Title Tag CTR Test

What it produced for CGH Injury Lawyers

A complete, runnable title-tag test for the CGH /denver/car-accident-lawyer/ money page. It follows the skill exactly: confirm a trigger, change ONE lever, record the change date, wait 2-4 weeks, then compare CTR for the same query set before vs after, normalized for position.

Target page

https://cghlawfirm.com/denver/car-accident-lawyer/ (the Denver car accident money page, the primary conversion destination for the firm's highest-value intent).

Trigger check (one is enough)

Run this against the firm's Search Console via seo-utils query_gsc (page = the Denver car accident lawyer URL, last 28 days, by query):

• Pull the page's query set, sort impressions descending, CTR ascending.
• The test is justified if the page sits top 10 (good rankings) but its CTR trails the position benchmark (e.g. an average position of 6-9 with CTR materially below the expected ~3-5% for that band), OR shows high impressions, low clicks, OR CTR trailing competitors on "car accident lawyer denver" and "denver car accident lawyer".

Per the skill, any single one of these is enough to test. In a high-competition PI SERP crowded with aggressive competitor titles, a strong-rank/low-CTR pattern on the head money term is the expected trigger.

One lever to change (per cycle)

Lever chosen: location position (front vs middle). Move the city to the front of the title so the geo-qualifier is the first thing the searcher sees in a local-intent query, where proximity language drives clicks.

• Control (current shape): Car Accident Lawyer in Denver, CO | CGH Injury Lawyers
• Variant (single lever moved, location to front): Denver Car Accident Lawyer | Free Case Review | CGH Injury Lawyers

Only the location position is the attributable change being measured. (The Free Case Review value-add is held constant across both ends if present; if the control lacks it, that becomes the lever for a separate later cycle, never combined here. One lever per cycle is the rule.)

To keep the test clean, the canonical single-lever pair is:

• Control: Car Accident Lawyer in Denver, CO | CGH Injury Lawyers
• Variant: Denver, CO Car Accident Lawyer | CGH Injury Lawyers

Identical words, only the city's position moves front vs middle. That is the one attributable lever.

Protocol (the close-the-loop steps)

1. Before snapshot. seo-utils query_gsc, page-filtered to /denver/car-accident-lawyer/, date range = the 28 days before the change, dimensions = query. Save: impressions, clicks, CTR, average position per query. This is the baseline; without it there is no test, only an edit.
2. Ship the variant and record the change date (ISO 8601, e.g. 2026-06-23). Log it in the title formula bank row for city/practice money page.
3. Wait 2 to 4 weeks for stable data (PI is not heavily seasonal, so 3 weeks is the planned window; extend to 4 if click volume is thin).
4. After snapshot. Same query_gsc pull, same page, same query set, date range = the 28 days after the change.
5. Compare, normalized for position. For the same query set, compute CTR before vs CTR after. Then normalize for average position: if average position improved alongside the CTR lift, the lift is NOT attributable to the title (a higher slot earns more clicks on its own). A CTR lift with flat or unchanged position is the win. Also normalize for any obvious seasonality in click volume.

Worked verdict frame (the shape the result is read in)

Query (same set both ends)	CTR before	Avg pos before	CTR after	Avg pos after	Position-normalized verdict
denver car accident lawyer	[GSC]	[GSC]	[GSC]	[GSC]	Win only if CTR up AND position flat
car accident lawyer denver	[GSC]	[GSC]	[GSC]	[GSC]	Win only if CTR up AND position flat
car accident attorney denver	[GSC]	[GSC]	[GSC]	[GSC]	Win only if CTR up AND position flat

[GSC] cells are filled from the real seo-utils query_gsc pulls at each end of the window. No values are invented; the firm's own Search Console supplies the proof.

Close the loop

• If the location-front variant wins (CTR up, position flat), roll "location front" into the title formula bank row for the city/practice money page page type, so every future CGH city money page (truck, motorcycle, premises across Denver and the expansion metros) starts from the proven shape instead of a fresh guess.
• If it loses or is inconclusive, revert and queue the next single lever (power-word/value-add swap, or number present vs absent) for a fresh cycle.

Note on numbers: no CGH CTR/impression/position figure is asserted here; all metric cells are pulled live from CGH's own Google Search Console. PI-vertical competition is the documented assumption behind choosing the head money term as the test target.