Control set

Without one, every SEO claim is a before-and-after chart, and a before-and-after chart cannot separate your work from a seasonal swing, a competitor's outage or a core update. With one, the question becomes whether the changed pages moved differently from comparable pages that did not change.

The set is chosen at landing time by similarity: comparable crawl depth, similar impression volume over the 28 days before landing, indexable and returning 200s. Then it is frozen and never recomputed, because a control that moves with the treatment is not a control.

Sometimes there is no adequate set. A small site may not contain enough comparable untouched pages, and a sitewide change contaminates its own control by definition. In those cases the fix is recorded as having no clean comparison rather than being given a verdict it did not earn.