Upload an inspection log, map the lot, the units inspected and the defects found, and get the operating-characteristic curve, the producer's and consumer's risk, the AOQ curve with its AOQL, every lot's accept or reject decision, and the smallest plan that would meet your AQL. Free.
Free analyses run on up to 10,000 rows. Larger files are randomly sampled to that size — sign up to analyze your full dataset.
Computing the operating characteristic curve and lot decisions...
Sent to — OC curve, producer's and consumer's risk, AOQ curve and AOQL, per-lot decisions with confidence bounds, plan design search, R code, and AI insights.
Analyze another fileThe record is aggregated to one row per lot (repeated identifiers are summed, preserving the inspected total) and the plan in force is resolved as a sample size n and an acceptance number c, either from module parameters or from the data itself. The probability model is chosen from the sample fraction: hypergeometric when the sample exceeds about a tenth of the lot, binomial otherwise, and the choice is stated. The operating-characteristic curve Pa(p) is then computed exactly — pbinom(c, n, p), or phyper(c, D, N-D, n) with D the whole number of defectives a lot of N at rate p would hold. The producer's risk is 1 - Pa(AQL), the consumer's risk is Pa(LTPD), and the defect rates at which the curve crosses 95, 50 and 10 percent acceptance are found by bisection. The AOQ curve is p x Pa(p) with the rectifying correction (N-n)/N where the lot size is known, and its maximum — found on a dense grid and refined — is the AOQL. Average total inspection is n + (1 - Pa(p))(N - n). Each lot gets a one-sided 95 percent Clopper-Pearson upper bound on its true defect rate. The inverse problem is solved by searching sample sizes upward, taking for each the smallest acceptance number that keeps acceptance at the AQL above the producer's-risk target, and stopping at the first plan whose acceptance at the rejectable level falls to the consumer's-risk target.
Use it when lots of discrete units arrive and each has to be taken or returned on the evidence of an inspected sample, and you need to know what the sampling plan actually does: how often it will return good lots, how often it will take bad ones, how much defective work still gets through, and what plan would meet the risks you agreed.
Not for asking whether a process is stable over time (that is a control chart), not for asking whether its output fits a specification window (that is a capability study), and not for a measured characteristic rather than a defective count (that is variables sampling). It also cannot certify conformance to a published sampling scheme.
Built for: Quality engineers, incoming-inspection and receiving teams, supplier-quality managers, and auditors who have to justify a sampling plan
Typical data source: An incoming-inspection log or receiving report: one row per lot with the lot number, how many units were inspected, and how many failed — or one row per inspected unit with a pass/fail result
One row per lot: what was inspected and what was found. A lot-size column is optional but worth mapping.
Minimum 5 rows · Best with 20-500 lots
Standard-library analysis: should I accept this lot? Map an inspection record — the lot identifier, how many units were inspected, and how many were found defective — and get the operating-characteristic curve of the sampling plan in force, the producer's risk (alpha at the AQL) and the consumer's risk (beta at the LTPD), the average outgoing quality curve with its worst point (the AOQL), the average sample number and average total inspection, the accept/reject decision for every lot with a one-sided upper confidence bound on each lot's true defect rate, and the inverse problem solved by search — the smallest (n, c) plan that meets a target AQL and LTPD at the two risks you name. Binomial or hypergeometric probabilities are selected automatically and the choice is stated.
The probability the plan accepts a lot, at every true defect rate it might have — with the AQL and the rejectable level marked, so both risks can be read straight off the curve.
The producer's risk of returning a good lot and the consumer's risk of taking a bad one, with the quality levels each is measured at.
How much defective work still reaches downstream once the plan has done its sorting; the peak is the AOQL, the worst long-run average the plan can produce.
Each lot's defect count against the acceptance number, coloured by the decision that follows.
Per-lot decisions with a 95 percent upper bound on each lot's true defect rate — because a clean sample bounds the rate rather than zeroing it.
The smallest sample size and acceptance number that meet your AQL, rejectable level and both risk targets at once.
Every formula in full, plus what acceptance sampling does not do.
Plain-English interpretation — what the numbers mean, what's significant, and what to do next.
Is the plan we inherited doing anything?
Map your inspection log and get the operating-characteristic curve of the plan you are actually running. Many inherited plans inspect a handful of units and reject almost nothing: if a lot has to be 58 percent defective before the plan catches it nine times in ten, the analysis says so instead of reporting the accept decisions as if they meant something.
See our FAQ for details on pricing, data privacy, and how the analysis works. Every report includes a Methodology section showing the statistical test, assumptions checked, and diagnostics run.
Run any analysis on your own data — validated R analyses, interactive reports, AI insights, and PDF export.
Try Free — No Credit CardTell us what went wrong, in your own words. We capture the page you're on automatically, so no need to describe where you are.