Three months ago I decided to stop guessing about click-through rates and start measuring them properly. The setup was straightforward: 40 pages split into two groups, one group with rewritten title tags emphasising numbers and specificity, the other left unchanged as a control. Google Search Console was the primary data source, pulling weekly CTR averages for each URL cluster.

The Experiment Setup

  1. Export all pages with more than 200 impressions per week from Search Console.
  2. Segment them into two equal groups matched by average position and topic category.
  3. Rewrite title tags for Group A using a consistent structure: primary keyword, a specific number, and a clear outcome phrase.
  4. Log the baseline CTR for both groups over two weeks before making any changes.
  5. Run for 60 days, pulling weekly reports every Monday morning.

Pages in Group A showed a measurable shift by week three. Average CTR moved from 3.1 percent to 4.6 percent across the test group, while the control held steady between 3.0 and 3.2 percent throughout the period.

The most consistent gains came from pages that had strong impressions but weak CTR — a combination that signals a ranking without persuasion.

What did not work: adding emotional language to titles showed no consistent improvement. Specificity outperformed sentiment every time.

Steps to Replicate This

  • Filter Search Console for high-impression, low-CTR URLs first.
  • Rewrite titles with a measurable element — a year, a count, a timeframe.
  • Set a fixed review date and do not change anything mid-experiment.
  • Compare against a genuine control group, not just your own historical average.

This kind of experiment takes patience. Results do not arrive in days, and a single month of data is rarely enough to draw solid conclusions.