A/B Test Group Assigner
Assign a list of users, records, or IDs into randomized A/B test groups at any split ratio — a fair building block for experiment design.
Generate
How the A/B Test Group Assigner works
Valid A/B testing depends on random assignment: every unit in your experiment population needs an equal, independent chance of landing in each group so that any measured difference in outcomes can be attributed to the treatment rather than to some hidden pattern in how groups were formed. This tool shuffles your full list with an unbiased Fisher–Yates shuffle, then splits it at your chosen percentage.
Because the split is based on percentage rather than a fixed count, this tool naturally supports both classic 50/50 splits and unequal ratios, such as a 90/10 split for a cautious rollout of a new feature to a small initial test group.
How to use it
Frequently asked questions
Can I run an A/B/C test with three groups?
Not directly in this tool; run it twice — first splitting into your combined B+C pool versus A, then splitting that pool again for B versus C — to build a three-way split.
Does group size divide exactly at the percentage I choose?
As closely as rounding allows — group sizes are rounded to the nearest whole number of your total list, so very small lists may show a slightly different ratio than requested.
Is this method suitable for a real production experiment?
The random assignment mechanic is sound, but production A/B testing systems typically also handle consistent bucketing across sessions, statistical power calculations, and guardrail metrics that this simple tool doesn't provide.