Control Contamination (SEO)
Also known as contaminated control, control group contamination
The unintentional modification of an observed variable before a baseline measurement is complete, making it impossible to attribute later changes to a single cause.
What it means in the system.
Control contamination happens when the thing being observed gets changed before a clean baseline is recorded, usually because the observer reacts to what they found before documenting it first. In SEO this is common: a practitioner notices a weak page, immediately improves it, and only afterward realizes there's no clean before-state to compare the after-state against. Once that happens, later changes in visibility can't be cleanly attributed to the intervention alone, because the true starting point was never captured.
Where it shows up.
A thin glossary page is found holding a SERP position. Instead of screenshotting it first, the page gets rebuilt on the spot. Any later movement in rankings can no longer be cleanly separated from what the page looked like at the moment it was actually observed, only from whatever historical data (search history, warehouse logs) happened to already exist before the change.
Contaminated control doesn't make an observation worthless, it just downgrades the confidence of anything measured afterward. The honest response is to say so explicitly, fall back on whatever historical data already existed as a proxy baseline, and treat the resulting conclusions as directional rather than proven.