Stop guessing. Start proving.

Connect content changes to measurable outcomes. GSC search data, GA4 engagement, real statistical significance, and an automatic verdict you can defend.

app.liquichart.com/experiments/living-content-engagement

Does Living Content improve time on page?

Confirmed

Hypothesis: Posts with Living Content blocks show 15%+ improvement in average session duration compared to static posts.

Statistical Analysis

high confidence
Metric
Before
After
Change
Effect
p-value
Sig.
Clicks(gsc)661816+23%1.12 (large)0.018
Impressions(gsc)11,12512,300+11%0.68 (medium)0.042
CTR(gsc)5.90%6.60%+12%0.72 (medium)0.038
Avg Position(gsc)7.86.5+16%0.85 (large)0.024
Pageviews(ga4)1,9702,450+24%1.05 (large)0.022
Sessions(ga4)1,5181,900+25%1.10 (large)0.019
Engagement Rate(ga4)61.0%69.0%+13%0.78 (medium)0.035
Avg Duration(ga4)155s185s+19%0.90 (large)0.028

Verdict: Confirmed

6 of 8 tracked metrics show medium or large positive effects. Primary metric (clicks) p-value: 0.018. Confidence: high.

Caveats

  • No control group. Observed changes may be due to external factors (seasonality, algorithm updates)

Metrics

+5%

890

+2%

12,800

+3%

7.00%

+5%

6.1

+5%

2,650

+5%

2,050

+3%

72.0%

+4%

198s

Search Performance

Traffic & Engagement

Four honest verdicts, no false certainty.

Every experiment resolves to one of four verdicts, each with its rationale, warnings, and the raw numbers behind it.

Confirmed

Multiple significant metrics and a clear positive effect, with no major warnings. Ship it, then apply it where it works.

Refuted

The data moves against the hypothesis with significance. The change did not help, so revert or rethink it.

Partially confirmed

Some metrics confirm while others stay flat. A qualified win worth a closer read before you roll it out.

Inconclusive

Effect sizes too small or the sample too thin to call. No false certainty, just an honest not yet.

Five capabilities, one closed loop.

Hypothesis to verdict

Close the loop without the spreadsheet.

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 and GA4 out of the box

Eight metrics, one click to authorize.

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.

Real statistics

Permutation tests, p-values, effect size.

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."

Four honest verdicts

Confirmed, refuted, partial, inconclusive.

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.

Verdict to recommendation

A confirmed test becomes a playbook.

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.

From hypothesis to recommendation, all through one API.

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.

  • GET /api/v1/experiments returns status and verdict for every completed test
  • experiment.completed webhook fires the moment a verdict is reached
  • One-time GSC and GA4 auth, reused across every experiment
  • Each verdict emits an experiment_verdict beat you can filter or route to Slack
# 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"
  }]
}
REST API · webhook · GSC + GA4 · permutation tests

Six experiments worth
running this quarter.

01
Test a rewrite, prove the lift

Edit a high-traffic post. Eight weeks later, the verdict says whether the change moved clicks, sessions, or session duration, and by how much.

02
A/B titles before you commit

Run two hypotheses in parallel: "title A lifts CTR" and "title B lifts position." The system reports which won, with statistical significance.

03
Validate the Living Content thesis

Apply Living Content to one article, hold a similar post static. Measure whether the variant-switching paragraph actually paid off in engagement.

04
Pre-commit a content bet

Before a publish push, run an experiment on a representative sample. Let the verdict say "ship it" or "rewrite first."

05
Ship evidence with the review

When a stakeholder asks whether the content audit worked, paste the verdict link. Confidence interval and all.

06
Compound the wins

Confirmed verdicts auto-generate Living Content recommendations for similar articles: apply what worked, where it works, with one click.

FAQ

Turn content strategy into evidence.

Test hypotheses with real search and engagement data. Automatic verdicts, automatic recommendations, on a free plan to start.