**To:** Freya Lindgren
**From:** Zara Ahmed
**Re:** Catalog health — first-week findings

**1. Catalog shape is driven by vertical, not merchant skill.** (*Merchant Catalog Profile*, *Catalog Structure: Top vs Tail*) Variant depth tracks category: Apparel runs ~12-14 variants/product, B2B (Office/Industrial) ~1-2. Image presence is near-universal everywhere (97-100%). The one lever with real spread is description completeness (57-86%). High confidence — large, consistent sample.

**2. "More images = more sales" oversimplifies.** (*Image-Count Bucket*, repeated within Apparel) The real jump is 0→1 image (Apparel: €109→€355/product); beyond that, 2-3/4-6/7+ bounce around rather than climbing steadily. I wouldn't tell merchants "add more photos," I'd tell them "don't ship with zero." Medium confidence — correlational and likely still confounded by price point.

**3. Review growth is just order growth.** (*Review Engagement*) Reviews-per-100-orders has stayed flat (~10-20%) for three years; rising raw review counts simply track rising order volume, not improving engagement. High confidence.

**4. December 2025 had a real rating dip** (3.71 avg, 25.8% 1-2-star — the low point of 36 months; *Monthly Trend*, *Dec Dip by Vertical/Merchant*). Drilling in: most of Home & Living's apparent decline is noise from 1-4 reviews per merchant, except two merchants with real volume — **Fennel & Flint** (14 reviews, 3.14 avg, -0.97 vs. baseline) and **Wiesgrund Wohnen** (12 reviews, 3.5, -0.6) — worth a direct check-in. Apparel & Footwear's decline is milder (-0.11) but carries half of all December reviews (219/454), so it matters in absolute terms. Medium confidence — single data point, needs a second month to confirm it's not one-off.

**5. Quality problems concentrate in specific merchants, including our #2 by revenue.** (*Poorly-Rated Product Concentration*) **Cragsley Outfitters**: 23 of 105 reviewed products (21.9%) average below 3.0 — a real tail, not isolated incidents. Studio Vrij (40%, n=10) and Brennquell Kaffee (13.1%, n=61) also stand out. Several of these bad-rated items still sell well (Brennquell's Aeropress kit: 128 units at 2.83; Cragsley's Corduroy Linen Shirt: 110 units at 2.86) — demand hasn't caught up to the reviews yet, but risk is building. High confidence for Cragsley given sample size.

**6. "High sales, no reviews" is mostly a B2B pattern, not a listing gap.** (*High-Selling, Few Reviews*) The list is dominated by Castell Fasteners, OfficeHub Nordic, Deskline Supplies — bulk/repeat buyers who simply don't leave reviews (e.g., a cordless drill: 248 units, €61,953, zero reviews). I'd caution against pushing review-solicitation campaigns here; it's buyer behavior, not neglect.

**7. Collections are saturated; bundles are white space.** (*Collections & Bundles Adoption*) Most merchants already put 65-100% of products in a collection (oddly many land on exactly 8 — possibly a template default, worth asking engineering). Bundles are essentially unused even by our top 6 merchants (0%); only small merchants reach 5-9%. I can't yet show a performance correlation — bundle usage is too rare to test honestly.

**8. Top 10 vs. the rest.** (*Catalog Structure*) The Top 10 (87% of revenue) have bigger catalogs (110 vs. 66 products) and much deeper variant assortment (8.2 vs. 4.8 variants/product), but categorization, image coverage, media richness, and collection usage are nearly identical across all tiers. "What good looks like" is about assortment depth and quality control at scale, not catalog mechanics — most merchants already do the basics.

**Caveats:** several merchant-level percentages rest on small review counts (single digits) — treat as directional. "Top 50 by revenue" cuts off ~70 tail merchants from most merchant-grain reports.

**Next steps:** watch January for a repeat of the Dec rating dip; pull return/ticket data for Cragsley's worst products; pilot bundles with 2-3 top merchants and measure AOV lift.