Which products actually make you money? Ask Claude, not your gut

Ask most store owners which products make them money and they'll name their bestsellers. But units sold isn't profit. The bestseller might be the one with the thinnest margin, the highest return rate, and the slowest restock — while a quiet product in the back is carrying the whole month. The WooCommerce reports tab shows you revenue. It doesn't tell you what to do.

Claude does — if you point it at the same data you already have. Not a new analytics suite to learn; a conversation with your own orders.

The questions that actually change decisions

Hook Claude up to your order and product data and these stop being guesses:

• Real winners — "rank products by revenue, then again by revenue minus returns. What moved?"
• Dead weight — "which products haven't sold in 60 days but still hold stock budget?"
• Repeat drivers — "what's the first purchase of my customers who came back and bought again?"
• Restock signal — "compare stock on hand to 30-day sales velocity — what runs out in two weeks?"
• Discount math — "which slow movers have enough margin to discount 20% and still profit?"

Each of those is a decision — push, kill, restock, discount — backed by your numbers instead of your mood.

How to set it up

You don't need to export anything if you connect the store directly; you do if you'd rather keep it read-only and simple.

Direct (live): connect your store over MCP (read-only key is enough for analytics) and just ask. Setup is the same connector as the WooCommerce MCP tutorial — once it's wired, Claude reads orders, products and stock on demand.

Export (simple, zero write access): pull your orders via the WooCommerce REST API and hand the file to Claude.

import requests, csv
r = requests.get("https://yourstore.com/wp-json/wc/v3/orders",
                 params={"per_page": 100, "status": "completed"},
                 auth=("ck_…", "cs_…"))           # read-only REST key
orders = r.json()
with open("orders.csv", "w", newline="") as f:
    w = csv.writer(f); w.writerow(["id","total","date","items"])
    for o in orders:
        items = "; ".join(f"{li['name']} x{li['quantity']}" for li in o["line_items"])
        w.writerow([o["id"], o["total"], o["date_created"], items])

Drop orders.csv into Claude and ask the questions above in plain language. It'll do the grouping, the margins (give it your cost data), and the ranking — and explain the "why," not just the table.

Why it matters

You already paid for this data — it's sitting in your store, mostly unread, because real analysis meant a spreadsheet afternoon nobody books. Claude collapses that afternoon into a question. The owners who win in 2026 aren't the ones with more data; everyone has the data. They're the ones who actually ask it before deciding what to stock, push and cut. Stop deciding by gut on numbers you could just read.