DPPM, defective parts per million, is the count of defective units divided by the total units produced or shipped, times one million. It measures how many bad parts a customer receives per million, and it counts a part as one reject whether it has one flaw or five. That single choice, counting parts instead of defects, is what separates DPPM from every other million-scale quality metric.
If a customer sends you a supplier scorecard, DPPM is almost certainly the headline number on it. Understanding exactly what it counts, and what it deliberately ignores, is the difference between managing your score and being surprised by it. This guide calculates DPPM, shows why customers lean on it, and draws the line between DPPM, DPMO, and plain PPM.
How do you calculate DPPM?
DPPM needs two numbers: how many units you shipped, and how many of those were defective. A defective unit is any part the customer rejects, regardless of how many things are wrong with it. The formula is simple. DPPM = (defective units ÷ total units) × 1,000,000.
Ship 40,000 parts and have the customer reject 6, and your DPPM is (6 ÷ 40,000) × 1,000,000 = 150 DPPM. The million scale exists because modern reject rates are tiny; expressed as a percentage that is 0.015 percent, a number too small to manage or trend, while 150 DPPM is a figure a supplier quality engineer can act on and compare month to month. Note what the formula does not include: nowhere does it ask how many defects were on those 6 parts. Three of them could have had two flaws each and the DPPM is unchanged. DPPM counts rejects, full stop.
Because it counts parts, DPPM is closely related to first pass yield just expressed as a defect rate at the million scale rather than a percentage passing. A supplier running 150 DPPM is passing 99.985 percent of parts on the metric the customer cares about: did the part arrive good or not.
Why do customers track supplier DPPM?
Customers track DPPM because it answers the only question their receiving dock cares about: how many bad parts will I get? A defect count matters to the supplier improving its own process, but the customer lives with rejects, line stoppages, and sorts, all of which scale with defective parts, not with defects per part. DPPM is the metric that maps to the customer's pain, which is why it became the common language of supplier quality, especially in automotive where a single bad part can stop an assembly line.
On a supplier scorecard DPPM is usually the top-line quality number, tracked monthly and rolled into a rating that governs new business. Automotive customers commonly hold suppliers to aggressive targets, often in the low double digits or single digits of DPPM, with best-in-class parts expected near zero. Miss the target and you can land on a watch list, trigger a containment demand, or be asked for a corrective action. That is why DPPM is not just a quality metric; it is a commercial one, and it belongs at the center of any real supplier quality management program.
The DPPM definition and its role in supplier quality are documented by the recognized bodies:
- DPPM = (defective parts ÷ total parts) × 1,000,000 counting defective units rather than individual defects (ASQ, Defective Parts Per Million calculation).
- PPM is the standard language of automotive supplier quality governed alongside FMEA, SPC, and MSA by the automotive core tools (AIAG Quality Core Tools).
- For a part with a single defect opportunity, DPPM and DPMO are identical; they diverge only when a part carries multiple opportunities for a defect.
How is DPPM different from DPMO and PPM?
The three metrics look alike and mean different things, and mixing them up is a fast way to argue past your customer. DPMO defects per million opportunities, counts every defect against every opportunity for a defect, so a complex part with many features is scaled by its opportunity count. DPPM counts defective parts against total parts and ignores opportunities entirely. PPM, parts per million, is the loose umbrella term, and there is the trap: in a supplier quality context PPM almost always means DPPM, defective parts, but in Six Sigma shorthand people sometimes use PPM to mean DPMO. Always confirm which one a scorecard means before you compare numbers.
The numbers only coincide in one case. For a part with a single defect opportunity, one thing that can be wrong, DPPM and DPMO are the same, because one defect equals one defective part. As soon as a part has multiple opportunities, they separate, and they can separate by a lot. A board with 200 solder joints and one bad joint is one defective part but one defect in 200 opportunities, so its DPPM and DPMO differ by more than two orders of magnitude. Quote the wrong one and you can look a hundred times better or worse than you are.
Should DPPM use parts shipped or parts received?
This looks like a footnote and is actually the source of most DPPM disagreements. The supplier naturally computes DPPM against parts shipped, its own count. The customer computes it against parts received and rejected at their dock. Those denominators are close but not identical, and the rejects that count can differ too: does a part sorted and reworked at the customer before it reached the line count against you, or only a part that actually stopped production? Two honest teams using different rules will post different DPPM for the same month and each will be sure the other is wrong.
The fix is to write the rules down before the first number is reported: the denominator, what qualifies as a defective part, whether sort finds and field returns are included, and the exact reporting window. Automotive supplier agreements usually specify this, and where they do not, it is worth a one-page addendum. A DPPM figure without an agreed definition is not a measurement, it is an opening position in an argument, and the time to settle it is before a bad month, not during one.
The three metrics look alike and mean different things, and mixing them up is a fast way to argue past your customer. DPMO defects per million opportunities, counts every defect against every opportunity for a defect, so a complex part with many features is scaled by its opportunity count. DPPM counts defective parts against total parts and ignores opportunities entirely. PPM, parts per million, is the loose umbrella term, and there is the trap: in a supplier quality context PPM almost always means DPPM, defective parts, but in Six Sigma shorthand people sometimes use PPM to mean DPMO. Always confirm which one a scorecard means before you compare numbers.
The numbers only coincide in one case. For a part with a single defect opportunity, one thing that can be wrong, DPPM and DPMO are the same, because one defect equals one defective part. As soon as a part has multiple opportunities, they separate, and they can separate by a lot. A board with 200 solder joints and one bad joint is one defective part but one defect in 200 opportunities, so its DPPM and DPMO differ by more than two orders of magnitude. Quote the wrong one and you can look a hundred times better or worse than you are.
| Metric | Numerator | Denominator | Best for |
|---|---|---|---|
| DPPM | Defective parts (rejects) | Total parts | Customer receiving, supplier scorecards |
| DPMO | Total defects | Units × opportunities | Comparing processes of different complexity |
| PPM (general) | Usually defective parts | Total parts | Umbrella term; confirm the definition |
How do you use DPPM to manage suppliers?
DPPM is only useful if the count is clean and the response is disciplined. Run it in this order.
- Agree on the definition. Settle with the customer or supplier what counts as a defective part, whether the denominator is parts shipped or parts received, and the reporting window. Most DPPM disputes are definition disputes.
- Count rejects completely. Capture every rejected part, including returns, sort finds, and line rejects, not just formal complaints. An undercount flatters the number and hides a real trend.
- Compute and trend monthly. Calculate DPPM per part number and per supplier, and watch the trend, because a single month is noisy while a rising line is a signal.
- Tie misses to action. When DPPM breaches the target, trigger a non-conformance report and a corrective action, not just a red cell on a chart. The number exists to drive a fix.
- Segment by defect, not just by part. Because DPPM hides how many defects each reject had, pair it with defect tracking by type so you know which failure mode is driving the rejects.
What DPPM hides, and how to cover the gap
DPPM's simplicity is also its blind spot. By counting a part as one reject no matter what, it tells you how many bad parts shipped but nothing about why, or how close the good parts came to failing. A supplier can hold a low DPPM while its process quietly drifts toward the edge, because parts that are barely in spec still count as good. DPPM is a lagging outcome, not an early warning.
Cover the gap by pairing DPPM with leading indicators from the floor: defect type and location, near-miss and marginal readings, and process trends. That only works if the underlying count is complete and captured where the parts are made, not reconstructed from memory at month end. When rejects and defects are logged live at the source and rolled up automatically, DPPM reflects reality and you can see the drift behind it before it becomes a customer reject. That is the kind of real-time quality record Harmony builds for manufacturers on top of the systems they already run, with no rip-and-replace (see how CLS replaced paper logging), and it ties your reject data to the same place that holds your cost of quality. More on connected quality data is on our features overview.