The Old Way: Queries


Writing a separate query for each visualization means you are doing a lot manual and repetitive work.

Hard to control

Having duplicate logic embedded in a growing number of visualizations makes it very hard to control and maintain.

Specialists need to serve others

Having to build queries to generate charts and graph is enough of a barrier that most business users will just give up, which means data teams need to service everyone.

Messy and unstructured

When everything is a visualization with no underlying structure, it is very hard to stay organized and accurate across a whole organization.

Silos and confusion

Users are looking at different visualizations in different dashboards and often start to get conflicting views.

The New Way: Metrics

Set up a metric once, power many visualizations

Use your metric definition to power multiple visualizations with far less configuration and no need to repeat yourself.

Easy to control

Centralized metric definitions are a natural fit for control and maintainability. Change a metric definition and keep everything in sync everywhere else.

Natural self-service for everyone

Metrics is an intuitive concept for users of all types, and with built-in visualizations everyone can solve all their frequent needs.

Easy to structure

Your metrics library becomes the basis of a common structure and single-source-of-truth that scales well for larger organizations.


Everyone is looking at the same metrics, in one place, which enables collaboration and a shared understanding of the numbers.

Hang on, is this just for time series?

This is a common reaction since metrics have a required time dimension, but metrics suit all kinds of visualizations and analysis, not just time series. For snapshot-style visualizations (e.g. ranked horizontal bar chart, big number) the time dimension is used to define the period used, like the current month or quarter.

Metrics are super flexible. They can be used to produce all kinds of visualizations, and you can combine multiple metrics for more in-depth analysis.

Is this only for reporting or can I do real analysis too?

The philosophy behind Steep is that everyone should be able to do basic analysis tasks easily, and data specialists and power users should be able to go deeper for more involved insights work. So yes, Steep is built for serious analysis.

In both cases Steep removes the friction of analysis work that takes you out of the flow—writing queries and building graphs. Instead, it allows you to focus on looking at your data from multiple perspectives and tease out insights that help you make better decisions.