Here's a structure I've landed on after plenty of spreadsheet wrestling.
Start with two campaigns:
- One that includes every single product, fed 20% of the total budget.
- One that only holds your top 100 best sellers, given the remaining 80%.
Keep an eye on performance in the full catalogue campaign. As products there start showing solid ROAS or conversion rates, gradually shift them into the top‑100 campaign and drop the underperformers out. Over time you're essentially using the wider campaign as a testing ground and the narrow one as your main revenue engine.
This way you don't waste budget on duds early on, but you also don't miss a hidden gem that might only need a bit of exposure. Spreadsheets don't lie - test, move, repeat.