Table of Contents
Before/After Analysis Guide
Validate your process changes with statistical proof. Paste measurements, get instant results.
Start NowWhat Is This Tool?
The Before/After Analysis tool compares two groups of measurements -- one from before a process change and one from after -- to determine whether the change made a statistically significant difference. It uses the Welch's t-test to account for different group sizes and variances, and presents the result in plain language so you don't need to be a statistician to understand it.
When To Use It
- After a process change (new tooling, updated parameters, different material)
- After a training initiative to measure its effect on quality
- When a consultant or engineer needs to prove an improvement to management
Before You Start
You'll need two sets of measurements from the same process and characteristic:
- "Before" measurements collected before the process change
- "After" measurements collected after the process change
- Optional: specification limits (USL/LSL) if you want Cpk comparison
1 Setup
Enter a descriptive title for your analysis, optional specification limits (USL and LSL) for Cpk comparison, and any notes about the process change.
Click to view screenshot
2 Paste Your Data
Paste your measurements in the text area. You can use two columns (Before | After) side by side, or one value column with a group column (Group | Value). The tool auto-detects the format. Minimum ~10 values per group recommended for reliable results.
Click to view screenshot
3 Review Results
The results page shows a verdict card (Improved / No Change), an auto-interpretation paragraph, a box plot comparing both groups, and a significance card with the p-value and Cohen's d effect size. Click "Show Details" for the full statistical breakdown including mean shift, sigma reduction, and confidence interval.
Click to view screenshot
4 QC-Coach
Choose how you'd like the results explained by selecting a persona chip: Quality Engineer (technical details), Production Manager (practical impact), Executive Summary (business brief), Training Guide (operator-friendly), or Root Cause Next Steps (investigation plan). You can also type your own follow-up question.
Click to view screenshot
5 Download Report
Generate a professional PDF report with the analysis summary, charts, statistical details, and QC-Coach transcript. Save the analysis online to view or share later via a link.
Click to view screenshot
Tips & Best Practices
- Collect at least 20-30 values per group for reliable results. More data gives stronger conclusions.
- Ensure both groups measured the same characteristic on the same gauge. Mixing measurement methods invalidates the comparison.
- A non-significant result doesn't mean "no change" -- it means the data can't confirm one. Consider collecting more data.
- Share the Executive Summary QC-Coach output with stakeholders -- it translates p-values and effect sizes into business language.
Understanding the Statistics
Show statistical explainer
p-value: The probability the observed difference happened by chance. Below 0.05 means statistically significant. Cohen's d: Measures the practical size of the difference (small < 0.2, medium 0.5, large > 0.8). Confidence interval: The range where the true difference likely falls (95% confidence). Cpk delta: How much the process capability index changed -- positive means improvement toward meeting spec limits.