A Simple Yet Scientific Approach to Problem Identification

In various continuous improvement methodologies, we talk a lot about problem solving. How do we know if there is a problem or an opportunity? My coaches and mentors from Toyota define a problem as a measurable gap in performance. Determining the size and impact of that gap is a critical part of problem identification.

Somebody might come to you with something that sounds like a problem. They might say, "Our patients are unhappy because they're getting the wrong meals too often." A statement like that would be called a "big, vague concern" by my mentors. You have to go and investigate, looking at data and observing the actual problem; you may find that the person reporting the “big, vague concern” might simply be overreacting to one patient who got really upset.

To formally define a problem, we'd want to know the following:

• Target condition ("what should be happening?")
• Current condition ("what is actually happening?")
• The gap between those two
• The impact of that gap

For example, our target condition for patients getting the right meal would probably be 100%. Data might show that one patient out of 1,000 got the wrong meal yesterday, and there were only four complaints in the last month - this is our current condition. To define our problem, we’d identify the gap between those two, and determine the impact of that gap.

"What do we know and how do we know it?"

To truly define the problem, we need actual data or other evidence of the problem, not a guess or an assumption. We should ask, "What do we know and how do we know it?" For example, four complaints doesn't mean there were only four incorrect meals (some people don't want to complain, but the wrong food can cause health issues for some patients).

But, either way, we have a problem. It's confirmed and it's measurable:

• Target = 100% correct meals to patients

• Current = 99.2% correct meals to patients

• Gap = 0.8%

• Impact = patients get upset, the wrong food could jeopardize their health

Defining problems in this way helps us ensure there actually IS a problem, before we run to conduct root cause analysis or start brainstorming possible solutions to test.

Sometimes, what looks like a problem isn't a problem at all

There might be times where the big, vague concern turns out to not be a problem after all. Let's look at an example from manufacturing. Sales might complain, "Our customers are mad about our slow response in getting their orders shipped out."

Rather than jumping straight into action, it's helpful to confirm if there actually is a problem. Looking at data, we might see:

• Target = 95% of orders shipped within two days

• Current = 97% of orders shipped within two days

• Gap = None

Using that definition, there is no problem because the plant is exceeding the performance standard.

Opportunities for improvement

Of course, a single upset customer might define the problem differently! They wanted their order on time and they didn't get it. The plant can decide if they want to work toward 100% on-time shipment.

The idea of raising the performance standard or target, say setting a new goal of 99% on time delivery, would represent what my Toyota mentors would call an "opportunity for improvement."

Register for our free upcoming webinar: