Product Analysis — Query Workflow

Product Analysis — Query Workflow

Four-step workflow for full analysis. Field names are role-based — already confirmed during setup.

---

Step 1 — Broad overview (fire all four in one turn)

Query	Tool	Group by	Measures	Limit	Order
Products by views	Session search	Viewed product titles	Sessions, CVR, revenue/session	20	Sessions ↓
Product ATC opportunities	Page visits	URL (filter: CONTAINS '/products/')	Page views, ATC_LIFT_OPPORTUNITY, add-to-cart rate	10	ATC_LIFT_OPPORTUNITY ↓
Collections by conversion	Session search	Viewed collection titles	Sessions, CVR, revenue/session	15	Sessions ↓
Product types by conversion	Session search	Viewed product types	Sessions, CVR, revenue/session	10	Sessions ↓

Product ATC opportunities: ATC_LIFT_OPPORTUNITY is a pre-computed score combining traffic volume and ATC rate gap — higher = more opportunity. Use this ranking directly to identify low-ATC product candidates.

Products by purchase is not in Step 1. Only run it in Step 3 if the data suggests cart abandonment (a product with good ATC rates but weak purchase completion).

---

Step 2 — Cross-reference in post-processing

• ATC opportunity = use the ATC_LIFT_OPPORTUNITY ranking directly. Cross-check the top URLs against products-by-views to confirm they are high-traffic products.
• Collection CVR benchmark = compare collections against others in the same price tier, not the site-wide average. Two game-focused collections are valid benchmarks for each other; a premium hardware collection is not a valid benchmark for a accessories collection. Only flag low CVR when a same-tier benchmark exists.
• Viewed-only % = sessions with no funnel progression ÷ total view sessions. Flag products where 80%+ viewed-only, but check site-wide average first — high viewed-only is normal for low-intent browsing.
• Site-wide CVR benchmark = total purchases ÷ total sessions. Use only as context — never to flag high-ticket products, collections, or types as underperforming. Price data is not in Noibu session data; infer price tier from product/collection names.

---

Step 3 — Pick 2–4 signals to dig into

If you see this…	Follow-up
High ATC_LIFT_OPPORTUNITY product in top-20 views	Page deep-dive: scroll depth, time on page, click engagement, errors
Product with good ATC rate but weak purchase completion suspected	Run "Products by purchase" now (session search, purchased product titles, sessions + median cart value, limit 20, sessions ↓) — then funnel depth breakdown for that product
Collection CVR well below a comparable benchmark collection	Country breakdown: near-zero CVR across LATAM/non-primary = localization/checkout gap
Collection CVR below others in same category	Product mix drill-down within that collection
Product type significantly over/under	Break down by product title within the type
High-purchase product absent from views top	Journey path analysis — likely reached via search/direct links
Collections with non-English names at very low CVR	Check checkout/shipping availability in those markets

---

Step 4 — Deeper follow-up queries

Run only the follow-ups the signals call for. Apply traffic thresholds now:

• 500K sessions/month: threshold ~0.1–0.2% of total

• 50K–500K: ~0.3–0.5%
• <50K: very low or skip

Product page deep-dive (high ATC_LIFT_OPPORTUNITY)

• Use page visits tool. Filter URLs CONTAINS the product's SKU code or slug fragment (limit 15).
• Measures: page views, ATC_LIFT_OPPORTUNITY, add-to-cart rate, median duration, median max scroll depth ratio, median clicked selector count, total visual error count. Order by ATC_LIFT_OPPORTUNITY ↓.
• Use URL CONTAINS (not exact URL) — product pages appear under /products/, /collections/[name]/products/, /es/products/, etc.
• Scroll depth: <0.20 = users not reaching ATC button; 0.20–0.40 with low clicks = content/pricing concern; high errors on one URL variant = possible JS error blocking ATC.
• Compare ATC rate across URL variants — a rate that's low on one variant but normal on others points to a URL-specific issue (broken localised page, collection-scoped page with missing ATC).

Funnel depth breakdown (strong ATC, weak purchase)

• Session search tool. Filter to sessions where target product title was viewed. Group by funnel depth field. Measure session count. Order ↓.
• Depth values: null = viewed only, 1 = ATC, 2 = checkout started, 3 = payment submitted, 4 = completed.
• Lead with "Viewed only %" = null sessions ÷ total.
• Note: depth-4 outnumbering depth-2/3 is normal (Apple Pay, Shop Pay skip checkout pages — not a data error).

Country breakdown (collection CVR well below benchmark)

• Session search tool. Filter to sessions where collection title was viewed. Group by country code (limit 20). Measure sessions, CVR, revenue/session. Order ↓.
• Near-zero CVR across multiple LATAM/non-English markets = ops/localization gap, not a merchandising problem.

Product mix within a collection (collection underperforms, country breakdown is clean)

• Session search tool. Filter to collection title viewed. Group by viewed product titles (array join, limit 25). Measure sessions, CVR. Order ↓.

Journey paths (high views, high bounce)

• User journey tool. Anchor on product slug fragment, loose mode. Both directions, max depth 6.