Most organizations are either trying to empower more of their people to use data, or to improve the quality, efficiency, and consistency of the data they’re using. And at any sufficiently large organization, you can bet there are efforts to do both. But solving one inevitably exacerbates the other — wide empowerment leads to duplicated dashboards and inconsistent metrics, while data governance efforts introduce processes and standards that result in long backlogs of data tables and dashboards to be built.
It’s a problem we’ve seen first-hand while building products to help organizations use data for decades, most recently at Looker and Stitch. So today, we’re excited to introduce Omni: the first business intelligence tool that combines the speed and flexibility of analytics tools with the governance of a structured data model. We built Omni to address BI’s entrenched challenges, and we’ve raised $26.9 million from Redpoint, First Round, Google Ventures, and other great investors, to help us achieve our vision.
At the core of Omni is the idea that building a data model should be a collaborative, interactive experience — as easy as building a query or a chart. Our goal is to serve an organization’s entire BI needs, so people won’t be confronted with tradeoffs between freedom and consistency or tension between data and business teams. We’ve made a single product that can start with exploration, graduate to visualization, and then be promoted into a data model for reuse when appropriate.
For as long as organizations have used data, the pendulum has swung between user empowerment and organizational control. Spreadsheets begot monolithic data platforms built around ‘cubed’ data models: Business Objects, Cognos, Microstrategy, OBIEE. Tableau and PowerBI arose from end users’ need to answer their own questions faster than a cube could be built. Looker and dbt reintroduced data modeling to the cloud data stack based on coded data models. Each swing of the pendulum was a reaction to the struggle of the previous generation — "I don’t want to wait on my Business Objects developer to add a field", so "I’ll just use a different tool" become "my Tableau is messy, how do I manage it" and then "Looker is too slow, why can’t I just add this CSV I have?" Rigid centralization leaves people frustrated, while the self-service approach becomes unmanageable.
The reality is every organization requires both the structure of a vetted data model and the freedom to let people explore beyond it. No one wants to invest time to build a data model to answer one-off questions. But when sales needs pipeline data, you want cleanly modeled data, and you may want to materialize the tables to speed up queries.
The growth of the modern data stack has brought huge productivity gains for data workers — modern collaboration, software development lifecycles, fast queries, resilient pipelines — all offered as services that deliver value quickly and manage themselves over time. Cloud data warehouses like Snowflake, BigQuery, Redshift, and Databricks are the centerpiece of this stack — a fast single source of truth that is so easy to manage and scale that analysts and non-technical users can often handle it themselves.
But as the rest of the data stack has become both more powerful and accessible, we have struggled to strike the right balance in the business intelligence layer. Although freedom and control are naturally at odds, we believe they can complement each other when channeled correctly. In fact, we’ve proven that with the BI backbone we’ve built with Omni. Raw data can be modeled more accurately by empowering the relevant business users to do it directly, and new questions can be explored more quickly by building off the data model from previous questions. The right toolset is fluid and fast when it should be, with simple, guided paths to hardening business logic and optimizing queries.
We can’t wait to show you how it works. Reach out today.