// REPORT_ID: MENFIT-2026
// DATA THRU: 27 Oct 2026
// PREP: PROPAGANDA SOLUTIONS

// Performance Dossier · Wednesday 28 Oct 2026 · Confidential

MENFIT — Week by Week · After the Restart

Last week (Oct 21–27, the first full week since the 21 Oct restart) vs the failed week (Oct 14–20) and the pre-failure week (Oct 7–13) · October 1–27 vs September 1–27 · day by day · week by week · complete months · Meta ad-level read

Scope & definitions: MENFIT web shop only (Shopware · menfit.com); excludes marketplace and retail/wholesale. Currency EUR. Revenue = Shopware amountTotal (gross, incl. VAT and shipping), all order states counted, cancelled/unpaid stated where present; no refunds recorded since 1 Sep. All windows are Europe/Berlin calendar days, start-inclusive and end-exclusive ("Oct 21–27" = 21 Oct 00:00 to 28 Oct 00:00); Shopware stores UTC and is converted explicitly. GA4 property and Meta account both report in Europe/Berlin. "Revenue ÷ spend" is webshop revenue divided by Meta spend on the same days — a blended ratio, not attributed ROAS and not profit.

Delivery held all week: €648.37 on Oct 21–27, €95 → €113 → €113 → €112 → €81 → €64 → €71 a day. Orders: 8 last week (7 paid, €475.90; one Klarna order cancelled at payment), €614.86 gross, 6 new customers — the same order count as the pre-failure week (8), but on €648 instead of €399 of spend: revenue ÷ spend 0.95 gross, 0.73 on paid orders (pre-failure 1.15). All 8 orders fell on Thu 22 – Sun 25 Oct; Mon 26, Tue 27 and this morning have none. The delivering ad set is optimising for add-to-cart, not purchase (224 add-to-carts, 2 pixel purchases), and the purchase pixel event is reporting an error (found 28 Oct 07:45) — that is the next fix. October 1–27: €1,343 on 20 orders vs €1,654 on 20 for Sep 1–27 (-19% revenue, orders level) after seven dark days.

Sources. Shopware Admin API extracted Wed 28 Oct 07:34 CET (order-level window fetched with total-count verification, 8 orders = 8 distinct ids, no duplicates; latest order #21471 at 25 Oct 15:06; no order since); GA4 Data API extracted 28 Oct 07:36 CET, daily through 27 Oct plus 28 Oct intraday; Meta read from Ads Manager on 28 Oct 08:40 CET, campaign / ad set / ad level with a day breakdown (the Madgicx API token was invalid this morning, so no API read; account currency EUR, timezone Europe/Berlin): September €1,641.87, Sep 1–27 €1,479.04, Oct 1–27 €1,386.29 (€737.92 on Oct 1–13, €0.00 on Oct 14–20, €648.37 on Oct 21–27); 28 Oct intraday not read; Google Ads spend not readable (no API access; GA4 shows google/cpc sessions only). ROAS = Shopware revenue ÷ Meta spend on matched dates. Two measurement notes. (1) The Meta pixel recorded 3 purchases in Oct 1–27 against 20 real orders; since the restart the delivering ad set optimises for add-to-cart (Ads Manager result column), and Sabina reported on 28 Oct that the purchase event itself is throwing an error — Meta is optimising blind. (2) GA4 recorded 336 Paid Social sessions in Oct 1–27 while Meta counted 2,894 landing-page views — GA4 is seeing roughly one in nine Meta visitors (cookie consent on mobile is the likely reason). GA4 session counts are therefore a sample, and the conversion rates shown are higher than the true rate on all visitors. The trend across months is still valid (the sampling was the same in June); the absolute level is not. Shopware is the truth for orders and revenue.

At a glance — last week (Oct 21–27) vs the pre-failure week (Oct 7–13) · the failed week Oct 14–20 was zero on every source

Revenue (Shopware)

€615

8 orders (7 paid, €475.90; 1 Klarna cancelled) · Oct 7–13: €459 · Oct 14–20: €0

▲ +34% · vs Oct 7–13

Orders

Oct 7–13: 8 · Oct 14–20: 0 · AOV €76.86 · all on Thu 22 – Sun 25

▲ +0% · vs Oct 7–13

New customers

+2 returning · Oct 7–13: 5 new + 3 returning · Oct 1–27: 14 new

▲ +20% · vs Oct 7–13

October 1–27 vs Sep 1–27

€1,343

20 vs 20 orders · €1,654 in Sep (19 paid, €1,560.54) · Oct paid €1,204.43 (19)

▼ -19% · vs Sep 1–27

Meta spend last week

€648.37

revenue ÷ spend 0.95 gross / 0.73 paid · pre-failure week €399 at 1.15 · Oct 1–27 €1,386 (0.97 gross, 0.87 paid)

Sessions (GA4 sample)

310

failed week 9 · pre-failure week 59 · consent-gated sample; GA4 recorded 6 of 7 paid orders

▲ +425% · vs Oct 7–13

▸ Day by day since the restart, and October to date

Day (Berlin)	Meta spend	Impr.	Link clicks	LPV	Add-to-cart (pixel)	Checkout start	Pixel purch.	Shopware orders	Gross	Paid	GA4 sess.*
Wed 14 – Tue 20 Oct	€0.00	0	0	1	0	0	0	0	€0.00	€0.00	9
Wed 21 Oct (from 11:07)	€95.16	4,994	327	227	18	2	0	0	€0.00	€0.00	14
Thu 22 Oct	€112.76	6,471	429	382	42	9	0	1 cancelled	€138.96	€0.00	39
Fri 23 Oct	€112.95	7,189	480	429	33	6	1	4	€243.21	€243.21	54
Sat 24 Oct	€111.89	6,188	454	406	48	6	1	2	€172.81	€172.81	51
Sun 25 Oct	€81.34	5,607	439	396	36	0	0	1	€59.88	€59.88	39
Mon 26 Oct	€63.69	4,099	309	286	27	2	0	0	€0.00	€0.00	61
Tue 27 Oct	€70.58	4,094	348	316	20	1	0	0	€0.00	€0.00	52
Wed 28 Oct · intraday to 07:35 provisional	n/r	n/r	n/r	n/r	n/r	n/r	n/r	0	€0.00	€0.00	9
Week Oct 21–27	€648.37	38,640	2,786	2,442	224	26	2	8 (7 paid)	€614.86	€475.90	310

Last week in one line. 8 orders (7 paid, €475.90; order #21464 cancelled at Klarna payment), €615 gross, 6 new + 2 returning customers, AOV €76.86, on €648.37 Meta → revenue ÷ spend 0.95 gross / 0.73 paid; cost per paid order €92.62 (pre-failure week: €49.91). Product mix (line value = totals apart from €4.90 shipping on one order): Power 3-month ×3 €208.44 (one of them cancelled), Power 1-month ×2 €57.90, Protection 3-month ×2 €119.76, Routine bundle €59.95, Performance 1-month €24.95. GA4 recorded 6 of the 7 paid orders: 5 booked to Direct (€267.46) and 1 to Paid Search (€69.48, a Google click on 23 Oct); 0 to Paid Social despite 263 Paid Social sessions — Meta traffic buys later, as Direct.

The shape of the week. Spend ran at ~€112/day for four days, then Meta throttled to €64–81/day from Sunday (campaign budget reads €65/day; the first days over-delivered against it). Orders came in a four-day cluster (Thu–Sun, 8 orders) and stopped: Mon 26, Tue 27 and this morning have none, although landing-page views (286 and 316) and add-to-carts (27 and 20) were normal. Checkout starts fell from 6–9 a day to 0–2 from Sunday. One week is not a trend, but the pattern — plenty of carts, few checkouts, no purchases — matches an ad set that is paid to find cart-adders, not buyers.

What is set up (Ads Manager, 28 Oct 08:40). Delivering ad set "Static – ATC – Copy": Active, optimisation = website add-to-cart (Results column: 224 add-to-carts at €2.89), uses the campaign budget (€65/day), 5 ads; it carried 100% of spend since 21 Oct. "Purchase Ad set" (Advantage+, created 21 Oct 11:19, optimised for purchase): still Learning, €0.01, 2 impressions — not delivering, same symptom as the original "Static – Purchase" set that stopped on 13 Oct. Sabina, 28 Oct 07:45: a pixel error on the purchase event is the likely reason the purchase-optimised sets cannot deliver; Prosper is checking the shop code. Until the purchase event fires, the only ad set Meta will serve is the add-to-cart one.

October vs September to date (same 27 days). Oct 1–27 = €1,343 / 20 orders (19 paid, €1,204.43) / 14 new / AOV €67.17, against Sep 1–27 = €1,654 / 20 orders (19 paid, €1,560.54) / 12 new / AOV €82.71 — revenue -19%, orders +0%, AOV −19% (the 1 Oct price cuts and smaller packs). Meta Oct 1–27 €1,386 vs €1,479 → revenue ÷ spend 0.97 vs 1.12. Without the seven dark days October would be tracking ahead of September on orders; with them it finishes roughly level on orders and below on revenue unless the last four days produce 6+ orders (September's last four days: €143.42, 4 orders).

The campaign — ad-level read, last week (Oct 21–27)

Ad	Spend	Share	Impr.	Link clicks	CTR	LPV	€ / LPV	Add to cart	Init. checkout	Pixel purchases
New Ad 11 – Copy active · ATC-Copy set	€459.62	71%	28,850	2218	7.69%	1969	€0.23	165	17	1
New Ad 10 – Copy + New Ad 14 – Copy (combined) active · ATC-Copy set	€147.79	23%	7,918	474	5.99%	399	€0.37	51	5	1
New Ad 6 Video – Copy active · ATC-Copy set	€29.25	5%	1,252	75	5.99%	60	€0.49	5	2	0
New Ad 13 – Copy active · ATC-Copy set	€11.70	2%	620	19	3.06%	14	€0.84	3	2	0

One campaign ("Libido steigern", objective Sales). Delivering ad set: "Static – ATC – Copy", optimised for add-to-cart (€648.36 of €648.37 since the restart); the new "Purchase Ad set" has 2 impressions in a week and is not delivering. "Share" is share of spend, not of sales. Green = produced a pixel purchase; red = >€100 spent, no pixel purchase. "New Ad 11 – Copy" takes 71% of spend and delivers landing-page views at €0.23; Ads 10 and 14 are shown combined (Ads Manager rendered their totals but not the individual rows in this read). Pixel purchases (2) understate real orders (7 paid). Omitted: Purchase Ad set €0.01.

▸ Status of the open actions

1 · Delivery — restored 21 Oct 11:07 and stable for a week (€648 spent, 38,640 impressions, 2,442 landing-page views at €0.27 each vs €1.64 before the stop). The cost of the dark week was ~€400 of spend not placed and ~8 orders not won; last week recovered the order count, not yet the efficiency.

2 · Purchase signal — the open problem. The delivering ad set optimises for add-to-cart; the purchase-optimised set cannot leave Learning; the pixel's purchase event is erroring (Sabina/Prosper, 28 Oct). Pixel saw 2 purchases vs 7 paid orders last week; 3 vs 19 paid in October. Fix order: repair the purchase event in the shop code → confirm in Events Manager → let the purchase-optimised set deliver → Conversions API from Shopware so this cannot silently break again.

3 · Audience. The old audience is being re-served (3.5–3.9 frequency before the stop); the Advantage+ set has not served a single impression beyond two. Add-to-cart rate is healthy (9% of landing-page views) but checkout starts collapsed from Sunday — watch Wed–Thu; if purchases do not return with the pixel fix, the audience is the next suspect.

4 · Google / PMax. Reactivation agreed 21 Oct; GA4 shows one Google-click purchase (€69.48, 23 Oct) and 3 Paid Search sessions last week — consistent with Google serving again at a low level. Spend still not readable (API access 6964220345 not granted). 5 · OpenAI / ChatGPT Ads: account ready since 14 Oct, pixel blocked on GTM publish rights (Prosper and Thorsten have user-only access to GTM-P3XVCD92) — not live. 6 · Protection with Michael; Power sells remaining stock ahead of Power 2.0.

Week by week — four completed Wed–Tue windows (to 27 Oct)

Week	Sessions*	Orders	Revenue	Conv.*	AOV	New	Meta spend	ROAS
Sep 30 – Oct 6	62	7	€390	11.3%	€55.66	6	€398	0.98
Oct 7–13 pre-failure	59	8▲+14%	€459▲+18%	13.6%	€57.34	5	€399	1.15
Oct 14–20 no delivery	9	0▼-100%	€0▼-100%	0.0%	—	0	€0	n/a
Oct 21–27 last week · restart	310	8	€615	2.6%	€76.86	6	€648	0.95

*GA4 sessions are a consent-gated sample (see measurement note). Orders 7 → 8 → 0 → 8; revenue €390 → €459 → €0 → €615; Meta €398 → €399 → €0.00 → €648. Revenue ÷ spend 0.98 → 1.15 → n/a → 0.95 gross (0.73 paid-only). Last week bought the same 8 orders as the pre-failure week for 62% more spend; the add-to-cart optimisation is the obvious reason.

Orders per Wed–Tue window.

Revenue per sampled session by month — the trend is real even though the level is inflated by GA4 sampling.

Complete calendar months — June to September (October is in progress)

Month	Revenue	Orders	AOV	New	Sessions*	Conv.*	€ / session*	Power %	Meta spend	ROAS
June	€1,852	25	€74.09	18	717	3.5%	€2.58	60%	€1,079	1.72×
July	€2,012	24	€83.82	12	707	3.4%	€2.85	68%	€1,567	1.28×
August	€1,717	21	€81.78	16	490	4.3%	€3.50	94%	€1,728	0.99×
September	€1,798	24	€74.90	16	267	9.0%	€6.73	49%	€1,642	1.09×

September: €1,798 / 24 orders — +5% revenue, +14% orders vs August, the first month to grow since May; ROAS 1.09× (Aug 0.99×). Second half 17 orders / €1,313 after guest checkout (11 Sep). October to the 27th: €1,343 / 20 orders, Meta €1,386, revenue ÷ spend 0.97 — level with September on orders, -19% on revenue after seven dark days and the October price cuts; four days left.

Insights — what the numbers are telling us

The restart worked as a restart. Seven days of delivery, €648 spent, 8 orders — the pre-failure order rate is back. What is not back is efficiency: 0.95 gross / 0.73 paid revenue ÷ spend against 1.15, and €92.62 per paid order against €49.91.
The reason is structural, not creative. The only ad set that delivers is optimised for add-to-cart. Meta did exactly what it was asked: 224 add-to-carts (9% of landing-page views, double the pre-failure rate) and 26 checkout starts — but checkouts and purchases do not follow carts at the old rate. The purchase-optimised set cannot start because the purchase event is broken.
Orders came in one cluster. Thu 22 – Sun 25 Oct: 8 orders; Mon 26 – Wed 28 morning: none, on normal traffic. Three empty days after a week of normal delivery is the thing to watch today and tomorrow, not yet a conclusion.
Traffic is cheap now. Landing-page views at €0.27 (pre-failure €1.64), CTR 7.2%, CPM €16.78. Whether cheap cart-adders are the right people is what the purchase signal will tell us once it fires.
October 1–27 vs September 1–27: orders level (20 vs 20), revenue -19% (AOV €67 vs €83 after the price cuts), spend €1,386 vs €1,479, revenue ÷ spend 0.97 vs 1.12. Through 13 Oct October was +86% ahead; the dark week and the cut prices took that back.
Measurement is the bottleneck on every side: pixel purchase event erroring; GA4 sampling ~1 in 9 Meta visitors and booking Meta buyers as Direct; Google Ads spend unreadable; ChatGPT Ads pixel blocked on GTM publish rights. Every one of these is an access or code fix, not a media decision.
Klarna: second Klarna order cancelled at payment since September (#21464, 22 Oct); 2 of 44 orders since 1 Sep. Small, but worth a look at the Klarna step.

Recommendations — this week, in order

1 · Fix the purchase event first (shop code, Prosper; verify in Events Manager with a test order), then let the purchase-optimised set take over. Until it fires, Meta is buying carts, not customers.
2 · Conversions API from Shopware right behind it, so the purchase signal no longer depends on the browser pixel and consent. This has been the single biggest lever since 12 Oct.
3 · Keep the add-to-cart set running meanwhile at the current budget — it is the only delivery there is — and do not touch its ads this week; judge Oct 28 – Nov 3 on Shopware paid orders per euro.
4 · GTM publish rights for PS (Michael's side): unblocks the GA4 consent fix, Hotjar, and the ChatGPT Ads pixel in one go.
5 · Google Ads API access for 6964220345 so PMax/Search spend can be costed against the Google-click orders GA4 now records.
6 · Daily 09:00 delivery check stays in place (green/amber/red, Sabina) — the dark week must not repeat.

// SOURCES · SALES: Shopware 6 Admin API (menfit.com), extracted 28 Oct 07:34 CET, complete days through 27 Oct 23:59 Europe/Berlin (gross amountTotal incl. VAT + shipping; all order states, paid-only stated; no refunds since 1 Sep; pagination and duplicates verified against total-count) · TRAFFIC: GA4 property 518453521 (Europe/Berlin), extracted 28 Oct 07:36 CET, daily through 27 Oct + 28 Oct intraday — consent-gated sample, purchase events incomplete · META: act_257999488596895 via Madgicx API (account / ad set / ad level) — Meta read from Ads Manager on 28 Oct 08:40 CET, campaign / ad set / ad level with a day breakdown (the Madgicx API token was invalid this morning, so no API read; account currency EUR, timezone Europe/Berlin): September €1,641.87, Sep 1–27 €1,479.04, Oct 1–27 €1,386.29 (€737.92 on Oct 1–13, €0.00 on Oct 14–20, €648.37 on Oct 21–27); 28 Oct intraday not read · GOOGLE ADS: not available (API access not granted); GA4-sampled Google sessions only
// WINDOWS (Europe/Berlin, start-inclusive, end-exclusive; Shopware UTC converted explicitly) · Last week = Oct 21–27 vs Oct 14–20 (zero) and Oct 7–13 (pre-failure) · October = Oct 1–27 vs Sep 1–27 · Weeks = Wed–Tue windows Sep 30 – Oct 27 · Months = complete calendar months Jun–Sep · Apr–Jun sales validated vs known figures · META via Ads Manager day breakdown (API token invalid 28 Oct) · CANONICAL MX²/BigQuery not verified this run
// Generated 2026-10-28 · Propaganda Solutions · updates auto-deploy to menfit.mx2.ai