How to Measure Cannabis Content Performance

Start with the GSC page report, not the site total

Site-wide traffic hides everything useful. Open Search Console's Pages report and read each guide on its own: clicks, impressions, average position. A page with 4,000 impressions and 12 clicks has a title and snippet problem, not a content problem. A page sitting at position 8 with impressions climbing is one refresh away from page-one traffic. The site total tells you none of this.

Beat the 16-month window

Search Console keeps 16 months of data and silently drops the rest. For a seasonal business — and cannabis is seasonal, with 4/20 and the winter holidays moving real volume — losing the year-before-last means losing your baseline. Export monthly to a spreadsheet, or connect the free BigQuery bulk export and forget about it. Either way, do it before you need the history, because there is no recovering it after.

Use GA4 to check whether anyone reads

GSC ends at the click; GA4 says what happened next. The metric that matters for content is engagement rate — the share of sessions that lasted 10 seconds, converted, or viewed a second page — plus average engagement time on the page itself. A guide pulling clicks with 15 seconds of average engagement is winning the ranking and losing the reader, which usually means the page answers a different question than its title promises.

Judge each page against its own job

A glossary entry, a city landing page, and a pillar hub should never be graded on one scale. The glossary entry succeeds on impressions and AI citations; the landing page on direction requests and menu clicks; the pillar on the rankings of the guides it links. Before calling a page a failure, write down what it was for. Most "underperforming" content is just content measured against another page's job.

Run a quarterly keep-refresh-merge-retire pass

Once a quarter, pull the page report and sort every guide into four buckets. Keep: performing, leave alone. Refresh: positions 5 through 15 with real impressions — update and expand these first, they pay fastest. Merge: two pages splitting one query, combine into the stronger URL with a redirect. Retire: no impressions in a year and no supporting role in a cluster. The discipline is doing it every quarter and acting on the list.

A worked example

A dispensary blog with 120 posts ran this exact pass. Forty posts kept. Twenty-two sat in refresh range — position 6 to 12 on queries with volume — and got updated over two months. Eighteen were merged down to seven URLs that had been cannibalizing each other's rankings on the same strain questions. Forty were retired with redirects to their category hubs. Net result inside two quarters: fewer pages, and clicks up on the pages that mattered, visible right in the same GSC report that flagged the problem.

Frequently asked questions

What three GSC numbers matter most for content?

Clicks, impressions, and average position, read together per page. Impressions without clicks point to a title or snippet problem; position 8 with rising impressions marks a page one content refresh from real traffic.

Why export Search Console data?

Search Console only retains 16 months of history, so year-over-year comparisons quietly fall off the back end. A monthly export to a spreadsheet or BigQuery costs minutes and preserves the trend line your seasonal traffic depends on.

Which GA4 metric replaces bounce rate?

Engagement rate — the share of sessions lasting 10+ seconds, converting, or viewing a second page. For blog content, engaged sessions and average engagement time per page tell you whether people who landed actually read.

How often should content be reviewed against the numbers?

Quarterly is the working rhythm. Monthly reads react to noise; annual reads let dead pages sit for a year. Each quarter, sort pages into keep, refresh, merge, or retire, and act on the list before the next quarter starts.