Big news! We’ve announced that we’re growing our partnership with Google Cloud to maximize the AI powering Bloomreach Discovery, our search and merchandising platform used by ecommerce businesses worldwide.
Integrating Google Cloud’s AI with Bloomreach Discovery aligns with the ethos of innovation, collaboration, and accessibility that modern technology champions. The collaboration will improve functionality across our suite of solutions and push the boundaries of ecommerce product discovery technology. More specifically, we’ll combine our commerce dataset and AI, Loomi — along with Bloomreach Discovery’s recall, ranking, personalization, conversational, and merchandising engines — with Google’s Gemini models and Vertex AI platform.
Traditionally, relevance and personalization are viewed as two separate problems. With this partnership, we aim to change this by integrating models from Google Cloud directly into the Discovery platform. This way, customers will benefit from the combination of Google Cloud’s sophisticated AI innovations with Bloomreach’s extensive ecommerce know-how, comprehensive data sets, and AI capabilities that drive revenue.
How Bloomreach Plans To Transform Product Discovery With Google Cloud
Instead of merely building Bloomreach’s technology on top of Google Cloud’s AI capabilities as so many have done, we’re integrating their models with our own artificial intelligence and proprietary data on a foundational level. Furthermore, we’re customizing Google Cloud’s general-purpose AI models specifically for ecommerce.
What sets this partnership apart is the pioneering spirit of both companies. Bloomreach is proud to utilize Google Cloud’s Vertex AI platform to bring AI to ecommerce customers. From recall and ranking to personalization and conversational interfaces, Bloomreach will work with Google Cloud to tackle the most prevalent challenges in product discovery today. With new features — like an LLM-based precision mode — we’re already seeing incredible preliminary results, including a 25% increase in search revenue per visitor (RPV) and an average conversion rate of 4-7%.
Let’s break down how each of Bloomreach’s five search engines will be improved by our expanded partnership with Google Cloud.
Recall
At the forefront of our expanded partnership is the pursuit of a hybrid search model that can transition between vector, semantic, lexical, and other query evaluation models to deliver optimal search results tailored to the query, the customer, and the moment. For example, vector models are indispensable for unraveling the complexities of long-tail queries with three or more words (which is ~30% of total queries), while semantic understanding is better for tight queries of one to two words. By prioritizing the development of a hybrid model, Bloomreach is ushering in an era of cost-effective solutions that equip retailers and brands with precision tools to address each query type individually instead of using a one-size-fits-all approach.
In addition to our advancements in query evaluation, we’re also exploring avenues for catalog enrichment. By leveraging generative AI (GenAI) like large language models (LLMs) and computer vision capabilities, we can enrich catalog data by extracting new product attributes from images to give customers a deeper understanding of product offerings.
Ranking
With ranking, Bloomreach also champions customization over a blanket approach. Unlike many vendors in the industry — that require retailers to construct their own algorithms from scratch or confine them to inflexible black box algorithms — Bloomreach empowers merchandisers with user-friendly tools to influence our AI. Our ranking technology now promises unprecedented levels of flexibility to craft custom algorithms at the site, category, or even query level. Merchandisers can ultimately choose their path, whether it’s activating Loomi to implement ranking weights or overriding our AI with their preferred rules. Better yet, merchandising practitioners can rigorously A/B test these algorithms within an easy-to-use interface before pushing them into production, making Bloomreach a solution that continues to be experimentation-forward (without having to wade through the complexities of algorithmic development).
Today, it’s more important than ever for merchandisers to keep up with trends. Bloomreach will allow business users to exercise granular control over their ranking algorithms where they see fit, supporting agile adjustments that cater to customers’ preferences at a moment’s notice.
Personalization
Last year, Bloomreach took a significant step forward by integrating Bloomreach Engagement with Bloomreach Discovery, creating our real-time segments feature. This innovation enabled Bloomreach Discovery to utilize customer data from the Engagement product. Every other solution on the market uses clickstream data or user session data for personalization, which is limited to what you can collect from a website and only lasts 30 days. By incorporating data from our customer data engine, Bloomreach surfaces demographic and multichannel data that lasts indefinitely in a customer profile, allowing us to create personalization that customers will notice when they shop.
In our next phase of personalization, we’ll create user embeddings to work with the product embeddings we’ve already enhanced for recall. Again, using Google Cloud’s Vertex AI platform, our new approach to 1:1 personalization involves creating a user embedding, which encapsulates a user’s browsing behavior and various types of history. This user-specific data is then compared with the product information contained within a product embedding, utilizing the concept of vector dissonance to craft a highly personalized experience for each individual on your ecommerce website at scale.
Conversational
In 2023, Bloomreach also teased the release of Bloomreach Clarity at The Edge Summit. Clarity capitalizes on the innovative potential of conversational shopping as another revenue channel for your ecommerce business. Our goal with this product is to enable businesses to use a conversational interface as an effective shopping assistant to interact with shoppers in targeted ways. For instance, if a customer is about to bounce, Clarity can be triggered to step in with an observation or question before they leave your site, mimicking the level of customer service shoppers historically experience in a physical store.
Beyond our new AI shopping assistant, Bloomreach is also exploring other ways to introduce conversational experiences within search and product discovery while keeping cost-effectiveness in mind. Imagine a scenario where your customer searches for “levis skinny jeans” and receives alternative options within the same category. This results set could also now trigger a conversational opportunity, providing shoppers with a deeper explanation of why these alternatives are recommended. Even though conversational queries only represent about 1-2% of queries, this modest fraction is rapidly expanding with the proliferation of GenAI, allowing your brand to foster engaging interactions and build trust and transparency.
Merchandising
Bloomreach has consistently set itself apart in the realm of ecommerce technology with its standout merchandising capabilities, blending a healthy mix of AI with manual controls to empower merchandisers to excel in their roles. We leverage real-time, first-party data to directly influence merchandising, presenting the top opportunities to practitioners through Loomi. This reflects a conscious effort to combine the precision of AI with the nuanced insights of human expertise.
Now, there’s an exciting opportunity to take advantage of copilot experiences that grant merchandisers even greater intelligence and automation to truly excel in their roles. With a copilot, merchandisers can navigate complex data landscapes more effectively, uncover valuable insights and trends, and ultimately drive better results for the business.
The Next Iteration of Bloomreach Discovery
At our core, Bloomreach has always been an AI company. And we remain committed to providing technology that delivers real impact to businesses. Our focus now with this growing partnership is to continue driving tangible business outcomes and enhancing customer experiences. Fortunately, our early experiments with the next iteration of Bloomreach Discovery have shown promising results for driving AI-powered impact.
Want to see what’s next with Bloomreach’s AI? Check out our latest AI product releases and innovations.
No, Loomi is the connective tissue across our suite of products and can power seamless ecommerce experiences on every channel — in real time and at scale. Bloomreach customers turn to Loomi to help with a wide range of use cases, such as preventing null results by automatically recalling similar items or ranking products based on their performance.
There’s a lot of talk in the ecommerce product discovery space about clickstream data, which focuses on browsing behavior like search input focuses, add-to-carts, purchases, and more. However, this data is limited due to the age of the cookie (30 days). Bloomreach believes that customers should have the ability to store their first-party data (e.g., email interactions, loyalty status, most commonly purchased) indefinitely and securely to provide a true lifetime view of the prospect or customer. With Google Cloud, we’ll only continue to build upon our personalization promise to the industry.
Bloomreach Discovery’s expanded partnership with Google Cloud brings exciting new capabilities to the table, but it’s important to note that accessing most of these features will require purchasing a premium package. While this may present an additional investment, it also unlocks a wealth of powerful tools and functionalities that can significantly enhance your online presence and performance. By investing in this premium package, your business can maximize the value it derives from Bloomreach Discovery.