48.3k Log in Get started
Metric · Marketing

What is landing-page conversion rate, and how do you measure it in Metabase?

Landing-page conversion rate is conversions divided by sessions, per landing page. The per-page part is the point: sitewide conversion rates move with channel mix, not page quality, so the actionable number lives at the page level. Measure it in Metabase from session data synced from Google Analytics 4 or Plausible.

TL;DR — conversions / sessions grouped by landing page, with a minimum-sessions threshold (HAVING sessions >= 100) before ranking anything. Read rate and volume together: a 20% page with 50 sessions loses to a 5% page with 10,000.

What landing-page conversion rate measures

How well a specific page turns its visitors into signups, leads, or purchases. Because it's scoped to one page, it isolates page quality from traffic quality — mostly. A page's rate still reflects who lands on it: a pricing page converts high-intent visitors a blog post never sees. Compare pages against their own history and against pages with similar intent, and segment by source via UTM parameters when a rate moves and you need to know whether the page changed or the traffic did.

And always read rate next to volume. Ranking pages by rate alone crowns low-traffic flukes; the pages worth optimizing are the ones where volume times rate-improvement is largest.

What data does it need?

  • A sessions table with landing_page, session_start, and a converted flag (or a join to a conversions table keyed by session).
  • UTM fields (utm_source, utm_medium, utm_campaign) captured on the session, so email and ad clicks credit their sources.
  • One conversion definition, owned in SQL — not one number from GA4, another from the CRM, and a third from the ad platform.

SQL patterns

Conversion rate by landing page (trailing 30 days, min 100 sessions)PostgreSQL
SELECT
  landing_page,
  COUNT(*) AS sessions,
  COUNT(*) FILTER (WHERE converted) AS conversions,
  ROUND(
    100.0 * COUNT(*) FILTER (WHERE converted) / COUNT(*), 2
  ) AS conversion_rate_pct
FROM sessions
WHERE session_start >= CURRENT_DATE - INTERVAL '30 days'
GROUP BY 1
HAVING COUNT(*) >= 100
ORDER BY conversions DESC;
Landing-page conversion by UTM source and campaignPostgreSQL
SELECT
  landing_page,
  utm_source,
  utm_campaign,
  COUNT(*) AS sessions,
  COUNT(*) FILTER (WHERE converted) AS conversions,
  ROUND(
    100.0 * COUNT(*) FILTER (WHERE converted) / COUNT(*), 2
  ) AS conversion_rate_pct
FROM sessions
WHERE session_start >= CURRENT_DATE - INTERVAL '30 days'
  AND utm_source IS NOT NULL
GROUP BY 1, 2, 3
HAVING COUNT(*) >= 50
ORDER BY sessions DESC;

Pitfalls

Ranking pages below sample-size thresholds.→ A page with 3 conversions from 15 sessions tops the leaderboard every time — and means nothing. Apply a HAVING floor before sorting by rate, and raise it when the base rate is low.
Using the sitewide rate as a KPI.→ It's a channel-mix artifact. More blog traffic drops it, more branded search lifts it, and neither says anything about the site. Keep it as context; optimize per page.
Ignoring intent differences between pages.→ A blog post converting at 0.5% may be outperforming; a pricing page at 3% may be underperforming. Compare pages within an intent tier — top-of-funnel content against content, high-intent pages against high-intent pages.
Letting the conversion definition drift between tools.→ GA4 counts a key event, the ad platform counts a pixel fire, the CRM counts a qualified lead — three "conversion rates" that never reconcile. Define the event once in your warehouse and compute every rate from it.

Where this metric applies

Metrics

Conversion rateClick-through rateCost per lead

Dashboards

Marketing funnelPaid vs. organic mix

FAQ

How do you calculate landing-page conversion rate?
Conversions divided by sessions, grouped by the landing page — the first page of the session, not every page the visitor touched. In SQL: COUNT(*) FILTER (WHERE converted) / COUNT(*) per landing_page. Define one conversion event, count it once per session, and hold that definition steady — a rate whose numerator quietly changed meaning is worse than no rate at all.
Why page-level instead of a sitewide conversion rate?
Because a sitewide rate is mostly a channel-mix report in disguise. Send more blog traffic and the sitewide number drops with nothing on the site getting worse; run a branded-search push and it rises with nothing getting better. Page-level rates isolate what each page actually does with its visitors. Keep the sitewide number as context on a marketing funnel dashboard, but diagnose and optimize at the page level with data from Google Analytics 4 or Plausible.
How many sessions before a page's rate is trustworthy?
Set a floor before you rank anything — 100 sessions is a reasonable default, higher if your base rate is low. At 30 sessions, one conversion swings the rate by three points; the top of an unthresholded leaderboard is always small-sample noise. A HAVING COUNT(*) >= 100 clause keeps flukes off the chart. For A/B-level decisions you need proper significance testing, but for monitoring, a minimum-sessions threshold removes most of the lying.
How do you credit email and ad traffic to landing pages?
Through UTM parameters. Tag every campaign link, capture utm_source, utm_medium, and utm_campaign on the session, and group conversion rate by landing page and source. That joins clicks from Mailchimp campaigns or paid ads to the sessions and conversions they produced — so you can see that the same page converts newsletter traffic at 8% and cold display traffic at 1%, which is a targeting insight, not a page problem. It's the same join that powers cost per lead by channel.
How do you track landing-page conversion rate in Metabase?
Sync sessions with landing page, UTM fields, and a conversion flag from GA4 (BigQuery export) or Plausible into a SQL database, then build a table of rate and volume per page with a minimum-sessions threshold. Chart rate and session count together — never rate alone — and pin it next to CTR on the marketing funnel dashboard so the click-to-conversion path reads in one place.