Define the product fields you need
Start with the decision the dataset will support. Typical fields include product title, current price, original price, discount, rating, sold count, shop, location, URL, image URL, and capture time.
Collect a consistent result set
Search results change while scrolling and may vary by market, account, and campaign. Record the query, market, sort order, filters, and capture time so another person can understand how the dataset was produced.
Normalize before export
Convert prices and sold counts into consistent numeric fields while retaining the original displayed value when useful. Deduplicate by stable product URL or product ID instead of title alone.
Choose CSV or XLSX
CSV is portable and easy to load into scripts or databases. XLSX is easier for analysts who need typed columns, filters, multiple sheets, and human-readable formatting. Keep JSON when nested source details must be preserved.