Thousands Of Variations To Test & Learn

Client
Baylis & Harding
‍
Project Type
Creative Automation, SEO,
DSP Advertising

Results
+23% Higher CTR,
Insight Into 5+ Variables

Description
‍Baylis & Harding were getting started with Amazon DSP, allocating a sizable budget for broad reach. However, without prior research on the best creative approach for the platform, they needed to uncover it mid-campaign and act fast.

Key Objectives

  • Identify aspects of the creatives assets that will lead to increased CTR.
  • Test variations of these aspects utilising DSP's bulk-creative-upload feature.
  • Provide near real-time performance insights, allowing for quick adaptation.

Process

  • Created a strategy brief, defining key messaging and visual variables to test, drafting initial assets as templates.
  • Automated asset creation with Python and Photoshop, making 1034 creatives (with embedded metadata) for every possible variable-combination.
  • Bulk uploaded assets to DSP, receiving performance data per-creative, then leveraging the image metadata for per-variable analysis.

Variation-Strategy Brief

Asset Template

Automated Output With Metadata

Reporting By Variable

Outcome

By analysing the data by variable, we quickly identified winning approaches, like light colour-palettes outperforming darker ones. Iteratively refining this way for each variable led to only using the most optimal creatives, with a 23% increase in average CTR from the campaign's outset.

For Baylis & Harding's next DSP campaign, they could use all the insights learned here as a basis, whilst still being able to choose other variables to test upon for further streamlined performance.

Want to get in touch?
Drop me a line!

Thank you! Your submission has been received!
Oops! Something went wrong while submitting the form.