> For the complete documentation index, see [llms.txt](https://docs.reportpundit.com/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://docs.reportpundit.com/knowledge-base/settings/mcp-connector/questions-to-ask-claude-about-your-shopify-reports.md).

# Questions to ask Claude about your Shopify reports

Once a report is in your conversation, the quality of the answer depends on the question. This page is a library of questions that work, organised by the kind of report you already have.

Find your report family below, copy a question, and change the details to match your store. Everything here assumes Report Pundit is connected. If it is not, see [How to connect Report Pundit to Claude](https://claude.ai/connect-report-pundit-to-claude).

### How to phrase a question that works

Three things make the difference between a vague answer and a useful one.

**Name the report.** "Run my Sales by Product report and..." beats "look at my sales." Claude then knows exactly which rows it is working from.

**Name the period.** Reports run on their own saved date range unless you say otherwise. If the period matters, say it. See [How to run a report for a different date range in Claude](https://claude.ai/change-date-range-in-claude).

**Name the level of detail.** By product, by variant, by day, by order, by country. Without this, Claude picks a level, and it may not be yours.

You do not need to write a perfect prompt. Ask, look at the answer, then refine. Follow-up questions are answered from the data already in the conversation, so they are instant.

### Sales and orders

*Assumes a Sales or Order report with revenue, quantity and dates.*

> Which products drove the revenue change between this month and last month?

> What share of my revenue comes from my top 10 products, and how concentrated is that?

> Which days of the week and times of day produce the most orders?

> Which discount codes were used most, and what did each one cost me in discount value against the revenue it brought in?

> Is my average order value rising or falling, and is that driven by price or by items per order?

> Break revenue down by sales channel and tell me which is growing fastest.

### Products and variants

*Assumes a Product Variant or Total Products Sold report.*

> Which variants sell best within each product, and are there variants I should stop carrying?

> Which products have I never sold, and how many are there?

> Which products sell well in units but poorly in revenue, and the other way around?

> Group my products by vendor and tell me which vendors earn their shelf space.

> Which product tags are attached to my best sellers?

### Inventory

*Assumes an Inventory on Hand or inventory snapshot report. For inventory over time, use a snapshot report.*

> Based on my current stock and my sales rate over the last 30 days, which products will run out first?

> Which SKUs have not sold in 90 days, and how much stock value is sitting in them?

> Which of my top 20 sellers are currently low or out of stock?

> Compare stock levels by location and tell me where stock is sitting away from where it sells.

> Which variants keep going out of stock, based on my snapshot report over the last three months?

### Customers

*Assumes a Customer report with order counts and spend.*

> What proportion of my customers have ordered more than once?

> How long is the typical gap between a customer's first and second order?

> Which customers spent the most in the last 12 months and have not ordered in the last three?

> Split revenue by new versus returning customers, by month.

> Which products do my repeat customers buy first, compared to one-time buyers?

### Fulfilment and shipping

*Assumes a Fulfillment report, or an unfulfilled or partially shipped orders report.*

> Which orders are paid but not yet fulfilled, and how old is the oldest?

> What is my average time from order to fulfilment, and is it getting longer?

> Are fulfilment delays concentrated in a particular location, shipping method or product?

> List partially shipped orders and tell me what is holding each one up.

> Which shipping methods are customers actually choosing, and how does that split by country?

### Refunds and returns

*Assumes a Refund report, ideally alongside a Sales report.*

> Which products have the highest return rate relative to units sold?

> What do returns do to my margin per product once I account for refunded value?

> Are returns concentrated in particular variants, sizes or colours?

> How has my return rate changed over the last six months?

> Which refund reasons come up most often?

### Payouts and transactions

*Assumes a Payout, Payout Line Item or Transaction report.*

> Does this payout reconcile to the orders inside it, and what explains any difference?

> What am I paying in payment processing fees as a percentage of revenue, broken down by gateway?

> Summarise my payouts by month, with gross, fees, refunds and net.

> Which transactions failed, and is there a pattern in them?

> Show me the order line item detail behind the largest payout in this period.

### Tax

*Assumes a tax report, such as tax by country or a regional tax report.*

> Summarise tax collected by country for this quarter.

> Break tax down by state or province and flag anything that looks unusual against the rest.

> Compare tax collected this quarter against last quarter and explain any large movement.

> Which orders have no tax recorded, and is that expected?

Always reconcile tax figures against a Report Pundit export before you file or pay. See [Getting accurate results from large reports in Claude](https://claude.ai/claude-large-reports-best-practices).

### Abandoned checkouts and draft orders

*Assumes an Abandoned Checkout or Draft Order report.*

> How much value is sitting in abandoned checkouts this month, and which products appear most?

> At what point are customers abandoning, and does it differ by device or country?

> Which abandoned checkouts have the highest value and a contactable email?

> Which draft orders are still pending, and how old are they?

### Marketing and attribution

*Assumes a UTM campaign report, or an ads or analytics report.*

> Which UTM campaigns produce the highest revenue per order, not just the most orders?

> Which channels bring customers who go on to order again?

> Compare ad spend against attributed revenue by campaign and tell me which are underwater.

> Which sources send traffic that converts on first visit versus later?

### Combine two reports for questions one cannot answer

Some of the most useful questions need two reports. Ask for both in one message and Claude will run each in turn.

**Sales plus inventory,** for reorder decisions:

> Run my Sales by Product report for the last 60 days and my Inventory on Hand report, then tell me which products I should reorder this week and roughly how many.

**Sales plus refunds,** for true profitability:

> Run my Sales by Product and Refund reports for last quarter and rank products by net revenue after refunds.

**Sales plus fulfilment,** for operational strain:

> Run my Sales and Fulfillment reports for the last 30 days and tell me whether fulfilment times get worse on my highest volume days.

### Four question patterns that beat "summarise this"

These work on any report, and they are where the connector earns its keep.

**Explain the change.** Do not ask what happened, ask what caused it.

> Revenue fell 12% in August. Break that down into how much came from fewer orders, lower average order value, and product mix.

**Find the anomaly.** Let Claude look first, before you tell it what you suspect.

> Look at this report and tell me the three things you think are worth investigating, before I tell you what I think.

**Decide something.** The analysis exists to produce a decision, so ask for the decision.

> Based on this, which five SKUs should I discount out and which should I reorder? Give me a reason for each.

**Stress-test the read.** This is the question most people never ask, and it is the one that protects you.

> What would make that conclusion wrong, and which other report would confirm or contradict it?

### Ask for the output you actually need

Claude does not have to answer in prose.

> Chart revenue by month for the last 12 months.

> Put the top 50 products into a spreadsheet with revenue, units and return rate, and give me a download.

> Write me a five-line summary of this I can paste into Slack for my team.

### Make a good question repeatable

When a question works, keep it. Paste the same prompt next month and change the dates. Merchants who get the most from this end up with a handful of prompts they run on a schedule, in the same way they used to run a handful of reports.

If you find yourself running the same report and question every week, it may be worth building the report in Report Pundit and putting it on a schedule as well, so the numbers arrive whether or not you ask.

### Related articles

* [How to pull a report into Claude](https://claude.ai/pull-saved-report-into-claude)
* [How to run a report for a different date range in Claude](https://claude.ai/change-date-range-in-claude)
* [How to run a report from the Library in Claude](https://claude.ai/run-pre-created-report-in-claude)
* [Getting accurate results from large reports in Claude](https://claude.ai/claude-large-reports-best-practices)
