Amazon DSP Performance+ and Brand+ Campaign Management

Amazon Performance+ vs. Brand+: which goals each one serves

Performance+ may look like a single campaign type, but in practice it splits into three goals; the fourth goal sits with Brand+. The one we choose changes the audience signal used and the success metric. Brand+ works at the top of the funnel to find new audiences likely to engage over time. We do not turn on all four goals for a brand at once; we start with the goal where conversion data is strongest.

• New customer acquisition: finding new audiences similar to existing customers with predictive models
• Remarketing: bringing back users who showed interest in your site or app
• Loyalty and cross-sell: showing existing customers a second product or a renewal
• Brand+ prospecting: broad, qualified reach for awareness and consideration

Feeding Performance+: Amazon Ads tag, Conversions API and MMP

Performance+ combines the brand's first-party conversion data with Amazon's shopping, streaming and browsing signals to build predictive models. On a website, this data comes through the Amazon Ads tag or the server-side Conversions API; in a mobile app, through an MMP. We install the tag with Google Tag Manager, but the real work is deciding which event counts as a "conversion".

The most common mistake we see is choosing easy events such as page views or add to cart as the main goal. The model then drifts toward audiences that produce cheap, plentiful events, and real sales do not rise. We choose the goal event as the step closest to the sale with enough volume and, where needed, send quality fields from lead forms back through the Conversions API.

What we control and what we leave to the model

As Amazon describes it, setup takes a few steps; geo targeting and brand safety controls stay available, and reporting is at campaign level. For teams used to seeing audience and inventory breakdowns at line item level, that is a loss. We keep control in three places: the goal event, geography and brand safety. We leave audience selection to the model, because narrow audiences added by hand slow down learning.

The approach will feel familiar to teams who run Performance Max on Google: we provide the goal and the signal, the system handles delivery. The difference is that Amazon's shopping data goes into the model, and creative comes from the display and video set you upload rather than being generated from a product feed. That makes creative variety even more important here; a Performance+ campaign launched with a single banner saturates quickly.

How we test Performance+ incrementality

Automation looking good in its own report does not mean it brings extra sales; especially with the remarketing goal, impressions may be going to users who would buy anyway. So we compare the Performance+ campaign with a classic line item for the same goal, format and period, and set up a geo-split test where possible. We run the test with growth experiment discipline, with a success metric written in advance, and confirm the result with GA4 data.

Performance+ and Brand+: when do we recommend it?

How we manage Performance+ and Brand+

1. Goal event selectionWe pick the event closest to the sale with enough volume and keep secondary events separate.
2. Signal setupWe install the Amazon Ads tag, Conversions API or MMP connection and test event matching.
3. Goal separationWe open new customer, remarketing and loyalty goals in separate campaigns with separate budgets.
4. Setting limitsWe set geography, brand safety and existing customer exclusions from day one.
5. Comparison testWe measure incremental impact against a classic line item or a geo control group.

Performance+ and Brand+: metrics and requirements

The Brand Move Amazon DSP Team

To learn more about Performance+ and Brand+ or other Amazon DSP ad formats, contact The Brand Move Amazon DSP team today.

Contact us now