Why we wait three weeks before telling you whether a change worked

When you change a page, the temptation is to check it the next morning. We built SearchLift to refuse to answer that question for three weeks, and it's worth explaining why, because it's the single most common way SEO effort gets wasted.

Rankings are noisy at short timescales

A page's average position moves day to day for reasons that have nothing to do with your edit: who else published, which queries happened to fire, personalised and localised results, and Google's own re-crawls. Look at any page's daily position line and you'll see a band, not a value.

Read that band two days after a change and you are reading the band, not the change.

The failure this causes

The damage isn't that you learn nothing. It's that you learn something false and act on it.

A page improves, dips for two days on noise, and gets reverted. Or a page is unchanged, drifts up for two days on noise, and the change gets called a success and copied across the site. Both are common, and both are worse than having waited.

Three weeks, and what we compare

Three weeks is long enough for a re-crawl and for daily variance to average out on most pages. We record impressions, clicks, click-through rate and average position before the change, then compare the same window afterward — not the day before against the day after.

Then we say plainly whether it helped, hurt, or made no measurable difference. That last verdict is the one most tools won't give you, and it's the most common honest answer.

What this means in practice

Make several changes, then wait. Batching is fine — the comparison is per page, so pages don't interfere with each other. What doesn't work is changing a page, staring at it, and changing it again on Thursday.