# How do I automatically export Amazon sales and ad data into Google Sheets?

A playbook for an AI assistant connected to Nova. It reads your daily sales, units, ad spend, ad sales and profit per product, and your campaigns, and lays them out as tables ready for Google Sheets. It only reads your data.

Every step calls a Nova tool that reads your live Amazon data, so paste this into Claude, ChatGPT or any assistant connected to Nova's MCP server. Run it once to design the sheet, then let a Nova scheduled export keep it current.

```text
Use Nova to build the daily sales and advertising tables I will keep in Google Sheets, and tell me how to keep them updated.

Period: the last 14 complete days. If I name other dates, use those.

Scope: all my Seller Central accounts unless I name some.

1. List the accounts. Call Nova's list_accounts. Give each account's marketplaces and channel. Say if an account is Vendor Central: scheduled exports do not cover Vendor Central yet.

2. Read the daily totals. Call Nova's query_metrics with groupBy "day", excludeUnfinished true, and the metrics netSales, units, adSpend, sponsoredSales, organicSales, sponsoredTacos and netProfit. One row per day.

3. Read each product by day. Call Nova's query_metrics with groupBy "childAsin", timeGrain "day", limit 50, and the same metrics. Lay it out long: one row per day and product, with the date, the child ASIN, then one column per metric.

4. Read the campaigns. Call Nova's query_advertising with groupBy "campaign", the metrics spend, sales, orders, clicks and acos, sorted by spend, limit 50, for the same period.

5. Format the tables for Google Sheets. Give each one as CSV in its own block: one header row, dates as YYYY-MM-DD, numbers without currency symbols or thousands separators, and the currency code Nova returns in its own column.

6. Tell me how to keep it current. Map the columns of steps 2 and 3 to a Nova scheduled export: the metrics to pick in the Sales, PPC and Profitability groups, Time grain Daily, Split by Child ASIN, and a rolling date range. Then give me the IMPORTDATA formula to paste in the sheet, with the export link left as a placeholder.

Rules: never add amounts of different currencies. Campaign figures from query_advertising use Amazon's attribution windows and never add up with adSpend from query_metrics, so keep them on a separate tab. Relay every notice Nova returns, such as cogsNotice, adsHistoryNotice, marketplaceNotice and currencyNotice, before the figures it qualifies. A day with no ad data before the account's ads history is missing, not zero. You cannot create the scheduled export or write to my sheet: give me the settings and I set them up.
```

## How to read the result

| What you see | What it usually means | What to do |
| --- | --- | --- |
| netProfit higher than you expect | Some units sold have no product cost in Nova | Enter the missing costs, then rebuild the export |
| Ad spend missing on the first days | The period starts before the account's ads history | Start the range later, or read those days as missing |
| Campaign spend differs from adSpend | Campaigns use Amazon's attribution windows, adSpend is allocated to products | Keep the two on separate tabs |
| Totals higher than one Amazon screen | The account adds several marketplaces | Add Split by Marketplace, or filter the marketplaces |
| Yesterday lower than the days before | The day is still filling as orders and ad reports arrive | Read it again after the next export rebuild |

## Limits

- Nova rebuilds a scheduled export every 3 hours. Google Sheets then reads the link on its own schedule.
- Vendor Central accounts are not in scheduled exports yet. An assistant connected to Nova can still read them and return the tables.
- Sessions, conversion rate and reviews cannot be split by SKU. Use Child ASIN.
- Anyone who has the export link can download the file. Keep it inside your team.
- The assistant reads Nova. It does not create exports or write to your spreadsheet.