Escaping Continuous Improvement
Most continuous improvement programs look like circling the origin
For most business leaders expecting tangible results, continuous improvement (CI) initiatives look like circling the origin, starting all over again, fixing issues that were thought to be fixed, and so-called improvements that don’t appear in the financial aggregates. This scene contrasts the idealized image of continuous improvement as a consistent incline going higher and better.
The main reason behind this phenomenon is that CI practitioners lack the analytical tools, analytical definitions, and frameworks that integrate into other functions of the organization. Without an extensible analytical framework, there is always way more gray area than black & white, and it is hard to make clear improvements and institutionalize improvements.
For the sake of simplicity, I directly dive into a framework that works like magic. This is a novel piece that you haven’t heard of, and it is explained in 5 logical steps.
CI improves efficiency. Efficiency contributes to productivity.
Efficiency is the ratio of reference inputs to actual inputs.
This means efficiency is the variance around the as-planned and as-budgeted consumption of inputs such as labor, material, tool, energy, machine time, etc. Therefore, manufacturing engineering and operations engineering are key functions for a healthy CI.
The key here is twofold: First, there must be a specific traceable measurement unit on which the plans and budgets are made. In discrete manufacturing, the simplest version of this unit is the operation step in the routings. Second, this unit of planning and budgeting can have multiple reference inputs such as labor, material, tool life, electricity, gas, machine time, etc.
This calls for a view as follows:
2. Efficiency has its internal and external trade-offs.
Most operations have multiple inputs, and the efficiency of these inputs usually impose trade-offs. For example, using higher-grade materials can reduce material efficiency due to the increased unit cost of these high-grade materials, but it can improve machine efficiency by preventing small stops, and it can improve labor efficiency by reducing the need for machine attendance. Another internal trade-off example is that running a machine faster might increase machine efficiency, but it can reduce energy efficiency due to exponential power draw and reduce material efficiency due to more defects. These are examples of internal efficiency trade-offs.
On the other hand, external trade-offs are driven by factors outside the operation. For example, set-up times are not only driven by the set-up SOP of the operation but also driven by how optimal the job sequence is. Other examples of external drivers are upstream delays, upstream defects, and downstream clogging.
So, using planning and budgeting standards with a ground-up activity-based approach is the best thing you can do to enable analytical conduct and persistent and measurable outcomes that don’t get lost in these trade-offs.
This calls for a view as follows:
© Saip Eren Yilmaz, 2023
3. External and internal efficiency trade-offs impose different efficiency curves for different objective functions
Remember the first picture in this article. Productivity is an aggregate outcome, which is more than the linear sum of local efficiencies. The gap between the linear sum of local efficiencies and the aggregate productivity is one of the big gaps where CI practitioners are lost. Between the local efficiencies and aggregate productivity outcome, there lie internal and external trade-offs, which can only be analyzed with deductive solutions such as optimization and simulation models.
Unfortunately, the classic continuous improvement toolbox does not have tools for dynamically optimizing external trade-offs. Yes, there are heijunka and changeover wheels, but I disqualify them due to their static and analog nature. Classical continuous improvement also lacks the tools for dynamically optimizing the internal trade-offs, which must be optimized as a function of the broader operational context. Because the internal efficiency of a process can be optimized in many directions, such as the minimum total cost, max speed, min defect, or any combination of relevant efficiency parameters, all of which are context dependent and imposed top-down.
As of 2023, it is quite possible to simultaneously optimize both the process parameters and the operations in synergy.
This perspective provides a very clean analytical perspective that clearly exposes external and internal levers and breaks process efficiency into sub-levers.
Process improvement lever
Input reduction, variation reduction, input substitution, cost reduction, and activity elimination/reduction
Process optimization lever
Optimize for X given the specific optimality conditions of the operational context.
Asset uptime lever
Make assets available more % of the time at a lower cost of uptime
Another great thing about this approach is that it always provides the process memory and also gives CI practitioners a warm start with clear strategies to improve processes.
© Saip Eren Yilmaz, 2023
4. External trade-offs must be optimized, and they impose the optimality conditions on individual processes
Now, this is where we come close to the aggregate productivity phenomenon, which must be measured as an exhaustive and ultimate metric such as OTE or working capital profitability. So, no one can sweep the inefficiency under the carpet.
Also, holistic measures are great tools for automated value discovery because these measures enable automated sensitivity analysis and what-if analysis on aggregate productivity and always provide a fresh list of the next best CI initiatives. But more importantly, aggregates, when decomposed into mutually exclusive and collectively exhaustive (MECE) subcomponents, you can do any fancy benchmarking at scale with ease and double click to any detail input efficiency parameter. The magic of benchmarking is that it boosts value discovery, raises the bar in visible ways, and disseminates best practices at scale.
The below example shows a case of how OTE can be used for this approach. Using working capital profitability integrated down into individual efficiency and optimization levers is a way greater magic, but it is too good to give away for free in a blog post.
© Saip Eren Yilmaz, 2023
5. OEE is myopic, and OEE is NOT about efficiency
Yes, OEE is not about efficiency. It has nothing to do with efficiency. Thus, it cannot be parsed to aggregate productivity.
OEE stands for overall equipment effectiveness. As the name suggests, it measures your effectiveness of getting the most out of your equipment.
Yes, it is myopic because it looks at one piece of equipment at a time, and it measures waste, ignoring why good performance is good when it is.
Most experts are indoctrinated with OEE in very misleading ways. This is a totally different topic, and I do not want to add insult to injury. But one thing I am very allergic to is when OEE is supposedly rolled up to factory level, which is insanely wrong. If you want to have a factory-level metric, OTE is the easiest.
OTE has its own deficiencies, but it is a quantum leap from OEE.
Take a look at the below pictures.
© Saip Eren Yilmaz, 2023
© Saip Eren Yilmaz, 2023
© Saip Eren Yilmaz, 2023
© Saip Eren Yilmaz, 2023
