Ten Steps To Improve The Efficiency of a System
Understanding a system well enough that you can improve its efficiency is challenging, rewarding, and can actually be a lot of fun. Over ten years and hundreds of projects, I have come to the conclusion that most of the methodologies and tools used to improve the efficiency of systems all function on the similar base concepts.
The goal of this article is to provide a free distillation of those concepts into 10 actionable steps. These steps are focused on increasing a systems ability to produce X per unit of energy or time.
Lets get started!
Step 1 - Measure how the whole system is performing currently.
Measure the entire system. You will usually have 1 total system metric for efficiency. This is typically X created per (time/energy unit). An example here could be through the lens of running a company that makes pizzas. Lets call this company “Pizza House”. In this case your total system metric for efficiency could be “Total Location Pizza Produced Per Labor Hour”. For each Y (unit of time) we get X (pizza). This metric represent all the pizza and all the labor hours used to create that pizza over a span of time.
Step 2 - Measure performance of subcomponents
Now you need to figure out the 3 -10(ish) subcomponents that will dictate the performance of the whole system. These are considered leading indicators of performance. If the sub component stations are doing well, we can gather that the lagging indicator (total pizzas produced per labor hour) will be in a healthy place.
For Pizza House, here are a few we could use;
Volume of cheese prepped per labor hour (by weight, per cheese type)
Sauce prepped per labor hour (by volume, per sauce)
Ingredients chopped per labor hour (by weight, per ingredient)
Pizzas produced by the chef per labor hour (this is where all the ingredients come together)
Pizza boxes folded per labor hour
Pizzas delivered per labor hour.
Error rate
*And the list could go on.
If these metrics are all in a good spot, so will your total system productivity.
An important note on data collection: You will change some aspects of what you measure over time. So don’t get too caught up in the weeds. Just do your best and start measuring now. Even if decide you were measuring the wrong thing after a month that is okay. Congratulations you are using data to make a decision! You just learned what not to measure, and are on the right track. Only way to lose here is to not measure.
Step 3 - Organize the data.
In this step you collect all the data you have been capturing, and organize it into a dataset. Doing so will help you zoom out and see the bigger picture. Usually some type of data table can be helpful here.
Step 4 - Understand the data you just organized.
Here you take the recently organized data, zoom out and see what the heck is going on. The more data, is typically better. But lets be real here. Unless you are dealing with a system that has lives in the balance, 30+ data points is probably enough to get cooking. The objective when improving a system is to make it better, not perfect. Clarity is gained through action and analysis.
When analyzing the data, your goal is to understand what performance looks like on a great day, a poor day, and an average day. Also, you will want to understand the variables that caused that type of performance.
Jumping back to the Pizza House example. Maybe a good day is 3 pizzas per labor hour, a poor day is 1 pizza per labor hour and, and a typical day is 2 pizzas per labor hour.
Step 5 - Figure out where effort will provide disproportionate returns.
Alright, this is one of the key steps in understanding how to improve a system.
It is critical to understand which subcomponent drives the majority of the system performance. If you have a sub component where all the raw goods come together to make the product. This is that step.
At Pizza House the key subcomponent would be where the chef builds the pizzas. And the associated metric would be “Pizzas produced by the chef per labor hour”.
Step 6 - Support The Key Subcomponent
Once you have identified the key subcomponent. Every other subcomponent in the system should align its work to support efforts at the key subcomponent.
That means proving the key subcomponent what it needs, when it needs it, how it needs it. Everyone in the system serves and supports the key subcomponent. This means any burden the system encounters must be put anywhere except the key subcomponent. This sounds obvious, but is missed a surprising amount of the time. Breaching this rule is typically done with the best of intentions, but it actually works against efficiency.
This is because the key subcomponent dictates the output of the system. It is impossible to have a whole system efficiency that is greater than your key subcomponent efficiency.
The reason for this is because all other subcomponents essentially drain from that key subcomponent from an efficiency metrics standpoint. If you have a Chef at Pizza House, that can make 10 pizzas an hour, their pizzas per labor hour is 10. But if you have 4 other kitchen staff supporting the Chef. Your total system efficiency is 2 pizza per labor hour
Heres the math:
Total pizzas/total staff hours = whole system yield
10 total pizzas / 5 staff hours = 2 pizzas per labor hour.
Step 7 - Find the Bottleneck Hampering Performance of The Key Subcomponent
This is where the action happens, and honestly never really stops. Once you have found the key subcomponent (chef), you need to understand where THE bottleneck is.
The bottleneck is the subcomponent that is choking performance of the system because it is starving the key subcomponent. If the Pizza Chef is waiting on cheese from the cheese prepping station, there’s your bottleneck.
Step 8 - Improve The Key Subcomponent.
To improve the throughput of the system, you will need to improve throughput of the key subcomponent. To do this you will need to ask yourself this important question.
“Is the key subcomponent waiting on raw goods?”
If the answer is “Yes”, then effort needs to be applied to the previously mentioned cheesy bottleneck. That is because the key subcomponent has capacity to spare, it just can not utilize that capacity. So to improve the whole system efficiency the bottleneck needs additional capacity.
To gain capacity, you apply process improvements and equipment/technology improvements. Additionally, you should look at the other support tasks and see if any of those tasks can better support the bottleneck. In the Pizza House example, if the cheese prep station is the bottleneck, but sauce prep has extra capacity, then the sauce prep-er should help the cheese prep station so that the chef is never waiting on materials.
With that covered, now we circle back to the question above.
“Is the key subcomponent waiting on raw goods?”
If the answer is “No”, then that means all the support systems have capacity and the bottleneck is actually the key subcomponent. If that is the case, that is where you provide process improvements and equipment upgrades. Once the bottleneck has been improved, it can make sense to expand and hire a second chef. This will allow Pizza House to take advantage of its improved efficiency and increase its total output.
Step 9 - Track Progress of Your Improvement Efforts, Adjust as Necessary.
As you make system improvements, you should see total system throughput increase. If you do not see improvements, that typically means you added capacity somewhere that it wasn’t needed. For Pizza House this would mean we should see “Total pizzas produced per labor hour” increase.
If the total system metric increases, great! That is what you were looking for. If it didn’t, well you learned where you have more capacity in your system and you can go back to the drawing board. Regardless of the outcome, the actions are close to the same. Keep finding the highest leverage tasks and processess that will give you your target result.
Another important note. As time continues, its complete common to work harder for each incremental improvement. In the Pizza House example, the first series of improvement projects may improve efficiencies say 40%, but after a handful of successful projects, that number will draw closer to say 5%. And after years of pushing performance, Pizza House may be targeting .5% improvements.
Step 10 - Celebrate the Wins
Make sure to celebrate! It is easy to expect the wins and just keep moving forward without looking back. That is a mistake. In life, wins are not guaranteed so we need to celebrate when they happen. Additionally, reflecting on what went well, and what could have gone better will help future improvements run a bit smoother. Improving systems can be a whole team effort and that needs to be acknowledged. Even through we are talking about systems, those systems are typically administered by people, and we can never forget that!
So there you go, a decade of projects boiled down into 10 steps. I hope they serve you well!