What is a data canvas?

Google launched the BigQuery Data Canvas in pre-release at Google Cloud Next a few weeks ago. We’re excited to see a tech giant like Google finally see what we’ve believed for a while - the canvas is the interface of the future for data.

But what is a data canvas? Why is it better? and How do the Count canvas and Google Data Canvas differ? Let’s dig in…

User interfaces matter

Anyone who’s switched from a keypad (e.g. Blackberry) to an iPhone knows that user interfaces matter. To move from a mini joystick and a dozen or so buttons to a fully tactile, responsive touchscreen was a key innovation that launched the mobile web and a trillion-dollar market with endless possibilities and applications.

Or to put it another way, though your last Blackberry had an internet browser and some simple application utilities the full potential of these technologies was completely held back by the way the user had to interact with them.

User interfaces bound function.

This is as true in the data industry as it is in mobile technology. Dashboards, pivot tables, and notebooks are the interfaces that we still use. Despite the new technologies, in-memory compute engines, or coding languages data vendors talk about, these interfaces are broadly the same as the ones we had over 20 years ago. (If you don’t believe me look at some screenshots of IBM Cognos and tell me it doesn’t look like PowerBI or any other dashboarding tool).

What’s also true is that despite all these new technologies, organizations still suffer from the same challenges - how to tap into the value of data, and make their teams more data-literate.

This is because the methods in which we interact with the data haven’t broadly changed. Or to put it another way, many data challenges are about user interfaces.

The canvas is the interfacial leap data needs

The challenge of interfaces was something my co-founder Oli and I realized back in 2019. A year ago we launched a collaborative notebook platform and realized how limiting this was for both the analyst and the business users they work with.

Around this time, we were increasingly using collaborative canvas tools for our internal workflows. We saw solutions like Figma for designers provided a space where designers could think and work unconstrained but also bring their wider teams into their workflows. We realized these were the interaction paradigms that data desperately needed too.

Canvas interfaces are powerful interaction paradigms for several different reasons:

1. Workflow optimization: The canvas is an interface optimized for complex workflows. A canvas can contain objects of various types and show their connections. Moreover, these objects naturally change and flex as you pan around the canvas and interact with them. This means the workflow dictates the UI, not the other way around.
2. It creates a place to think: The infinite canvas allows users to use space as a means of organization. Fragmented information and concepts can be grouped, and half-finished ideas can be parked to be returned to later, simple things like comparing two versions of the same thing are as easy as zooming out. This is not only a much more natural and fluid way to work, but it also dramatically reduces the user’s cognitive load (ie the amount of information they have to keep in their head as they move between different interfaces to complete a task).
3. It allows true collaboration: The canvas provides the space for users to genuinely collaborate without conflict. Users can work side by side, sharing their thinking and methodology as they solve a problem. As we move into the world of AI this also makes the canvas perfect for working with AI assistants.

Google’s BigQuery Canvas - An early start

We’ve been on this journey with the canvas for over 3 years and are fortunate to have 100s of customers using the canvas to work more efficiently and drive greater value from data. So we’re excited to see Google starting to make strides into the same future recently.

We believe the canvas is the interface of the future for data and that means we, perhaps surprisingly, want and expect other vendors to see the benefits and help educate the market.

The benefits of the canvas for working with AI are certainly clear to Google as this is currently the primary use case for their canvas. However, their canvas is missing several very important features that we know are essential for the canvas to drive true value:

1. Google’s canvas isn’t collaborative meaning users can't work together.
2. Their canvas architecture isn’t built for scale. In Count we regularly see users creating 100s of separate SQL, python, and visual cells in one canvas something that Google’s canvas would really struggle with.
3. Google’s canvas has neglected the primary user. One of the reasons tools like Figma are so successful is how they not only create a collaborative space but also absolutely nail the experience for the power user. In Count, this has meant building a tool not just for SQL but for Python, advanced visuals, a dbt integration, and 1000s of other little features that delight our analyst and data engineer users each day. We don’t have an NPS of 70 for nothing…

Looking into the future

There’s nothing like having a heavyweight like Google get stuck in to give you a sense you’re on the right track. However, if there’s one negative of the canvas paradigm, it’s that it is a much less forgiving interface if you don’t get it spot on, which our blood, sweat, and tears can attest to. So though we hope the data canvas category builds, we don’t recommend the wider market to take on the task lightly!

‍