Drag-and-drop has its place, but Evidence is all about the code
From Tableau and Looker to PowerBI and beyond, there’s an abundance of business intelligence (BI) tools designed to help companies extract valuable insights from their extensive data resources. However, a newcomer has emerged with fresh BI capabilities targeted at data teams with a strong technical inclination.
Founded in 2021 in Toronto, Canada, Evidence made its debut as part of Y Combinator’s summer ’21 startup cohort, offering a modern alternative to well-established BI tools. While BI tools share common features, they often differ in their intended audience. Some prioritize code-based workflows for data ingestion, like Google’s Looker, while others offer a user-friendly drag-and-drop interface for less technical data analysts. Some tools even combine both approaches.
Evidence takes a decidedly code-based approach, allowing teams to construct data products using SQL and markdown. What sets it apart is its open-source nature.
To expand its reach in the commercial market, Evidence recently announced a successful seed funding round and unveiled its premium cloud product for businesses that lack the resources to self-host Evidence.
Moving away from drag-and-drop interfaces, Evidence believes that while these interfaces are useful for data teams in certain scenarios, they lack the precision and granularity that more manual coding approaches provide. According to Evidence’s COO, Sean Hughes, drag-and-drop methods often result in data products that are challenging for end-users to navigate and for data teams to maintain.
Within the Evidence platform, every step, from data acquisition to report definition, relies on code. This approach appeals to modern data teams that prefer to operate similarly to software engineers. It facilitates version control, governance, and collaboration through Git, enabling teams to create a comprehensive and accurate project history. This also allows teams to revisit previous versions, copy and paste code snippets, and repurpose old code.
Hughes highlighted a common issue with most BI tools: they accumulate outdated, broken, or irrelevant reports due to the difficulty of reusing components in different contexts. Evidence, on the other hand, avoids this problem.
Additionally, a code-based approach supports broader continuous integration and deployment (CI/CD) initiatives, allowing teams to work on development versions of projects, run tests on changes, and release updates to production through pull requests.
While Evidence may seem like a response to the growing no-code/low-code movement, Hughes sees it more as an extension of a separate trend in the analytics field. Data teams increasingly want to operate like software engineers, adopting code-driven and open-source products in their data stacks.
Hughes drew an analogy with Squarespace, a platform for building websites. While Squarespace serves many users, it isn’t suitable for every scenario. Similarly, no-code/low-code reporting tools work well for some, but they can be limiting for more technical data teams. Evidence aims to provide an advanced solution for technically-inclined data teams who require capabilities beyond what no-code/low-code tools offer.
One of Evidence’s key selling points is its open-source nature, setting it apart from industry giants like Looker and Tableau. Other open-source tools, such as Lightdash, Metabase, and Apache Superset, offer similar functionality, but they can be self-hosted. Evidence believes that combining an open-source approach with a code-based workflow will make it attractive to businesses worldwide.
After an initial period in early-access mode, Evidence is now expanding access to its cloud service through an invite-based program, supported by $2.1 million in seed funding from A Capital, Y Combinator, SV Angel, and several angel investors. The cloud plan includes a free starter tier with up to five viewer accounts and a team plan priced at $500 per month, offering up to 50 viewer accounts. Customizable plans can accommodate enterprise-grade requirements like single sign-on (SSO) and additional viewer accounts.
