When to split a campaign, and when to leave it alone

Every account we take over has the same instinct baked in somewhere: split the campaign. Split by match type, split by ASIN, split by placement. It’s usually offered as a default best practice. It isn’t one — it’s a decision with a real cost, and most of the time that cost gets ignored.

What a split actually resets

A campaign’s history is the reason it performs. Amazon’s algorithm has learned which search terms convert, which placements are worth a higher bid, what time of day and day of week the account does best. Split a campaign and that learning doesn’t transfer. The new campaign starts cold — no conversion history, no placement data, nothing to bid confidently against. For two to four weeks, sometimes longer, performance dips while the algorithm relearns what the old campaign already knew.

That’s the trade sellers rarely price in. “More granular control” sounds like a pure upside. It isn’t free.

When the split is worth it

A few situations genuinely call for it:

• The campaign is masking a real difference in intent. Broad match and exact match terms sitting in the same campaign, competing for the same budget, is a common one — the broad terms burn spend the exact terms could have converted.
• One ASIN or match type is starving the rest of the budget. If a single high-volume term consistently eats the daily budget before the rest of the campaign gets a look, the others never get a fair test.
• You need separate reporting for a real business reason — a new product launch you want isolated, or a category you’re deliberately scaling differently from the rest.

In each case, the split solves a problem the current structure is actually causing.

When to leave it alone

Most of the time, the answer is no:

• The campaign is performing fine and the “problem” is really just a desire for tidier reporting.
• The account is already under-spending its daily budgets — more campaigns means more budgets to manage, not more control.
• The real issue is bid or placement optimization within the existing structure, not the structure itself. Splitting won’t fix a bid problem.

The short rule

Before splitting, ask what specific decision the split lets you make that you can’t make today. If there’s a real answer — reporting you need, budget being starved, intent being blended — it’s worth the reset. If the answer is “it would just feel cleaner,” leave it alone and let the campaign keep the history it’s already earned.

← Back to Notes