Package insert A/B test planner and results worksheet
Plan an insert card experiment, compare observed response rates and costs, and download a CSV with your assumptions. Free and local to your browser.
Plan your comparison, then record observed outcomes. Count each responding recipient once, using the same definition for A and B. Repeated scans are not unique recipients. Distributed inserts means those delivered to recipients, not the quantity ordered from a printer.
Enter total attributable costs for each group consistently. Use 0 for confirmed zero; leave an unknown cost blank. No currency conversion or tax calculation is performed.
Observed results
Distributed inserts: enter a whole number from 0 to 1,000,000. Complete valid counts and costs to see results.
More responses can mean a lower rate
In the fictional example, A receives 10 responses from 200 recipients (5%), while B receives 12 from 400 (3%). B has more responses but a rate 2 percentage points lower. Neither observation establishes that the artwork caused the difference. Keep assignment, delivery conditions and the observation window with your report.
Keep the interpretation honest
Choose the outcome and closing rule before reviewing results. Random assignment, when feasible, helps separate the treatment from audience differences. These descriptive calculations do not provide a sample size recommendation, a significance test, revenue or ROI. Zero distributed recipients makes the rate unavailable; zero responses makes cost per response unavailable.
NIST: completely randomized designs explains random assignment of a factor to experimental units. This worksheet records your assignment method; it does not perform randomization or a statistical test.
Your inputs stay in this page's memory until reload or navigation. The CSV contains what you entered. No account save, tracking service or campaign submission is performed.
Build distinct campaign links →
How this tool works
Observed response rate = unique responding recipients ÷ distributed inserts. B minus A is reported in percentage points, not causal lift or statistical significance.
Use the actual specification from your selected printer. Defaults are examples, not universal production requirements. These tools do not inspect a PDF, certify a print file, provide a supplier quote or predict sales.