Connect your Shopify store to Claude and stop guessing why last month dipped

You have access to my Shopify store through the connector. Act as a commercial analyst who has run a small e-commerce business, not a report generator.

Pull the data yourself. Where a number is unavailable, say so rather than estimating.

Compare the last 30 days to the 30 days before it, and to the same 30 days last year if that data exists. Then work through this:

1. WHAT MOVED. The 5 metrics that changed most, with the actual numbers and the percentage change. Ignore anything that moved less than 5%, it's noise.

2. THE CAUSE CHAIN. For the biggest change, trace it back as far as the data allows. If revenue dropped, was it traffic, conversion rate, average order value, or repeat rate? Then go one level deeper on whichever one it was. Keep going until you hit either a real cause or the limit of what the data can tell you. Say clearly which one you hit.

3. WHAT'S QUIETLY WORKING. Find the product, collection, customer segment or traffic source that's growing but too small for me to have noticed. Tell me what it's growing at and what it would be worth if I gave it real attention.

4. WHAT'S QUIETLY LEAKING. Products with high views and low conversion. Customers who bought twice and then stopped. Stock sitting still. Anything where money is already on the table and falling off it.

5. THE THREE QUESTIONS I SHOULD BE ASKING. Based on what you found, what am I not curious about that I should be? These should be questions I can answer with the next month of data.

Rules:
- Show me the numbers behind every claim. If you can't show a number, label it a hypothesis.
- Do not recommend actions in this response. I want the diagnosis first, uncontaminated by suggestions.
- Where two explanations fit the same data, give me both and tell me which evidence would separate them.
- Australian English. Plain language. No dashboards, no bullet soup, write it like you're explaining it to me over a coffee.

Pull everything you can on [product name] for the last 90 days: units, revenue, views, conversion rate, returns, and which other products it gets bought alongside. Then tell me whether this product has a traffic problem, a listing problem, a price problem or a demand problem. Be specific about which evidence points where, and tell me what you'd need to see to be sure.

Using my order data, segment my customers by purchase behaviour, not by demographics. Show me: one-time buyers, people who bought twice within 90 days, and anyone who's bought 3 or more times. For each group, tell me what their first purchase usually was and their average lifetime value. Then tell me which first purchase most reliably predicts a repeat customer, and how confident you are given the sample size.

I'm deciding whether to discontinue [product]. Pull its full performance history, its stock position, and what it contributes beyond direct revenue (bundles, first purchases, repeat triggers). Lay out the case for keeping it and the case for cutting it, with numbers on both sides. Tell me what I'd be giving up that isn't obvious from the sales figure. Do not tell me what to do, I'm making this call.