Allbirds — Shopify catalog pull

What was asked

Show what the system produces for a DTC / ecom product research brief: pick a store, pull the full catalog, summarize the price shape and category mix. No vendor data, no paid marketplace API.

What the system did

Every Shopify store exposes its entire catalog at <domain>/products.json — public, unauthenticated, paginated at up to 250 per page. The pipeline:

1. Hit allbirds.com/products.json?page=N&limit=250 until the store returned a short page (5 passes, 987 products total)
2. For each product, normalized: title, first-variant price, currency, product_type (the store’s own category), product handle URL, Shopify product id
3. Persisted raw Shopify JSON blobs into scraped_products.raw_json so every variant, image, and tag stays queryable after the fact

Total runtime: under 30 seconds, zero credentials, zero risk of getting blocked.

Headline numbers

• 987 products in the live Allbirds catalog
• Price band: $2 – $160
  • p10 $22 / p25 $49 / median $75 / p75 $105 / p90 $135
• Category mix: Shoes 813 · Apparel 95 · Socks 42 · Accessories 34 · Other 3
• Every product has a resolvable /products/<handle> URL → feeds direct attribution for any downstream tracking

What a pilot engagement gets on top of this

This sample stops at raw catalog data + a price shape. A real product-research retainer layers on:

1. Review ingest — pull public reviews (Amazon, Shopify review apps, third-party sites) per SKU and cluster pain points.
2. Cross-store benchmarking — scrape 5–10 comparable stores in the category and compute margin / price-position deltas.
3. Opportunity briefs — Claude-drafted, 1-page summary per SKU cluster: pricing gaps, recurring review themes, angle ideas a competitor could take to market.
4. Monitoring — weekly diff reports when new SKUs drop, prices move, or out-of-stock signals appear.

Why Shopify instead of Amazon / Lazada / Shopee

The same pipeline type supports Amazon and the PH marketplaces (Shopee, Lazada) — we have scrapers for those too. But those sources use heavy bot-detection and require proxy rotation + captcha handling for stable runs, so they’re not a good fit for a public demo you can reproduce. The Shopify /products.json endpoint is the cleanest possible illustration of the shape of data a product-research pilot starts from, and every number on this page can be verified by hitting the URL yourself.

Caveats

• /products.json exposes what the store wants public — locked SKUs, wholesale-only items, and some draft products are hidden. For a brand’s full catalog, a pilot pairs this with a sitemap crawl.
• price here is the first variant’s price. Multi-variant SKUs (size / color / bundle) aren’t broken out in this sample — they sit inside raw_json.variants[] in the database.
• category is the store’s own product_type field — consistent inside Allbirds, noisy across stores. Cross-store work uses a normalized taxonomy.