvement and give feedback to the manager and the employees who made that change.
Before ending this post I want to show you numbers from a real project that I was part of. Our goal was to reduce the cycle time as fast as possible. The project started in mid-January 2019. The numbers in this chart are real:
The charts and data can be downloaded by clicking on the excel icon:
Summary
Cycle time is one of the advantages that an organization can have. To improve it we need to break it down to small segments that we can improve independently. Each segment improvement will shorten the overall cycle time. In this post, we dealt with the easiest problem – delivery from stock. But what if we also need to produce some of the orders? In the next posts, we will see how to break down the cycle time in organizations that produce the goods and organizations that deliver from stock and produce it if it is not in stock.
I will appreciate it very much if you add a like next to this post on the Facebook page and also if you add like to the Facebook page as well.
I will be glad if you press the like button, so I will know that you liked this post and I will write more.
Any remarks or comments feel free to send to:
ThePlanningMaster@gmail.com