Decide what would interrupt an A/B test before the first visitor arrives.

A fictional studio compares two consultation buttons. Halfway through the planned window, its form starts failing on mobile. A higher completion rate on one button is now only part of the story.

Write the interruption rules alongside the experiment plan: • Define the eligible audience, assignment method and exclusions before the run • Use one reporting timezone and a planned start and end • Name the functional failures or data-quality issues that require an early stop • If a guardrail is breached, record when it happened, which observations may be affected and who owns the investigation • Keep the original plan and source exports with the deviation record so the next reviewer can understand what changed

Choose the next step after reviewing the issue and the evidence. Reaching a minimum count does not resolve a broken form or make an interrupted run a reliable winner.

The preview is the actual Predeclare the plan worksheet from the Business Experiment and A/B Results Kit. The download includes an editable Word packet, Excel tracker, printable PDFs, a fictional completed example and Quick Start.

The kit documents a manual comparison. It does not assign website traffic, detect failures or perform a statistical significance test.

Explore the kit: https://contra.com/products/DT2ZXGKY-business-experiment-and-ab-results-kit