Playbook: measure recovery without fooling yourself

Playbook: measure recovery without fooling yourself
Most 'we recovered' claims are noise. A disciplined measurement protocol:

— Step 1: Define recovery upfront as impressions+clicks regaining 80% of pre-drop baseline for 14 consecutive days, not a one-day spike.
— Step 2: Track at the page-cluster level — site-wide averages hide that some pages recovered and others stayed flat.
— Step 3: Wait for the next confirmed core update to validate — Google has repeatedly said HCS-style recovery often lands on a subsequent update, per their own statements.
— Step 4: Hold a control set of untouched pages to separate your fixes from rising tides.
— Step 5: Annotate your charts with every deploy date.

Don't celebrate intra-rollout bounces; they reverse.

One to bookmark: Google's ranking updates page — the canonical timeline to align recovery windows against.