Pattern Revolutionizes Ecommerce Optimization Using Amazon Nova Foundation Models

Powered by AWS, Content Brief lets Pattern’s customers analyze a wide range of elements, including keywords, search terms, product images, and customer reviews. To process the vast amounts of data required for this kind of comprehensive ecommerce analysis, efficiency is crucial. Seeking an optimal solution, Pattern turned to Amazon Bedrock—a fully managed service that offers a choice of high-performing foundation models—for efficient inference capabilities to optimize token usage, reducing costs while maintaining high-quality outputs. And using Amazon Bedrock, the Pattern team also gets access to various foundation models tailored to different tasks, including Amazon Nova Foundation Models which offer frontier intelligence and industry-leading price performance. These efforts help Pattern’s customers see significant improvements in traffic, conversion rates, and revenue.

In ecommerce, optimizing product content such as imagery, video, and copy for online listings is crucial for driving traffic and conversions. Managing this effectively at scale is challenging. According to Pattern’s 2024 Ecommerce Executive Strategy Snapshot, 58 percent of executives plan on increasing their investment in product imagery, video, and written copy for online listings to maximize ecommerce opportunities. The traditional approach of conducting competitive research and content optimization can take a week or more, with multiple team members working on a single product. This is virtually impossible to scale, especially for companies managing thousands of products across multiple marketplaces.

For some Pattern customers who have many brands and SKUs, the ability to rapidly analyze market data and optimize content is essential for maintaining a competitive advantage. Pattern identified the opportunity to use its vast dataset of over 44 trillion ecommerce data points to help brands transform content optimization, helping teams to move from conducting time-intensive research to focusing on rapid content updates and testing.

Pattern’s team needed to develop a sophisticated prompt-engineering process that could continually refine prompts to optimize both quality and efficiency. The Pattern team found that using Amazon Bedrock let them tailor the model’s behavior precisely to their needs, improving the accuracy and relevance of AI-generated insights.

This fueled the development of Content Brief, Pattern’s powerful AI-driven tool built on AWS. Content Brief uses Amazon Bedrock to provide access to multiple foundation models, including Amazon Nova Foundation Models, which Pattern employs for high-volume data processing tasks such as keyword classification and search-term analysis. The cost-efficiency and superior performance of Amazon Nova Foundation Models helps Pattern to process massive amounts of data quickly and affordably, facilitating rapid analysis of market share, consumer demographics, discovery behavior, and content performance across multiple marketplaces simultaneously.

Pattern’s solution uses a suite of AWS services to create a comprehensive and scalable application. Amazon Simple Storage Service (Amazon S3), an object storage service, stores vast amounts of product images and data, forming the foundation of Pattern PXM’s data-point repository. Amazon Textract—a machine learning service that automatically extracts text, handwriting, layout elements, and data from scanned documents—extracts and analyzes text from product images, facilitating detailed comparisons with competitor listings. Meanwhile Amazon DynamoDB—a serverless, NoSQL database service—powers Content Brief’s rapid data retrieval and processing capabilities, storing both structured and unstructured data. For computationally intensive tasks, Amazon Elastic Container Service (Amazon ECS), a fully managed container orchestration service, with GPU support provides dynamic scaling based on demand. Pattern also uses security features from AWS, such as AWS PrivateLink, which provides private connectivity between virtual private clouds, to verify that data transfers occur over private IP addresses, enhancing security and maintaining compliance with data protection regulations.

The business impact for Pattern PXM’s brand partners and customers has been transformative, providing rapid access to data-driven insights and comprehensive analytics that were previously scattered across multiple sources. As demonstrated in one study, Content Brief helped achieve a 21 percent month-over-month revenue surge, 14.5 percent more traffic, and a 21 basis-point conversion lift. Pattern’s ability to process and analyze massive amounts of data quickly and affordably through Amazon Nova Foundation Models has democratized access to sophisticated ecommerce optimization tools, making enterprise-level insights available to brands of all sizes. The platform has become an indispensable tool for ecommerce operations, helping brands to make data-backed decisions at scale.

Additionally, using Amazon Nova Foundation Models through Content Brief has delivered significant performance improvements in both efficiency and cost-effectiveness. For example, Pattern was able to reduce the cost of its keyword classification process by 76 percent after implementing Amazon Nova Foundation Models. This dramatic cost reduction directly benefits Pattern’s customers, as the company is now able to offer advanced analytics at a more accessible price point for brands of all sizes. The solution has also dramatically reduced processing time for content analysis while maintaining high accuracy levels, helping Pattern to scale their content optimization capabilities across their extensive brand portfolio.