HorseTonic SEO/GEO analytics pull

Date range: 2026-06-23 to 2026-07-20 (last 28 days) unless stated otherwise Pulled: 2026-07-21 Sources: GA4 property 506939398 (G-7D3MRH0CH4) and PostHog EU project 191785 Status: GA4 is still blind for session attribution; PostHog is the only reliable traffic and funnel source. Live review: https://review-hub-black.vercel.app/horsetonic-seo-geo-2026-07-21/

1. Executive summary

GA4 captured only 3 sessions in 28 days, all channel Unassigned, yet recorded $1,208.85 revenue and 8 items purchased via server-side Measurement Protocol. The gtag fires, but sessions and acquisition are not stitching.
PostHog recorded 752 pageviews in the same period, with a clear funnel and strong DACH + Benelux geography. Purchase events are not reaching PostHog, so the bottom of the funnel is dark.
Organic search is the dominant referrer (Google 101 users / 241 views), followed by Instagram (50 users / 50 views).
English content is outperforming Dutch in raw views on the joint-supplement article, while the German locale is almost invisible (11 views).

One-line focus: fix measurement first, then push the English joint-supplement cluster into German-speaking Europe.

2. GA4 findings

Metric	Last 28 days	Note
Sessions	3	All `Unassigned`
Engaged sessions	0	Engagement rate 0%
Events per session	1.00	Minimal client data
Total revenue	$1,208.85	From server-side MP only
Items purchased	8	Ecommerce purchases report
Channel split	100% Unassigned	Attribution broken

Interpretation: the GA4 tag is present on the site, but client-side session initialization is still failing. Purchases are being reported because the Mollie webhook fires the Measurement Protocol hit independently of the client session. This makes GA4 unusable for traffic-source ROI or keyword-level SEO reporting.

Proof: browser pull of GA4 Traffic acquisition and Ecommerce purchases reports.

3. PostHog traffic sources

Referrer	Users	Views
`https://www.google.com/`	101	241
`https://instagram.com/`	50	50
`https://horsetonic.eu/en`	15	15
`http://m.facebook.com`	4	4
`http://m.facebook.com/`	3	11
`https://www.ecosia.org/`	3	5
`https://www.facebook.com/`	3	3
`https://www.bing.com/`	2	15
Other (ChatGPT, DuckDuckGo, Yahoo, etc.)	10	14

Takeaway: Google organic is the largest measurable acquisition channel. Instagram is second, but those sessions are likely from the paid social test (one URL carried utm_source=ig). Bing/Ecosia/DuckDuckGo show small but real search volume outside Google.

4. Top pages and products

Top pages by `$pageview` (28 days)

Page	Views
`/nl`	123
`/en`	76
`/nl/product`	51
`/en/kennisbank/best-joint-supplement-horses`	36
`/en/product`	28
`/en/product/artico-gold`	21
`/en/login`	20
`/en/contact`	20
`/nl/contact`	19
`/nl/login`	17

Top product-detail views (28 days)

Product page	Views
`/en/product/artico-gold`	21
`/en/product/tonic`	16
`/nl/product/artico-gold`	13
`/nl/product/tonic`	13
`/nl/product/lacta-ease`	11
`/nl/product/ph`	10
`/nl/product/elektro`	10
`/nl/product/zen`	7

Takeaway: the English joint-supplement article is the strongest content asset. Artico Gold and Tonic are the strongest products in both languages. The /en/login and /nl/login pages ranking high suggests informational traffic is landing on account pages; consider whether those pages should be noindexed.

5. Conversion funnel drop-off

Step	Events	Share of top step	Drop from previous
`$pageview`	752	100%	-
`product_viewed`	146	19.4%	-80.6%
`product_added_to_cart`	13	1.7%	-91.1%
`checkout_started`	7	0.9%	-46.2%
`payment_info_added`	3	0.4%	-57.1%
`purchase` / `purchase_completed`	0	0%	-100%

Key drop-offs:

Product view to add-to-cart (91% drop) is the largest leak. Stronger product-page social proof, clearer value bullets, and a sticky add-to-cart CTA should be tested.
Checkout to payment (57% drop) and add-to-cart to checkout (46% drop) indicate checkout friction. Shipping cost surprise or payment-method uncertainty are the usual culprits.
Purchase event is missing entirely in PostHog, so the final conversion cannot be tied back to traffic source or landing page.

6. Geography split

Top countries by `$pageview` users (28 days)

Country	Users	Views
Belgium	135	272
United States	86	104
Germany	62	77
The Netherlands	48	61
Austria	27	30
Russia	22	22
Sweden	21	25
Canada	14	14
Luxembourg	13	13
France	12	14

Device split

Device	Users	Views
Desktop	326	444
Mobile	178	304
Tablet	4	4

Locale split

Locale	Views
`nl`	364
`en`	308
`fr`	69
`de`	11

Takeaway: DACH + Benelux is the natural market, but the German locale (/de) is only 11 views in 28 days despite German traffic from Germany/Austria. The English site is pulling significant US/UK/CA traffic, which is unlikely to convert due to shipping/payment limits. Prioritize European localization over English global growth.

7. Focus recommendations

Immediate (this week)

Fix PostHog purchase tracking. The bedankt page already calls posthog.capture("purchase", ...), but no events arrive. Verify the key, the provider initialization on the thank-you page, and that the event is not blocked by consent/CSP.
Re-verify GA4 session attribution. The gtag config is in SSR HTML, yet sessions are not recorded. Test with a fresh browser, check analytics_storage consent state, and confirm gtag('config', ...) is not being overridden.

SEO/GEO focus (next 2-4 weeks)

English joint-supplement cluster is the highest-leverage content. Expand /en/kennisbank/best-joint-supplement-horses with comparison tables, FAQ schema, and internal links to /en/product/artico-gold and /en/product/artico-support.
German expansion. Translate the top English article and the Artico Gold/Tonic product pages into German, ensure /de hreflang is correct, and resubmit the DE sitemap.
Product-page CRO. Test a stronger add-to-cart section on Artico Gold and Tonic pages; these are the only products with meaningful view volume.
Checkout friction. Add shipping estimator and trust badges before checkout start; the 46-57% checkout drop is costing more than the product-page leak in absolute terms.

What to stop

Do not spend effort growing English global traffic until shipping and payment support those regions.
Do not rely on GA4 for SEO reporting until session attribution is fixed; use PostHog + GSC as the interim source of truth.

8. Data sources and proof

Live review / shareable URL: https://review-hub-black.vercel.app/horsetonic-seo-geo-2026-07-21/
GA4 property: 506939398, stream 12228395046, measurement ID G-7D3MRH0CH4
PostHog project: 191785, host https://eu.posthog.com
Raw data pulled via GA4 OAuth (browser) and PostHog Personal API key kimi-analytics-pull.