AWS Machine Learning Blog

New Engen improves customer acquisition marketing campaigns using Amazon Rekognition

New Engen is a cross-channel performance marketing technology company that uses its proprietary software products and creative solutions to help their clients acquire new customers. New Engen integrates marketing, AI, and creative expertise to provide a one-stop solution that helps their customers optimize their digital marketing budget across Facebook, Google, Instagram, Snap, and more.

Improving ad campaign performance

To help improve ad campaign performance, New Engen uses Amazon Rekognition, a deep learning-based image and video analysis service, to extract metadata from images in their creative ads library. This metadata is then incorporated into internal reports, which organize images by common labels. The reports display how well each ad image performed. For example, photos that include labels such as person, cosmetics, and city are aggregated together along with a performance analysis using metrics such as engagement rate, views, and total clicks. This enables the creative team to easily filter for ad images that have similar labels to assess which common aspects of certain ad images caused some to perform well and others to perform poorly.

How it works

Building creative ad images is expensive, so New Engen wanted to ensure that creative decisions were being driven by data coupled with ad image analysis. To illustrate this, the simplified example that follows shows how the creative team uses image metadata for data-driven decisions. The report that aggregates all of the metadata allows the creative team to select the account they are working on and see how similar images have performed. In this example, the images that are marked “bad” and “average” have significant differences compared to the images marked “good.” For example, in the “average” image the dog isn’t looking at the camera, and in the “bad” image a silver bowl isn’t used and the dog’s head is turned. Bringing this information together in one place has enabled New Engen’s creative team to be more efficient and data-driven when creating ads for their clients or running A/B tests.

“Amazon Rekognition has helped our algorithms analyze hundreds of thousands of images to locate subjects and materials that resonate most with our customers, by delivering consistent image tagging and analysis in an efficient period of time,” says Charles Brophy, Lead Engineer at New Engen. “Rekognition has enabled us to streamline the process of developing creative ads and gives us back valuable time that was being spent on determining what image properties drive successful marketing campaigns.”


Using Amazon Rekognition has enabled New Engen to take the guesswork out of choosing what creative ads should be built and what images will resonate with particular audiences. Creative designers can now filter New Engen’s ad images by the labels generated from Amazon Rekognition, and easily find images that have already proven to be successful.

“This simple use of extracting metadata from Amazon Rekognition has given our creative teams the data they need to quickly find images that have proven success, as well as what type of images they should create,” says Brophy. “Using Rekognition has been simple and we’ve accomplished this outcome just by using the DetectLabels API.”

About the Author

Kaiser Larsen is a Product Marketing Manager for AWS artificial intelligence solutions. Outside of work you’ll find him hiking, cooking for family and friends, and eating ice cream whenever there’s an excuse to celebrate.