How advertisers can use Amazon Ads solutions as a creative testing engine

Most advertisers still make creative decisions the way they always have: through qualitative feedback, creative briefs built on assumptions, small test groups, and post-campaign summaries. These approaches show what happened, but they rarely explain what influenced audiences' behaviors or how creative decisions ripple through the full customer journey.

Amazon Ads offers something structurally different. You can test creative against shopping and streaming signals to understand which messages resonate with relevant audiences.

This enables creative testing that goes beyond attention metrics — connecting ad exposure to shopping outcomes like consideration and purchase, so every test reveals real business impact.

At Flywheel, we approach creative testing as part of your full-funnel strategy. Using Amazon Marketing Cloud (AMC) and proprietary measurement frameworks, we help brands understand how upper-funnel creative connects to awareness, consideration, and purchase outcomes, measuring impact across the entire customer journey. Based on our work with enterprise brands, we've compiled five tips to build an always-on creative learning system that helps improve performance across the funnel.

1. Start with the question

Strong creative tests begin with a focused hypothesis rooted in consumer behavior. Flywheel’s approach is unique in that we look across the entire funnel simultaneously and use Amazon’s retail signals to help our clients answer complex business questions. Are streaming audiences exposed to your brand later showing up in branded search? Are consumers browsing multipacks but buying single units? Are new-to-brand shoppers viewing detail pages but not converting?

With many advertising solutions, a creative test tells you what people engaged with and the story stops there. With AMC, you can understand how audiences move through their entire journey.

At Flywheel, we start by mapping your current funnel health. Where are audiences progressing? Where are they dropping off? Where might creative be the lever? We then design tests that measure how creative decisions influence funnel movement. Does this creative drive more new-to-brand awareness that converts downstream? Is it associated with increased repeat purchase among existing audiences?

Before launch, it’s important to align on the success metric tied to that funnel view: conversion rate, detail page view rate, units per order, new-to-brand share, branded search activity, or repeat behavior over 30-60 days. Then, connect it to a broader funnel question: is this creative connecting relevant audiences to purchase outcomes?

2. Limit the variables

Cleaner test design leads to better decisions, especially when you're measuring across the full funnel. Advertisers should isolate one or two creative variables at a time while holding the rest steady. If the headline, image, offer, and audiences all change at once, the result is hard to interpret. A better approach is to keep the product, audiences, budget, and landing experience constant while changing only one element, such as the opening video frame or the benefit statement.

Advertisers should also watch for retail variables that can distort results, such as out-of-stocks, limited availability, price promotions, or deal badges. In those cases, the strongest-performing asset may not be the strongest creative. A creative that performs well when inventory is high but fails during stockouts tells you something very different about consumer resonance than creative that performs consistently across conditions.

3. Match format to learning

The test format should match the signal you want to learn from, and this is where a full-funnel approach unlocks the greatest opportunity.

For lower-funnel and search, Sponsored Products help advertisers understand product-level performance close to purchase with creative tests across hero images, pack sizes, and variation strategy. Sponsored Brands and Sponsored Brands video allow for testing headlines, value propositions, and short-form storytelling where relevant audiences are exploring your category.

For upper-funnel and awareness, Amazon DSP and Prime Video ads help brands evaluate broader storytelling and messaging with audiences unfamiliar with your brand before they enter a search mindset. Dynamic TV Creative takes this further by allowing advertisers to generate multiple creative variations for different audiences, products, or moments across Amazon’s streaming TV inventory, which means you're testing messaging alignment with signaled consumer intent.

It’s important to not treat upper-funnel in isolation. A piece of streaming creative might generate strong engagement but fail to drive consumers downstream to branded search or product detail pages. Conversely, a display creative that seems underperforming on engagement might be perfectly calibrated if it's driving high-intent traffic to detail pages and converting at strong rates. The key is measuring the full path and whether the creative moves consumers through the funnel all the way to conversion.

Flywheel uses Amazon Marketing Cloud to connect these dots. This level of visibility lets you optimize each creative touchpoint knowing exactly how it influences the next stage and final business outcome. Generative AI ad solutions, like Amazon Ads Creative Agent and Dynamic TV Creative, help brands move faster by reducing production friction and making it easier to build variations for testing.

4. Measure what matters

Flywheel's full-funnel measurement methodology structures measurement in three layers, then connects those layers to your overall funnel health.

First, measure immediate response. Which creative drove stronger engagement, detail page views, branded search activity, or conversion during the active campaign window?

Second, evaluate audience engagement. Which creative variant brought in more new-to-brand consumers, higher units per order, stronger basket-building patterns, or movement into higher-value products? Two creatives might drive the same conversions, but one brings in consumers who buy more, repeat more, and have higher lifetime value.

Third, assess longer-term consumer value. Which exposed consumers returned to purchase, repeated within 30-60 days, or showed stronger loyalty signals?

These three layers feed directly into your overall “Return on Consumer” (RoC). Built on AMC and using Amazon's full-funnel measurement and Brand Metrics, Flywheel's proprietary RoC dashboard helps advertisers better understand funnel health, showing where audiences are progressing efficiently (opportunity, awareness, interest, purchase, loyalty) and where they're dropping off. When you layer creative testing insights informed by Amazon’s retail signals into your RoC analysis, you see exactly where creative decisions inform and contribute to funnel movement and where optimization is most urgent.

In practice, this approach drives measurable results. We worked with a leading household air care brand, who used Amazon's browsing, streaming, and shopping signals to inform a modular creative system that matched products to specific seasonal moments. They also leveraged Amazon Ads AI-powered creative tools to dynamically scale creative variations across Sponsored Brands, Display, and Prime Video. The outcome: Glance Views up 26% (vs. a 10% target), new-to-brand customers up 19% (vs. a 5% target), and ROAS up 32% (vs. a 5% target). Each metric mapped to a different layer of the funnel, turning a single creative test into a full-funnel optimization.¹

Here is the distinction between testing a creative asset and optimizing a funnel. Flywheel’s approach uses Amazon Marketing Cloud to measure creative effectiveness, then applies that insight to your brand’s holistic RoC to understand how creative decisions inform customer journeys and long-term outcomes. Once these patterns are clear, you activate campaigns across the funnel, building audiences based on signals such as exposure, engagement, and purchase status, informed by what you've learned about which creative moves the relevant audiences to the next stage.

5. Build a learning loop

The goal is to create a repeatable system that connects individual creative tests to your broader funnel strategy. Document each hypothesis, test structure, audience, metric, and result, then feed those learnings back into the next brief and funnel analysis. Over time, those results inform future creative development and broader campaign planning. Instead of relying on assumptions, brands understand how creative decisions influence overall RoC and funnel health. By treating creative tests as part of a full-funnel optimization engine powered by AMC insights and measured through a RoC lens, brands learn faster, activate smarter, and build stronger long-term consumer value.

Sources

¹ Partner-provided data, 2025.

About the author

Bernie Che is the Senior Director of Retail Acceleration at Flywheel, where she oversees relationships with all of Flywheel’s full-funnel retail media partners. With deep expertise in digital media and performance marketing, Bernie helps global brands navigate the complexities of today’s omnichannel landscape—driving growth through integrated, data-driven solutions. Since joining Flywheel in 2018, Bernie has led efforts to unify campaign planning, optimization, and measurement across platforms, ensuring brands can connect with shoppers wherever they browse and buy.