Design Meets Data
How systematic testing and algorithm adaptation drove 92% revenue growth for a premium furniture brand
Cozmo
Furniture & Home
12 Months
Paid Social, Paid Search, Performance Creative
Cozmo
Furniture & Home
12 Months
Paid Social, Paid Search, Performance Creative
Challenge
As a premium, design-led furniture brand with long consideration cycles, MyCozmo struggled to exit Meta’s learning phase and identify creative winners before final purchase. Faced with the challenge of scaling paid media whilst maintaining brand premium and adapting to Meta’s Andromeda algorithm update, the brand needed a testing framework that could identify high-performing creatives using early signals and drive sustainable new customer growth.
Solution
Our solution was centred on building a systematic testing and optimisation framework capable of adapting to Meta’s evolving algorithm while scaling high-performing creatives at pace.
Adaptive Creative Testing Framework:
We established a rigorous creative testing methodology designed specifically for long consideration cycles. Rather than waiting 35 days for purchase data, we used leading indicators, including engagement rates, Add to Cart events, and swatch requests, to identify promising creative directions early. This allowed us to iterate quickly while still validating performance through eventual conversion data.
The framework incorporated multiple creative formats and angles, testing user-generated content, lifestyle imagery, product-focused assets, and educational content about fabric quality and design processes. Each creative asset was tagged and tracked through our custom naming convention model and bespoke reporting, enabling granular performance analysis across the entire funnel.
Andromeda Algorithm Adaptation:
When Meta rolled out the Andromeda algorithm in Q2 of 2025, we rapidly adapted our campaign structure to align with the new optimisation approach. We consolidated previous Advantage+ campaigns following Meta’s update while maintaining clear funnel segmentation. Our prospecting efforts were allocated 70%+ of the budget, with retargeting capped at 30%, ensuring continuous audience expansion.
We restructured bid strategies to work with rather than against Andromeda’s machine learning, allowing the algorithm more flexibility while maintaining guardrails through strategic budget allocation and audience segmentation. This adaptive approach meant we could scale performance even as the underlying platform technology evolved.
Integrated Channel Strategy:
While Meta served as the primary growth driver, funnel leakage was prevented by Google Ads campaigns for search intent capture and engagement-focused retargeting. This multi-channel approach ensured we maintained visibility throughout the customer journey, from initial awareness through the lengthy consideration period to final purchase.
Performance Monitoring and Optimisation:
To preserve agility and ensure ad fatigue was identified early, we established optimisation routines focused on creative performance, audience saturation monitoring, and budget reallocation.
The creative testing framework enabled us to identify and scale winning assets quickly, while sunsetting underperformers before creative fatigue impacted results. This systematic approach to creative lifecycle management proved essential for maintaining performance throughout seasonal fluctuations.
New Customer Acquisition
ROAS
Increase in Ad Spend
More creative tested
Top Ads
They helped us really deliver on our growth targets this year. The team have the right balance between data, analysis and creative strategy. It's enabled us to make more assets and measure them even better than we have previously
