# Master Data Defect Rate, DPMO and Sigma Level Calculator > Source: https://netray.co/tools/mdm-record-defect-rate-sigma-calculator > Publisher: Netray (netray.co), AI and integration engineering for ERP and manufacturing systems. > Section: Free Tools ## Short answer Enter records tested, checks per record and defects found to get defect rate, DPMO, first-pass accuracy and sigma level. ## Category quality ## Audience - MDM analyst - Data steward - Quality engineer - ERP admin ## Tool **Kind:** calculator **Inputs:** #### Records tested **Key:** records **Type:** number **Default:** 5000 **Min:** 1 **Step:** 100 **Unit:** records **Help:** Number of active master records you ran the checks against. #### Rule checks per record **Key:** checks **Type:** number **Default:** 8 **Min:** 1 **Step:** 1 **Unit:** checks **Help:** Number of distinct rules tested on each record, such as completeness, validity and consistency tests. #### Failed checks found **Key:** defects **Type:** number **Default:** 360 **Min:** 0 **Step:** 10 **Unit:** defects **Help:** Total number of rule failures across all records. #### Records with at least one failure **Key:** bad_records **Type:** number **Default:** 290 **Min:** 0 **Step:** 10 **Unit:** records **Help:** Distinct records that failed one or more checks. #### Average cost to fix one record **Key:** fix_cost **Type:** number **Default:** 12 **Min:** 0 **Step:** 1 **Unit:** $ **Help:** Steward time plus downstream rework, per record. **Outputs:** #### Total opportunities **Key:** opportunities **Expr:** records * checks **Format:** integer #### Defect rate per check **Key:** defect_rate **Expr:** defects / opportunities * 100 **Format:** percent #### DPMO **Key:** dpmo **Expr:** defects / opportunities * 1000000 **Format:** integer **Primary:** true **Description:** Defects per million opportunities. #### Sigma level (with 1.5 shift) **Key:** sigma **Expr:** if(gt(defects,0), normsinv(1 - defects / opportunities) + 1.5, 6) **Format:** decimal **Primary:** true **Description:** Short-term sigma using the conventional 1.5 shift. #### Record accuracy (clean records) **Key:** record_accuracy **Expr:** (records - bad_records) / records * 100 **Format:** percent **Description:** Share of records passing every check. #### Estimated rework cost **Key:** rework_cost **Expr:** bad_records * fix_cost **Format:** currency **Chart:** #### DPMO as failed checks rise **Type:** line **X:** defects **From:** 0 **To:** 1500 **Steps:** 30 **Series:** - dpmo **Disclaimer:** Sigma is computed with the conventional 1.5 sigma shift used in Six Sigma tables and assumes defects are independent. Treat it as a comparison index across master data domains and months, not as a precise process capability figure. ## Content **Intro:** Quality teams know DPMO and sigma levels from the shop floor. The same maths works for master data: treat every rule check on a record as an opportunity and every failure as a defect. This calculator turns your profiling counts into a defect rate, DPMO, sigma level and rework cost. **Sections:** #### How to use the calculator **Body:** Run your data quality checks across a set of active records. Enter the number of records tested, the number of rule checks applied to each record, the total number of failed checks and the number of distinct records with at least one failure. Add an average cost to fix one record if you want a rework estimate. The calculator reports defects per million opportunities, a sigma level and the share of records that passed everything. The chart shows how DPMO climbs as failures increase, so you can see the impact of cleanup targets. #### The formulas **Body:** Opportunities equal records tested times checks per record. Defect rate equals defects divided by opportunities. DPMO equals that rate times one million. Sigma level uses the standard normal inverse of the yield, one minus the defect rate, plus the conventional 1.5 shift. Record accuracy is different from defect rate: it counts a record as bad if any check fails, so it is always lower than check-level yield. **Bullets:** - Opportunities = records x checks per record - DPMO = defects / opportunities x 1,000,000 - Sigma = NORMSINV(1 - defect rate) + 1.5 - Record accuracy = clean records / records tested #### Worked example **Body:** With 5,000 records and 8 checks each there are 40,000 opportunities. If 360 checks fail, the defect rate is 0.9 percent, DPMO is 9,000 and the sigma level is about 3.9. If 290 distinct records have a failure, record accuracy is 94.2 percent. Fixing 290 records at 12 dollars each costs about 3,480 dollars, which is the cleanup budget for that snapshot. A good use is comparing vendor, customer and item masters on the same scale, then focusing effort on the lowest sigma. #### Common mistakes **Body:** Do not change the number of checks between months and compare sigma, because adding easy checks inflates sigma. Do not count one root cause several times: if a bad unit of measure fails three rules, decide whether that is one defect or three and apply the choice consistently. Do not use sigma to compare master data to a manufacturing process in absolute terms. Finally, always report record accuracy next to sigma, since business users understand it more easily. **Cta:** If you would like these defect counts produced automatically from your ERP each month, we can build the checks and the scorecard on your own infrastructure. **Faq:** #### What is a good sigma level for master data? **Answer:** There is no universal target. Many teams treat 4 sigma, about 6,200 DPMO, as a reasonable first goal for transaction-critical rules and aim higher for fields that block orders or payments. Compare against your own trend. #### What counts as an opportunity? **Answer:** An opportunity is one rule check applied to one record. If you test eight rules on every record, each record offers eight opportunities. Keep the rule set fixed when you compare periods. #### Why is record accuracy lower than the yield? **Answer:** Yield counts failed checks across all opportunities. Record accuracy counts a whole record as bad if any one check fails. A few failures spread across many records lowers record accuracy much more than yield. #### What does the 1.5 shift mean? **Answer:** It is a Six Sigma convention that long-term performance drifts about 1.5 sigma from short-term performance, so published tables add 1.5. It is a convention rather than a measured value, which is why sigma is best used for comparison. ## Search terms dpmo calculator, data quality sigma level, master data error rate, record accuracy percentage calculator, defects per million opportunities, data defect rate formula, six sigma data quality --- Talk to an engineer: https://netray.co/contact?intent=erp_ai | Machine catalog: https://netray.co/agent/catalog.json