Count is a canvas-based BI tool that transforms data teams from support functions into growth engines. In this video, Co-Founder and CEO, Ollie Hughes, walks you through how it works.
Count is a new kind of BI tool based around an infinite collaborative canvas. Unlike traditional BI tools, Count is about far more than building dashboards. It combines all the tools needed for modern data work into one collaborative interface. And this means data engineers, data analysts, and business users can work together in one tool for the first time. Hundreds of teams use count for all their core data workflows. Be that building data models and defining metrics using SQL and DBT, Solving business problems by exploring data using a combination of low code visualizations in Python. Or building company reports which show metrics in new and powerful ways. Count is about far more than just visualizing your data. It gives you a way to visualize your business and see how it fits together. Bring together business context and data to solve problems as a team, and gives your data team the tools they need to maintain strong governance over your data models and metrics. The infinite canvas makes count one of the most powerful interfaces for data exploration. In Count, you explore data using cells. There are SQL cells, Python cells, and visual cells, alongside a number of other low code and filter cells as well. The power of cells is that they're modular. Every cell can use the result of every other cell, allowing you to break your analysis down into parts so you can use the best type of cell for each step of your analysis, and every step you're taking is easy to understand and can be documented using Count's whiteboarding tools. Cells can either run against an external database like the SQL cell or in counts in memory database. Regardless where the data is being calculated, cells update dynamically, allowing you to iterate through your work really quickly. For example, if I add a filter to this query, to a different time period, you'll see that all downstream cells which depend on this query are gonna update automatically to reflect that change. The magic really comes when you start to use the flexibility and space of the canvas to help you think through your work. For example, if I want to under understand not just the number of games played, but the average number of games played per player, I can simply duplicate this visual. And update the aggregation. And now I've got bet both sets of metrics side by side. And if I wanted to take this a bit further and maybe understand how all these visuals are impacted by this filter, I can select all of these objects, drag them below, and update the query here. And then I can see two sets of time periods side by side. What makes this kind of analysis even easier is using Count AI Copilot. You can use it to write queries or to change existing cells. For example, if I come in here and say, add games played to the color axis, It's gonna add in that change in here and make this child look even better. Once I finished my analysis, I can use frames to group my bits of analysis together into sections and keep everything neat and tidy. I can even use templates to help me tell my story quickly. I'm gonna add in a simple slide template, and then I can drag in this visual into this placeholder and update the title. There we go. So at the end of my exploration, not only do I have a really beautiful output, but all my thinking and methodology is clearly documented so I can review it all in one place. One of the biggest benefits of count is how it lets you visualize not just your data, but your business. You can use the canvas to work with your team to map out business processes with live data so that every metric you have is in context and the business performance can be seen in a really clear way. Here's an example of a user onboarding journey through a mobile app. As you can see, every step the user takes is mapped by a screenshot, and alongside that, we have a conversion metric and activity number. There are over twenty metrics in this view, but everything feels really simple and clear, and it's easy to see the area of opportunity that we can go after. Showing metrics in this kind of system way is a really pathway to show your business, and it applies to more than just mobile apps. Our customers use CAT to map out their internal business processes as well as their whole company growth model. For example, here's a way you can show your user retention model, map out a customer acquisition funnel, or show your top level metrics as a metric tree. When you've built a report like this, you can do two things. The first thing is turn this into an operational asset using report mode. Report mode takes frames like this and turns them into a fully operational dashboard that the whole company can see. And filter. Secondly, you can set up an alert. An alert set a schedule, which means this report can be seen on a regular basis in Slack or via email. Currently, this is set up to go out every day at nine AM, and I'm gonna make sure that the product design team can see that in their Slack channel. Presenting your metrics in this way gives your company an amazing way to see what's really going on in their business and work out what really matters. Data modeling and count is something really quite special. The combination of SQL cells in a two dimensional collaborative space is just incredibly powerful. To show you why, here's a typical data model with a hundred and forty nine line SQL query. And we can use the space of the canvas to see this code a bit differently. I can click on split out CTEs, and that explodes the model down into its constituent parts. I can then work through the lineage, see what each bit of code is doing, and start to refactor or debug different parts of my code. Account can even pull code in directly in from dbt core or dbt cloud. This database is connected to DBT Core, and that means I can come down to this particular model and import it into the canvas. And I'm gonna import not just this model, but its upstream parents as well. Seven models in total. As you can see, the code I've imported is actually uncompiled, and I can see the compiled version of the code on the right here. I can edit this code as I want, add in macros, new models as I choose. I'm gonna add a simple limit twenty. And then Kat will compile that code on the fly and show me the downstream effects. And when I'm ready I can simply commit this change to GitHub with a code diff. This makes count incredibly powerful as a DBT IDE or data modeling environment, but it's also hugely powerful for reporting as well. It allows analysts to not only, once they scoped out a report, start to build out the analysis and the business logic at the same time. And when the report is finished, you can simply come to the business logic and import that back to DBT. Just like that. One of the many things which separates Count from traditional BI tools is its combination of flexibility and governance. And Count's metrics layer lets you define your data model and company metrics in one place so everyone in your organization can explore them with complete freedom. As you can see, you can define your metrics in the catalog using code. Metrics can be defined as ratios or complex expressions. You can define relations between different tables in your database and even control the aggregations that are possible on a given metric to stop people doing things which don't make sense. And once you define your catalog, you can access them in a project or canvas just like they're on the database. For example, here's a project built for the product team, and I've given them access to the metrics catalog for product, not to the database that these metrics have been built off just yet. But the product team have used this metrics catalog to build a number of operational reports like this metric tree and this onboarding funnel alongside a number of one off analyses knowing that every metric they're using is governed and self consistent. And if I wanted to do some exploration myself, I could do that either using a new canvas or Explore. Explore is effectively a one cell canvas built for quick explorations. Here I can visualize all the metrics and dimensions I define in my data catalog. For example, I can look at number of users, split that by industry, and paid status, and then filter this for workspaces which have an integration with GitHub. Great. And let's say I wanna start sharing this insight with my team. I can simply go to save as Canvas. Count's now added that insight into a full Canvas so I can start to explore this insight with more flexibility and share what I found and collaborate with my wider team.