The Performance Paradox exists at the intersection of two diametrically opposed perspectives, both apparently rational. These are competing perspectives of business processes. The Performance Paradox suggests that operating problems can be solved locally without reference to the constraint of the production system. This is an illusion.

Local performance losses are the result of discarding excess production capacity. If production capacity can’t be used to produce profitable output, it must be discarded by one or another means. The operating problems we perceive as causes of lost performance turn out to be the valves required to allow excess capacity to escape the production system. These problems are not causes of lower performance. They reflect the collective means by which excess capacity leaves the production system. They are unavoidable. They are the effect of holding excess capacity.

The view of operating problems as causes of performance loss which can be repaired locally is associated with a view of business process as a collection of valuable operating assets, including human assets, each having its own independently realizable performance potential. This is the ‘continuous improvement’ fiction.

Another perspective of operating problems explains performance loss as the collective effect of needing to discard excess capacity. This perspective envisions business as a continuum or system where individual asset performances are interdependent and limited by process dependencies.

For example, here are the theoretical performance capacities and weekly throughputs at four different operating assets. We calculate the throughput potential at 100% realization of theoretical performance and at 75% of theoretical performance for comparison purposes:

These are the absolute limits for throughput at each of the four assets where each asset is effectively an independent production system. However, let’s now assume these four assets are interdependent steps in one sequential production system. Step 1 becomes the constraint of the system (no subsequent step can achieve a throughput greater than Step 1). Steps 2, 3, and 4 are non-constraints. They are limited by the constraint.

Let’s assume that OEE is 75% at the constraint (Step 1). Step 1 may process only 1,260 units per week at 75% performance. Each of the subsequent steps are therefore limited to process a maximum of 1,260 units per week, regardless of their independent theoretical capacities. So, all else held constant, OEE performance is strictly limited at each of the non-constraints. We know that OEE cannot be improved at non-constraints without either lifting the throughput of the constraint, or reducing operating hours at the non-constraints. Step 2 is limited to 62.5% OEE. Step 3 is limited to 53.57% OEE. Step 4 is limited to just 46.88% OEE:

The Performance Paradox kicks in when ‘local’ (functional managers) who own the assets at the non-constraints look independently at the OEE performances for their assets and determine that they are ridiculously low. Everyone knows that the opening gambit for serious OEE performances is to lift OEE to 75%. The ‘threshold’ for best practices in OEE is 85% in most industries. No serious manager is going to settle for an OEE performance of 46.88% without a fight (as in Step 4 of the above example).

Yet, without lifting throughput at Step 1 (the constraint) by raising OEE above 75%, or reducing the scheduled hours at their non-constraint, how can OEE be improved at Steps 2, 3, and 4? Put simply, it can’t. Lifting throughput at the constraint (Step 1) can be achieved by repairing a local performance issue. Lifting OEE at Steps 2, 3, and 4 is impossible without increasing throughput at the constraint or reducing scheduled hours at non-constraints, or some combination of both.

Let’s dig down a little further to examine the situation of the unfortunate managers responsible for performance at Step 4. Their OEE is 46.88%.

We will assume this is reflected in an Availability of 75%, Efficiency (Performance) of 85%, and a Quality of 76.5% (OEE = % Availability x % Efficiency x % Quality). Management decide to improve Quality by embarking on a well-resourced (expensive) Kaizen directed at reducing quality defects, first time through (FTT). They invest in creating standard operating procedures, operator training, and some capex designed to tighten process variability. Congratulations! Their efforts are rewarded when Quality performance increases to 85%…the threshold of best practice in their industry.

But wait! The % Availability at Step 4 has fallen to 77.9% and % Efficiency is just 70.8%. Factoring these two deteriorated outcomes with the newly improved Quality outcome, we see 77.9% x 70.8% x 85% = 46.88%.

Brutally, nothing has changed for the OEE outcome at Step 4. The Kaizen is locally successful (lifting Quality performance at Step 4), but it has failed in its enterprise objective to improve throughput across the entire system. OEE at Step 4 is directly dependent upon throughput and OEE performance at Step 1. The idea that this can be changed by acting locally to solve problems at Step 4 is mistaken.

There was never ANY chance of lifting OEE at Step 4, all else held constant. Management are drawn to an attempt to lift local performance by assuming local problems are the cause of local low performance. The real cause was an oversupply of capacity at the non-constraints in relation to a given throughput at the constraint. Cause and effect are confused.

Revisiting our data table, let’s see what happens to OEE at Step 4 when we lift throughput at Step 1 (the constraint). We will assume that OEE at Step 1 has increased to 85% through a range of efforts. OEE at Step 4 may now rise to 53.13%. That is an improvement, but we are still well south of the threshold for serious OEE outcomes at 75%:


What to do next? How do we rapidly get OEE to 75% at Step 4? Further improvement at the constraint may be expensive. Maybe, as we alluded to earlier, the real problem is a lack of profitable demand. In fact, even if OEE at the constraint was 100%, OEE at Step 4 would still hit a wall at just 62.5%:


So, something else must give. Let’s reduce the scheduled operating hours at the non-constraints:


We remove one 8-hour shift at Step 2, four 8-hour shifts at Step 3, and just over six 8-hour shifts at Step 4. Then it becomes apparent that 75% OEE performance is attainable at every step of the process and we can remove a lot of expenses at the same time. This has to be done in conjunction with allowing sufficient protective and recovery capacity…a subject for another time.

To clarify, reducing scheduled hours won’t necessarily result in immediately lifting OEE at non-constraints (even though it often does just that, almost magically). However, what can be said with certainty is that OEE can’t be improved at non-constraints without first lifting throughput at the constraint or reducing scheduled operating hours at non-constraints.

Introducing OEE to a production system can result in a type of collective insanity where people chase shadows at great expense and find only frustration and failure. It’s a culture killer where many people are ultimately defeated and deflated. Understanding process constraints is a prerequisite for embarking on the OEE voyage. One thing is certain. If you have been hitting a wall and OEE performance resists all attempts to climb above 75% for months or even years, you may be trapped inside the Performance Paradox. Look locally. Do you see cause or effect?


Productivity Step Change (PSC) is a global Management Consulting Group based in the US dedicated to maximizing Capacity Utilization, improving Return on Invested Capital (ROIC), and increasing productivity for its clients across industry. Our enterprise-level, data-led, cross-functional business analysis typically requires 4-6 calendar weeks and is designed to provide evidence of your opportunities for overall growth. Please contact us if you would like to schedule time for a discussion and presentation.

Pin It on Pinterest

Share This