Test before you trust: a control group for SEO changes
Rankings move every day for reasons that have nothing to do with your work. A control group separates the effect of a change from the background noise.
- Write the hypothesis first. For example: "New title tags will raise click-through rate on product pages."
- Pick similar pages. Same template, similar traffic, same season. Split them at random into a test group and a control group.
- Change only the test group. Leave the control pages untouched for the whole test.
- Wait long enough. Four to six weeks is usually the minimum for search data to settle.
- Compare the groups, not the dates. If both groups rose by 10%, the change did nothing; the whole market moved.
A change that does not beat its control group is not worth rolling out to the rest of the site.