Connect content changes to measurable outcomes. GSC search data, GA4 engagement, real statistical significance, and an automatic verdict you can defend.
Hypothesis: Posts with Living Content blocks show 15%+ improvement in average session duration compared to static posts.
Verdict: Confirmed
6 of 8 tracked metrics show medium or large positive effects. Primary metric (clicks) p-value: 0.018. Confidence: high.
Caveats
Every experiment resolves to one of four verdicts, each with its rationale, warnings, and the raw numbers behind it.
Multiple significant metrics and a clear positive effect, with no major warnings. Ship it, then apply it where it works.
The data moves against the hypothesis with significance. The change did not help, so revert or rethink it.
Some metrics confirm while others stay flat. A qualified win worth a closer read before you roll it out.
Effect sizes too small or the sample too thin to call. No false certainty, just an honest not yet.
Write a hypothesis. Link it to a post. The system collects weekly snapshots from GSC and GA4 across the whole experiment window, then delivers a verdict you can actually defend in a meeting.
GSC: clicks, impressions, CTR, average position. GA4: pageviews, sessions, engagement rate, average session duration. Snapshots run weekly and automatically, with no manual exports and no CSV gymnastics.
Every metric ships with a p-value, a confidence interval, an effect size (small, medium, or large), and a sample-size warning when the data is thin. No vibe-checking. No "looks like it went up."
No false certainty. A confirmed verdict requires multiple significant metrics and a clear effect; an inconclusive one says so plainly. Each verdict ships with rationale, warnings, and the raw numbers behind it.
When an experiment is confirmed, it auto-generates a Living Content recommendation for the linked post and surfaces candidates among similar articles. The win compounds: apply what worked, where it works.
Read verdicts from the public REST API, receive a webhook the moment an experiment completes, and watch every verdict land on your Pulse timeline. Connect GSC and GA4 once and they feed every test.
# list completed experiments via the public REST API curl "https://app.liquichart.com/api/v1/experiments?status=completed" \ -H "Authorization: Bearer lc_live_..." { "data": [{ "shortId": "exp_a1b2c3", "status": "completed", "resultVerdict": "confirmed" }] }
Edit a high-traffic post. Eight weeks later, the verdict says whether the change moved clicks, sessions, or session duration, and by how much.
Run two hypotheses in parallel: "title A lifts CTR" and "title B lifts position." The system reports which won, with statistical significance.
Apply Living Content to one article, hold a similar post static. Measure whether the variant-switching paragraph actually paid off in engagement.
Before a publish push, run an experiment on a representative sample. Let the verdict say "ship it" or "rewrite first."
When a stakeholder asks whether the content audit worked, paste the verdict link. Confidence interval and all.
Confirmed verdicts auto-generate Living Content recommendations for similar articles: apply what worked, where it works, with one click.
Experiments test claims, generate verdicts, and feed proven wins back into your content.
Test hypotheses with real search and engagement data. Automatic verdicts, automatic recommendations, on a free plan to start.