As the quality engineer for the microwave top-level assembly line, you find yourself disappointed in the failure rate of the cover installation process. The failure rate seems stuck around 6%, when looking at a defects per unit standpoint, which is much higher than expected. 

You dive into the data, hoping that the defects per opportunity tell a different story. Is there also a 6% chance for the failure to occur, pointing towards poor process design, or is it a far smaller percentage chance, possibly pointing towards poor training or an uncalibrated tool? What is the DPO, or defects per opportunity? 

An overview: What is DPO?

Defects per opportunity, commonly known as DPO, helps give proper context to how often an error is occurring within a process.

DPO is a ratio calculation. Using the calculation below, DPO gives the ratio between the total number of defects divided by the total possible opportunities that a defect could occur. Since you cannot have more occurrences than opportunities, the result will be a number between 0 and 1.

Using a manufacturing example, a top-level assembler for a microwave oven includes 10 screws used to fasten the case to the body of the microwave. During a shift, 50 microwaves have their cases fastened to the body. At the end of the shift, it is found that three microwaves have quality errors. The first failed unit had eight screws that were installed with too little torque. The second and third each had two screws that were installed with too little torque. We take the 12 total errors, and divide the number by 500, which is the total number of screws used during the shift for cover installation.

We conclude that the DPO for this process, over the timeframe measured, is 0.024 DPO. Put in simple terms, there is a 2.4% chance that a screw will be under-torqued every time a screw is tightened. This does not mean 2.4% of products will have defects, since the operation occurs multiple times per product.

3 benefits of monitoring DPO

Understanding failure rate from an opportunity point of view

Unlike DPU, which gives you a better understanding of how many units to expect to leave the process with errors, DPO gives you an understanding of the true failure chance for a defect to occur. 

In the example above, the DPU, or defects per unit, is 0.06, or a 6% chance of a unit having a failure. But from an opportunity point of view, the example above only shows a 0.024, or 2.4% chance of a failure per opportunity. 

Because the operation occurs multiple times for a single unit, the two measurements tell two different stories.

DPO is the starting point for calculating DPMO

DPMO, or defects per million operations, is derived by first calculating DPO, then multiplying the ratio by one million. This number is used to generate Six Sigma levels for a process.

Demonstrates the failure rate of different opportunities

Unlike other metrics, DPO helps demonstrate which types of failures have a higher rate of occurrence.

3 best practices for DPO

• While a great measurement for determining the likelihood of a failure occurring per opportunity, it does not give you an understanding of the failure rate per unit produced or service provided. You can have a low DPO and still have a startlingly high unit failure rate, or DPU.
• Other defect measurements to use in conjunction with defects per opportunity include defects per unit, defects per million opportunities, and parts per million (units defective). Note that the product or service provided is the unit, while customer specifications such as size, shape, material, and timeliness are opportunities.
• Be sure to understand what constitutes a defect. In general, a defect is defined as a property of the product or service that does not meet customer specification.

Frequently Asked Questions (FAQ) regarding DPO

Can I use defects per occurrence as a key process indicator?

While a useful metric for diagnosing issues, DPO is rarely a good key process indicator, since by itself it does not illustrate the rate at which a unit leaves the process without defects. If you’re looking for a quality-based key performance indicator at a higher level, it’s often recommended to use RTY, or rolled throughput yield, though this metric has its own limitations in telling the complete story of quality.

What are some tips for data collection when calculating DPO?

Start by determining your range, either by timeframe such as a specific shift, or by a quantity of units. Make it over 30 opportunities whenever possible for statistical significance, and as big as possible within the resources available for the study. Determine the types of defects you are including in the study and in which parts of the process they occur. Focus on common causes of failure versus special causes of failure.

What is a good DPO?

A process should always strive for a DPO of 0, representing no errors. This is rarely financially feasible, as a quality level that high could cost more than the repair of a low amount of errors. 

A good DPO is the value a company can get down to without improvements to the value costing more than the defects impact the company. Continuously strive to lower the number, and watch over time for any increases that might occur.

DPO demonstrates the chance of a failure to occur

Defects per opportunity gives you the failure rate for a particular failure to occur. Monitoring different time periods will allow for a better understanding if a particular failure rate is improving, but it may not show whether yields are improving on a unit basis. When used in conjunction with other quality metrics, it can give a clear picture regarding the quality of product produced or service being provided.