What is lead time and what does it tell you about? How do you calculate it and how can you use it? – Take a look at the answers to the most common questions people experience while learning about lead time with easy, clear and catchy examples.
1. What’s the difference between work item types and classes of service?
Imagine you own a café. You offer your clients espresso, cappuccino, frappe, and tea. These are your work item types – each product you offer is going through a different process of preparation.
You also have your clients with their different requests. Your regular morning starts with you serving coffee to your Standard clients. They come, choose their seat, and start reading their newspaper while waiting for coffee. Then, suddenly, you see a woman that runs into your café in a rush to order for takeaway. If she waits for too long, she will run away and you will lose the order. So you Expedite her order while your Standard clients are waiting.
When the woman happily leaves, you raise your eyes to a clock. There is a table booked for 11 a.m.; your clients want to have a meeting in your café. This is your Fixed Date request. Thinking about an upcoming order, your eyes meet the blue sticker on your planner saying «Explore the Chai Latte», – a new type of drink, that one of your clients suggested you have. It becomes very popular and it seems like a good idea to have it to attract new customers. But making it happen doesn`t affect your current processes, so this is your Intangible class of service.
2. How can you calculate your lead time?
Lead time is the time you need to deliver the service to the customer.
However, a customer might not always be the external one, who is actually paying money, but also it may be your colleague or another department who ordered the work to be done.
To know your lead time, you have to understand the time the request is in the process; i.e. how long does it take for the work item (ticket) to move from the commitment point to delivery on your Kanban board. To have an understanding of your data, you need to know the lead time of at least 30 work items (tickets). Knowing your lead time helps you to understand your processes and make proper commitments for your customers.
Imagine this, knowing your lead time you can easily say «Hey, Jeff, I give you a 90% guarantee that we will deliver this work in 14 days or less, you`ll see!». And you will deliver on time because you know that 90% of the work items go through your Kanban board in 14 days or less. Now you can be trustworthy and reliable in the eyes of your customers and clients.
3. Thin-tailed or fat-tailed?
Remember a simple mantra: “the problem is always in the tail”. Why so? A lead time probability function is a curve that shows the distribution of actual data for the lead time from a Kanban system. The x-axis shows the number of days of lead time and the y-axis – the occurrence of that time within the sample data set (the number of tickets pulled through the Kanban board). Therefore, the longer the tail of your curve is, the bigger your lead time and the less predictable/reliable your process.
“Fat-tailed” distribution has a long visible tail which means a poorly predictable and risky process where planning is difficult. Meanwhile, the “thin-tailed” curve has a short tail and reflects the process that is reliable, predictable, has shorter delays with lower impact.
How many data points are enough to know it? 11 data points as a minimum are enough to understand whether it is thin or fat-tailed. With 30 data points, you will have a robust understanding and a fairly accurate model. Knowing your lead time “tail” in your Kanban system is very important and makes a vital difference to planning, risk management, and the likelihood of being viewed as a trustworthy service provider.
4. How to understand what is “early” or “late” when you know your lead time?
First of all, what is your lead time? Imagine you have a thin-tailed lead time of 30 days.
The calendar shows the 1st of October and the delivery is planned for the 1st of December.
- This means you are currently in a «super early» period. This is 2×100% of your lead time before the desired delivery date (DDD) or more.
- It will change to just «early» next week because it is still more than 100% of the lead time.
- What you need is «normal» – a period between 100%-85% of your lead time, which will be the first days of November in case of our example.
- If we start «late«, 80-50% of the lead time, we may still be able to finalize the work and realize the profit, but it will already be affected by the impact of delay and we may need to change the class of service to higher.
- The last moment to react is a «last responsible moment» (LRM), an instant at 50% of the lead time before DDD which is the 15th of November for our example. If we start that day, we still have a change to finalize work and realize the profit, but when this moment is gone we are not able to finish on time.
- Now you`re in «irresponsibly late» period.
5. Why doesn’t forecasting make sense when you have fat-tailed distribution?
As you already know, a fat-tailed distribution means the process is not predictable and unreliable. For example, the fat-tailed lead time may be min. 1 day and max. 77 days (Flow.hamburg GbR real case study data). This means that in the best possible scenario the item may be delivered today; in the worst – during the next 77 days.
The analysis of the data showed that about 50% of all items were being taken care of in 6 days or less, but if you were unlucky and got into the other 50%, you might`ve ended up with the delivery in more than 2 months. That is why with the fat-tailed lead time we have a very fragile variability, so it is harder to predict reasonably from it.
Learn more about Kanban studies in Kanban Maturity Model book or get access to full book content online using kmm.plus. Attend training at the David J Anderson School of Management to learn more about advanced Kanban studies and how they can help your business, or find your local trainer at Kanban University to start your Kanban journey.
Recent Comments