In this post, Mercurius IT’s Operations Director, Lee Cridland, speaks to NETRONIC about his experiences of production and project task scheduling at Siemens and job scheduling at Mercurius IT, as well as provides insights into what good scheduling looks like.
Lee Cridland: My involvement in scheduling began when I was working at a company called Alstom, which is now Siemens. The site I worked at was primarily one where gas turbines were designed, manufactured and supported over their lifetime. Gas turbines are hugely complex and expensive machines, with contract values typically ranging between £2 million and £50 million. Gas turbines contracts often consist of hundreds of thousands of components, with some being bought out finished (BOF), some being assembled as kits and others being manufactured from raw materials. Many of these components were on long lead times, maybe 12 months or more. This meant that scheduling was massively important because if you got your scheduling wrong you couldn’t meet customer requirements on time in full (OTIF), and might not be able to take the order or might incur higher costs and penalties; meaning lower overall profit.
Having a schedule allows you to forecast demand and perform critical path analysis, which helps reduce the cost and risk of projects. It organizes your resources based on a work breakdown structure to work out the most efficient way of completing a project. The critical path shows dependencies and prioritizes important tasks in the project to create a sequence which determines the overall duration. It is also very valuable when you need to factor in interruptions, both planned and unexpected. Planned interruptions are things such as maintaining machine tools, operators going on holiday or turning the power off to get the supply tested and serviced annually—they’re the things you’re aware of. Unexpected interruptions are things such as a machine tool breaking down and taking a couple of days to fix. You need to know how you are going to catch up and how it is going to affect the schedule.
You need to have a method and a process to keep on top of all the contributing factors and that’s what scheduling is all about.
Scheduling can be a very complicated, sophisticated and mentally challenging role; however, using a computer to visualize your schedule makes it much more straightforward. It gives you the ability to drag and drop elements into the schedule, combining them and giving a clear picture of everything in front of you. It shows the critical path, helps distribute tasks efficiently, keeps costs under control and prepares you for unexpected interruptions.
A very important function that schedulers can use is a what if scenario. You ask a question, such as “If I were to change the delivery dates of these two contracts around, what impact would that have on my overall demand and capacity?”. The system then predicts what effect it will have on your schedule. It’s a very useful tool for optimizing and testing adjustments in your schedule to see if they’re worthwhile, rather than implementing them with no preparation. Overall, scheduling is essential for manufacturing organizations and most production scheduling is bottom-up, for all but the simplest products, using a finite capacity forward scheduling approach.
Lee Cridland: At a business like Mercurius IT, where our resources are people, we do infinite capacity backwards scheduling, which is a top-down approach. It’s much easier for us to alter our capacity as we can work weekends, bank holidays and overtime. We can also hire third-party resources with the skills we need and subcontract. We take a more rudimentary approach to scheduling as it's easier for us to change our capacity, meaning we can afford to be less accurate with our predictions and we don’t have to be as concerned about lead time. We schedule using a sort of milestone level, using a rule of thumb approach which is based on templates. There are different templates for particular project types. For instance, if a project is going to take 12 weeks, we know the duration of tasks within that 12 week period. If a different project is going to take 12 months, but it’s the same type of project, we just stretch the template out over that duration—the tasks stay the same but they have a longer duration in which to be completed.
Every morning we look at an Excel spreadsheet with all our resources, current customers and projects. We have 4 slots per day for 2 weeks and drag and drop tasks on for each consultant. We don’t worry very much about the size of the task, we make a judgement. Currently, our granularity is quarter days, so we look at tasks and estimate how many quarter days we think they’ll take. If you go back even a year ago, we would have allocated full time resources by project and managed their work time based on throughput versus where we needed to be. We wouldn’t try and level the load for each resource, which is what we’re working towards now.
As we get more sophisticated, we’re starting to better estimate the duration of tasks in our projects. We want to move to a finite capacity forwards scheduling approach eventually. Although, the problem with this is that you turn the entire process on its head. Using infinite capacity backwards scheduling is a top-down approach, whereas finite capacity forwards scheduling is bottom-up. You must choose one or the other, you can’t blend the two together very easily. It’s a big change and a very skilled job that takes time, but after you’ve done it once, it becomes a lot easier.
Lee Cridland: The answer here is really that it depends. If you have spare capacity then you don’t really care. For instance, if you have an experienced resource that will take two hours to complete a task but they’re busy for the next 12 hours and you have a less skilled resource that would take 10 hours but they’re completely free, you’d give it to the less experienced resource because you have that spare capacity. If they’re fixed cost then you’re paying them whether they are working or not, so you don’t really care which resource is quicker. What you want is to share the workload until you have your capacity used up effectively. However, if your capacity is filled and you need to subcontract, you need to make a judgement as there may be extra cost, training, admin and risk involved. You need to weigh up the pros and cons to see if it’s worth getting the task completed quicker through subcontracting or waiting until your resources are free.
Lee Cridland: Using Mercurius IT as an example, where we use a top-down approach, the main things we would need to be aware of to begin creating a quality schedule are our capacity and demand.
In order to understand your capacity, you first need to be aware of how many resources you have, then you’ll need to know what skills they have and how efficiently they complete work. To comprehend how efficient your resources are, you need to set a standard task which should take half a day, then get all your resources to complete the task and judge their efficiency by the time it takes to complete. It’s very important to understand your resources throughput as everyone is going to be different and will have varying work rates—knowing the efficiency helps you schedule more accurately.
You then need to set up a working calendar in which you specify whether you work weekends, bank holidays or overtime. You need to be aware if you allow paid overtime or not, as some businesses will pay a premium and others won’t. Once you have an understanding of overtime, absence and holiday allowances, you’ll have a picture of your forecast.
Once you understand your capacity, you need to get a picture of your demand. Firstly, you should look at how many customers and contracts (or sales orders) you have, then look at how much work is involved in each sales order. A work breakdown structure needs to be created by examining how the work is broken down into separate tasks and how they sequence together. You’ll need to know the dependencies because some tasks will have to be completed before others. Once you’ve established your work breakdown structure, you need to see what skills are required for each task and estimate how much time they’ll take, this will give you your demand.
Once you understand your capacity and demand for each sales order, you need to give them scheduled delivery dates. Put all this information into your schedule and work backwards from the delivery dates to create a work plan. Breaking down tasks day by day will help ensure that all scheduled delivery dates are met.
Lee Cridland: For me, the most difficult thing is estimating the effort required for each task. We already have templated work breakdown structures. We’re aware of how many resources we have, we’ve got a good understanding of what skills they have and we have a pretty good idea of how efficient they are. What you’ve got to do for each of those tasks is estimate how much work is involved and what skills are required to complete it. That is what I’d say is the hardest thing. There are so many varying factors contributing to the tasks that it is hard to be accurate when estimating the amount of work there is.
Lee Cridland: If you’ve not done any scheduling before, I’d suggest keeping it simple and using a top-down approach. That way you’ll learn where your strengths and weaknesses are. Incrementally, you can take a more granular view of your capacity, or time buckets, as you gain experience and learn more about what you’re scheduling. At Mercurius we used to use a day and then we went down to half days. Now we use quarter days and our scheduling is more accurate as a result of becoming more granular.
My other advice would be to keep it visual. Even if you schedule on a computer, have a screen or whiteboard that all your staff can see. Having it visual and shared means everyone knows what they’ve got to do and it’s more likely that if somethings going wrong it’ll trigger someone else to react to it. It might also sound a bit negative but having a schedule that is visible to everyone can motivate people to get the job done because everyone else will know they’ve failed—you don’t want to be the one person that’s late.
In 2010, Lee Cridland joined Mercurius IT as Operations Director, having climbed the ranks
He holds an Honours Degree in Engineering & Manufacturing, is a Chartered Engineer and has a Master’s Degree in Business Administration.