Send a conversion event per brand, and stop letting your best margin subsidise your worst

Multi-brand retailers optimise against a Purchase event and then wonder why the algorithm keeps finding customers for the cheap, high-return, low-margin brand. It is doing exactly what it was told. Every sale looked identical.

What a per-brand event changes

Fire a distinct conversion when someone buys a particular brand, and you can bid for the customers who buy that brand specifically. Not "people who buy from us" - people who buy the thing you actually want to sell more of.

• Build a lookalike from buyers of one brand rather than from all buyers blended together.
• Run a campaign that optimises for a brand you have an exclusive on, or excess stock of, or a co-op budget from.
• Exclude buyers of a brand from prospecting for it, without excluding them from everything.

Use cases worth the setup

1. 1Distributor or co-op funding. When a brand pays part of your media cost, they want to see conversions for their products, not your store total. A per-brand event is the report.
2. 2Margin steering. Your best-margin brand and your worst can have identical price points. Only a distinct event lets bidding tell them apart.
3. 3Exclusivity. If you are the only stockist for a brand in your market, the audience of people who buy it is genuinely yours to build.
4. 4Launch pushes. A new brand has no purchase history to optimise against until you start sending its event - so start sending it before the launch, not after.

What to do next

List your brands by margin rather than by revenue. If the order is different from what you would have guessed, that difference is what per-brand events are for.