Yobi Partners with Microsoft
Yobi partners with Microsoft to democratize responsible enterprise access to customer behavioral data
In today’s data-driven world, understanding customer behavior and having the ability to predict customer intent through data and analytics is a paramount advantage that businesses should embrace to remain competitive. Yobi, founded in 2019 by Max Snow, Bill Wise, and Tom Griffiths, is on a mission to enable companies to compete with powerhouses like Amazon and Google and their ability to predict their customers’ buying behaviors. Harnessing the power of Microsoft Azure Databricks and Azure Data Factory, Yobi is helping to democratize ethical access to customer third-party data for all.
To us, Microsoft represented more than a service provider. It was truly a partner, not only in driving our solution forward, but also in providing strategic go-to-market support. Yobi wouldn’t exist today if it weren’t for the partnership, leadership, and support that Microsoft has provided us.
Max Snow, Founder & CEO of Yobi AI
Using third-party data in an ethical, responsible, and secure manner
Using third-party data in an ethical, responsible, and secure manner
To create more personalized customer experiences and accurately predict shopper intent, large tech companies collect more data about their customers than many other industries today. How do they do this? They gather customer behavioral data across their omnichannel engagements. However, this is creating a widening disparity in competition between businesses. The overall shift to first-party data, in large part due to new privacy legislation, limits decision science capabilities and creates an insurmountable advantage for companies with access to massive amounts of first-party customer data—in terms of personalization and customer acquisition. Yobi, in partnership with Microsoft, is on a mission to change this narrative by creating parity between large tech companies and the rest of the world. Using the Azure platform for data services, machine learning technologies, and storage solutions, Yobi is forging a new path forward when it comes to customer behavioral data.
Yobi’s technology translates raw behavioral data into a unique, machine-readable string of numbers (such as a vector embedding) that can be directly deployed into existing company models. These embeddings are built on the unification and compression of trillions of raw behavioral events to statistically represent a customer’s dynamic behavior. For example, customer behaviors—such as “Kim watches action movies” or “Dave buys red flowers”—can be captured, analyzed against other customer behaviors, and used to generate a prediction. This prediction is represented numerically to protect customer privacy and prevent companies from accessing raw customer information. The raw data is compressed during the embedding process, preserving privacy and, in some cases, reducing storage fees by 10 times the typical costs. When Yobi’s technology is deployed into an existing company model, it helps businesses synthesize first-party data and customer relationship management data while streamlining the use of third-party information. This technology creates a complete picture for customer behavioral predictions that helps to drive up revenue by uncovering trends. It also provides personalized shopping experiences and helps businesses understand the impact of their media and ad campaigns.
Yobi is committed to humanity and ethics. Its strong ethical practices ensure that data aligns with the following pillars:
- Compliant: Ethically sourced, fully consented data complies with all privacy laws and best practices.
- Clean: Predictive signals from raw data are expertly extracted using behavioral science principles.
- Comprehensive: Identifiers across datasets are unified to create a holistic view of customer behavior.
- Trustworthy: All datasets undergo extensive testing for data bias, statistical signals, and provenance.
- Private: Customer representations protect individual privacy without sacrificing predictive signals.
Yobi has already observed significant results from its technology. Examples include the following:
A number of behavioral predictions were captured with 99-percent accuracy. This includes predicting the likelihood of an individual buying almond milk, motor oil, car tires, gift cards, a household pet (like a dog or cat), fruit, baby food, beer, hard seltzers, and more.
Customer acquisition for the History Channel increased by 500 percent.
Machine learning predictive performance for a Fortune 500 insurance company improved by 10 percent, leading to a three-times increase in website marketing conversions.
Seeking more than a cloud service provider, Yobi turns to Microsoft for partnership
When Yobi first launched in 2019, Amazon Web Services (AWS) was its cloud service provider. But with the customers it was trying to reach—who were either already on Azure or making their move to Azure—Yobi elected to pursue a deal with Microsoft. As fate would have it, the partnership between Yobi and Microsoft turned out to be synergistic. They made an immediate connection based on their strategic stances on individual privacy and data protection. Yobi prioritizes customer consent regarding data collection while Microsoft doesn’t believe in collecting customer data without permission nor sharing it with third-party companies. So, it was no surprise that when Yobi was struggling to find investment partners, Microsoft was there to take a risk and provide funding through its End Customer Investment Funds (ECIF) program, helping Yobi get off the ground.
Through the ECIF program, Yobi will take the following steps:
- Migrate its environment from AWS to Azure.
- Deploy and scale its offering on Azure using Azure-native services.
- Build integrations and product capabilities across several first-party Microsoft solutions.
Accomplishing the first step will enable Yobi to solidify its implementation strategy and secure scalable infrastructure. The second step includes expanding the distributed processing and training of Yobi’s machine-learning models and raw datasets. Also, Yobi and Microsoft are focused on building a scalable query engine that can support a high throughput of API requests in addition to internal research and development projects. The final step will enable Yobi to broaden its identity graph capabilities so it can process and resolve disparate identifiers in near real time. Once Yobi’s infrastructure is implemented and its offering is scaled, the focus will be to build custom capabilities and integrations for Microsoft products like Xandr, Microsoft Dynamics 365 Customer Insights, and Azure Machine Learning. These new capabilities will support processing a high throughput of programmatic user requests in near real time in addition to constructing an external feature store that gives Microsoft customers access to queries and allows them to retrieve data from Yobi.
Another exciting contribution that Yobi is bringing to Xandr includes the ability to effectively identify and reach prospective customers with a high likelihood to purchase—all at a fraction of the traditional advertising expenditure for a higher return on investment. For Dynamics 365 Customer Insights, Yobi will enable data analysts, data scientists, and marketing teams to integrate and access Yobi’s suite of out-of-the-box behavioral data predictive models and machine learning solutions, offering secure access to data that enhance marketing efforts, customer communications, personalization, and more.
Ensuring Yobi had the proper support it needed to accelerate this journey and meet its goals, the teams behind the Microsoft Global Partner Solutions (GPS) Program and Microsoft Retail & Consumer Product Goods (CPG) Solutions were there to help navigate a complex environment. “Working with the Microsoft GPS and CPG teams was an amazing experience. They provided incredible guidance on how to productize our solution, how to implement and drive adoption of the solution, and overall help with navigating the Microsoft landscape. Truly amazing thought leaders and an absolute joy to work with,” said Max Snow.
Yobi has also achieved Azure IP co-sell incentive status as a partner in Azure Marketplace, meaning Microsoft sales teams are incentivized to sell Yobi’s offering to increase revenue and customer awareness. This is an example of how Microsoft is more than a cloud service provider. It’s a partner committed to helping cloud-born companies build, scale, and grow their businesses.
Accelerating development timelines and enhancing developer productivity with Azure
Beyond its partnership with Microsoft, Yobi turned to Azure to help build its solution. As a cloud computing platform that provides more than 200 products and cloud services in more than 60 regions around the globe, Azure helps solve today’s challenges. Yobi used Azure Databricks as the backbone of its technology because it provided a one-stop shop where Yobi could access a unified set of tools for building, deploying, sharing, and maintaining enterprise-grade data solutions at scale. Yobi was also able to train its machine learning models using Azure Databricks Machine Learning, which provided an integrated machine learning environment that helped simplify and standardize Yobi’s machine learning development processes. Azure Databricks provided the foundation for Yobi’s groundbreaking technology and accelerated its path to production. “Azure Databricks was able to help us unblock a lot of things we were trying to do. We were really impressed with its ability to integrate easily with other tools we were using to help streamline production,” said Frank Portman, Founding Machine Learning Engineer at Yobi.
Yobi also took advantage of Azure Data Factory as its primary choice for triggering ingress, which enabled the company to create data-driven workflows for orchestrating data movement and transforming data at scale. Because the data it collects comes from several different sources, Azure Data Factory was instrumental in helping streamline the data collection process. Combining the power of Azure Databricks Machine Learning with Azure Data Factory, Yobi has built the largest consented predictive behavioral graph in the United States. “Azure Data Factory acted as a layer between us and the rest of the world. It really helped streamline the collection of data from multiple sources and allowed us to focus our efforts on improving functionality,” said Portman.
In terms of storage solutions, Yobi used Azure Blob Storage to store massive amounts of unstructured data for transformation. Blob storage was used for its elasticity and scalability, enabling Yobi to scale operations regardless of dataset sizes. And for security and identity management, Yobi also adopted Azure Active Directory and Microsoft Defender for Cloud.
Focusing on what’s next
With its current offering already returning significant results for several businesses, Yobi is preparing for next steps. The power of generative AI will enable Yobi to continue pushing the boundaries of what customers can do with their technologies using its offerings alongside Azure Machine Learning. The company also has plans to build its own generative enterprise technology using the Azure technology stack, which will ultimately enhance Microsoft offerings. As Yobi continues to grow and its environment becomes more complex, the company will turn to Azure Kubernetes Service to address these challenges using a fully managed Kubernetes service. In the coming months, Yobi hopes to announce a new set of capabilities, including marketing campaign generation and business enterprise process automation, among others. Stay tuned for more.
Azure Data Factory acted as a layer between us and the rest of the world. It really helped streamline the collection of data from multiple sources and allowed us to focus our efforts on improving functionality. Azure Databricks was able to help us unblock a lot of things we were trying to do. We were really impressed with its ability to integrate easily with other tools we were using to help streamline production.
Frank Portman, Founding Machine Learning Engineer at Yobi AI