CRO & Landing Pages
by
A conversion review for ecommerce teams: where the funnel leaks, whether the drop comes from your traffic or your site, and which campaigns send spend to pages that convert worse, with every fix priced in euros.
When conversion rate drops, most teams open the checkout and start guessing. Half the time the site didn't change: the traffic did. This skill reads your Polar pixel sessions and answers in the right order. It first splits the change into mix (new traffic that converts less) and rate (the same traffic converting worse), then locates the step that broke by device and country, and only then looks at landing pages: which campaigns land on pages that convert below the best alternative, which pages simply aren't live yet, and what each fix is worth in lost orders times your AOV.
What It Does
- Reads the funnel step by step, from session to add to cart to checkout to order, against the previous period.
- Splits any conversion change into mix effect and rate effect, so you know if it's the traffic or the site.
- Locates the step that broke by device and country, down to a missing payment method.
- Matches campaigns to landing pages and flags swaps only when both pages have enough sessions.
- Tells a broken page from a seasonal or pre-launch one, and an audience problem from a page problem.
- Prices every fix: lost orders times your AOV, per 28 days.
Use Cases
- Explain a conversion drop to the team in one message instead of a week of theories.
- Find the country where checkout completion fell and the payment step behind it.
- Check which campaigns send traffic to a weaker landing page than the best one.
- Build a CRO test backlog ranked by euros, not by opinions.
Example Session
/polar-cro-landing-pages Why did conversion drop over the last 28 days, and which landing pages should we fix?
CRO review · last 28 days vs previous 28 days
Source: Polar pixel sessions · 28 full days ending yesterday · AOV €96 · read-only
• Conversion fell from 3.4% to 2.8% while sessions grew 35%. It's a rate effect: every major channel converts worse, not just the new traffic.
• The step that broke is checkout completion, down 11 points on mobile in France and the UK.
• About €24k in lost orders over 28 days. Four fixes below, sorted by value.
- Finding | Where | Verdict | Now | Reference | € / 28 days
- Checkout completion | Mobile · FR + UK | Fix | 58% | 69% | +€9.8k
- Add to cart | All devices | Test | 6.9% | 8.1% | +€7.2k
- Landing page | Prospecting · PDP A vs B | Test | 2.7% | 3.6% | +€1.9k
- Homepage | Brand broad vs exact | Audience | 2.9% | 5.3% | n/a
- Gift set PDP | Organic + email | Pre-launch | 1.0% | n/a | n/a
Same homepage, two brand campaigns: exact converts 5.3%, broad 2.9%. The page is fine; the broad campaign brings a weaker audience, so the fix belongs in Google Ads, not in the CMS.
The gift set page converts at 1.0% because it isn't on sale yet. Flagged pre-launch, not broken.
Assumptions: CVR = completed checkouts ÷ sessions · swaps need 200+ sessions on both pages · lost orders × period AOV · nothing changed on the site or in ads.
Install It
- Copy the command.
One line in your terminal. Free, no account needed. - Run it.
npx skills add pulls the skill from Polar's GitHub and installs it for the agents you pick: Claude Code, Cursor, Codex and more. - Connect Polar and ask.
Connect the Polar connector, then ask why conversion moved. The skill does the rest.
skills folder
Copy
npx skills add \
https://github.com/polar-analytics/claude-skills \
--skill polar-cro-landing-pagesFrequently asked questions
How is this different from a GA4 funnel report?
GA4 shows you the funnel. This skill explains it: it separates traffic mix from conversion rate, finds the step and segment that broke, and prices each fix in euros, so you start with the problem worth the most.
What data does it need?
Session-level data from the Polar pixel or GA4, plus your store connector for orders and AOV. Without session data the funnel can't be measured, and the skill says so instead of guessing.
Does it change anything on my site or in my ads?
No. It reads your data through Polar's MCP and recommends. Swaps, fixes and tests are yours to ship.
When does it recommend swapping a landing page?
Only when both pages have at least 200 sessions from comparable traffic and the current one converts below 70% of the best. Below that, it proposes a test, not a swap.
What is the mix effect?
When new traffic converts less than your usual traffic, overall conversion drops even if the site hasn't changed. The skill measures how much of the change comes from that mix and how much from a real conversion loss.
Why are some pages labelled pre-launch?
A seasonal or upcoming product page gets visits before it's on sale, so its conversion rate is low by design. The skill flags it instead of sending you to redesign a page that isn't live yet.
A funnel review is worth what your session data is worth.
Polar's pixel follows every session from landing page to order and ties it to the campaign that sent it, so the skill can tell a traffic problem from a site problem.
<script type="application/ld+json">{"@context": "https://schema.org", "@type": "FAQPage", "mainEntity": [{"@type": "Question", "name": "How is this different from a GA4 funnel report?", "acceptedAnswer": {"@type": "Answer", "text": "GA4 shows you the funnel. This skill explains it: it separates traffic mix from conversion rate, finds the step and segment that broke, and prices each fix in euros, so you start with the problem worth the most."}}, {"@type": "Question", "name": "What data does it need?", "acceptedAnswer": {"@type": "Answer", "text": "Session-level data from the Polar pixel or GA4, plus your store connector for orders and AOV. Without session data the funnel can't be measured, and the skill says so instead of guessing."}}, {"@type": "Question", "name": "Does it change anything on my site or in my ads?", "acceptedAnswer": {"@type": "Answer", "text": "No. It reads your data through Polar's MCP and recommends. Swaps, fixes and tests are yours to ship."}}, {"@type": "Question", "name": "When does it recommend swapping a landing page?", "acceptedAnswer": {"@type": "Answer", "text": "Only when both pages have at least 200 sessions from comparable traffic and the current one converts below 70% of the best. Below that, it proposes a test, not a swap."}}, {"@type": "Question", "name": "What is the mix effect?", "acceptedAnswer": {"@type": "Answer", "text": "When new traffic converts less than your usual traffic, overall conversion drops even if the site hasn't changed. The skill measures how much of the change comes from that mix and how much from a real conversion loss."}}, {"@type": "Question", "name": "Why are some pages labelled pre-launch?", "acceptedAnswer": {"@type": "Answer", "text": "A seasonal or upcoming product page gets visits before it's on sale, so its conversion rate is low by design. The skill flags it instead of sending you to redesign a page that isn't live yet."}}]}</script>