Platforms for embedded analytics

There is no one way to deliver data

It has been really fun to get feedback on the book I wrote with Jim Horbury, Embedded Analytics: Integrating Analysis with the Business Workflow. We cover a lot of ground, as you would expect, but one area that always generates a lot of questions is the choice of different embedded architectures, especially for enterprises who want to add analytics to their internal apps. It’s not as if there is just one way to do it, because the specific needs of each organization vary so much. Some organizations may need a simple solution for reporting, while others may need a more powerful and flexible solution for data analysis. The budget and the level of technical expertise available will also affect the choice of an embedded analytics architecture.

Component libraries are a collection of pre-built visualization components that can be embedded in a host application, such as D3.js or Victory. Component libraries are a good option for organizations that need a quick and easy way to embed analytics. However, component libraries can be limited in terms of analytic functionality, such as drill-down or pivot tables, even though they may offer a ton of chart types.

Even now, a high proportion of business needs can be met with simple reporting. Enterprise reporting platforms are designed to provide a standardized view of business data. These platforms can be a good option for organizations that need a straightforward solution for reporting and analysis and most are embeddable in some way. Most enterprises will already have a platform deployed and running, so if that can be embedded, it can be a simple choice for simple embedding. But it’s not going to give you much analytic insight - just a predetermined view of what needs to be presented.

Business intelligence applications (think of PowerBI, Tableau and Qlik as the market leaders) typically include a wide range of features for reporting, analysis, and dashboarding. These applications can be a good option for organizations that need a powerful and flexible solution for data analysis. However, business intelligence applications can be more complex to manage and to learn than reporting. For example, they may require more specialized knowledge of data analysis, or they may have a more complex user interface. This can make it difficult for business users to use these applications without IT assistance.

There are a good range of purpose-built embedded analytics platforms such as SiSense, GoodData and Logi Analytics, designed specifically for embedding analytics in a host application. These platforms typically offer the range of business intelligence features as well as development and management features for integration with the host application. Purpose-built embedded analytics platforms can be a good option for organizations that need a comprehensive and easy-to-use solution for embedded analytics, especially if you are developing your own in-house applications.

Embedded self-service is a type of embedded analytics that allows business users to create and modify their own analytics without the need for IT assistance. This can be a valuable option for organizations that want to empower business users to make data-driven decisions. Embedded self-service may increase user adoption of exploratory analytics, not just reporting, and improve the quality of user engagement and decision-making. However, it can be difficult to ensure that the analytics are accurate and timely, or that the processes for sharing and collaboration are secure, and that the volume of self-service requests can be managed. And, like the BI applications which can be embedded, self-service requires a lot more training than a simple reporting system or visualizations.

Organizations that are considering embedded self-service should carefully consider their needs and requirements before implementing a solution. Key features of embedded self-service include ease of use, flexibility, security, and management.

Factors to consider when evaluating embedded self-service solutions include the skills and data literacy of business users, the capabilities of your existing data infrastructure, and - of course! - the budget.

We cover a lot of this in the book and we hope there is enough information to help with these decisions. Let us know! And if you do read the book, let others know through reviews too. Thanks!