Mason is shutting down
Mason is shutting down
Mason is shutting down
Aug 28, 2023
Mason is shutting down due to a lack of traction. Below is a summary that outlines why we founded Mason, what we built, and the lessons we've learned.
Mason is shutting down due to a lack of traction. Below is a summary that outlines why we founded Mason, what we built, and the lessons we've learned.
When we founded Mason in the fall of 2021, it stemmed from a shared pain. The three of us, coming from backgrounds in engineering, design, and product management, had spent nearly a decade wrestling with data analytics tools that felt more like obstacles than allies.
Our experiences from startups, followed by our tenure at Shopify, had taught us the value of speed: ship fast, learn quickly, and iterate at lightning speed. For this reason we needed to keep data analytics simple and optimize for flexibility. We wrote SQL, visualized the result, and shared the insights with our teams.
However, the best tools on the market did not support this way of working. They were designed for centralized data teams, with high levels of expertise, building data models and dashboards for others to consume. They were not optimized for the type of highly specific, ad hoc questions that we had seen deliver so much value.
Based on our own experiences, and after conducting over 100 interviews with people from similar backgrounds, we identified two significant problems with the tools we had been using:
They quickly turned into trash piles of abandoned reports where teams struggled to re-use prior work.
They lacked support for collaboration, so teams spent time trying to debug queries on Slack.
When we founded Mason in the fall of 2021, it stemmed from a shared pain. The three of us, coming from backgrounds in engineering, design, and product management, had spent nearly a decade wrestling with data analytics tools that felt more like obstacles than allies.
Our experiences from startups, followed by our tenure at Shopify, had taught us the value of speed: ship fast, learn quickly, and iterate at lightning speed. For this reason we needed to keep data analytics simple and optimize for flexibility. We wrote SQL, visualized the result, and shared the insights with our teams.
However, the best tools on the market did not support this way of working. They were designed for centralized data teams, with high levels of expertise, building data models and dashboards for others to consume. They were not optimized for the type of highly specific, ad hoc questions that we had seen deliver so much value.
Based on our own experiences, and after conducting over 100 interviews with people from similar backgrounds, we identified two significant problems with the tools we had been using:
They quickly turned into trash piles of abandoned reports where teams struggled to re-use prior work.
They lacked support for collaboration, so teams spent time trying to debug queries on Slack.
A retrospective
We set out with a vision to build a product for fast-moving product teams where analysts, engineers and product managers could collaborate and answer their own questions with SQL. From the moment we wrote our first line of code to the present, we hit several key milestones:
We built a data tool available both on the web and as a Mac app, which featured a collaborative SQL editor that learned from every query, a shared query library, realtime dashboards, and much more.
After 6 months of development, we launched our alpha, bringing on on our first customers—some of which stayed with us to the end.
Our waitlist grew organically to 1500 organizations, driven primarily by short video demos of our product that was posted to Twitter and LinkedIn.
We brought on a team four talented engineers, enabling us to ship new features weekly.
In January 2023, we shipped an AI experience and repositioned our product to an AI-powered SQL editor.
Yet, despite all our efforts, Mason failed to find success.
The problems we set out to solve were primarily evident to larger teams already using a data tool. Unfortunately, Mason was neither 10x better than their current solutions nor seen as a complement.
For a data startup, it's crucial either to offer a product that complements existing tools or to optimize for other startups where there’s low switching cost. Interestingly, we learned that startups often seek a data tool that does it all, and it was hard to identify a single job Mason could excel at to become the preferred choice.
Collaboration was our unique value proposition. We experimented with multiple collaborative features, starting with "pull requests for data" during our alpha launch and progressing to a more Figma-like experience with a multiplayer editor, code comments, and a shared query library. However, these features failed to resonate with startups, where the number of people writing SQL is limited.
A retrospective
We set out with a vision to build a product for fast-moving product teams where analysts, engineers and product managers could collaborate and answer their own questions with SQL. From the moment we wrote our first line of code to the present, we hit several key milestones:
We built a data tool available both on the web and as a Mac app, which featured a collaborative SQL editor that learned from every query, a shared query library, realtime dashboards, and much more.
After 6 months of development, we launched our alpha, bringing on on our first customers—some of which stayed with us to the end.
Our waitlist grew organically to 1500 organizations, driven primarily by short video demos of our product that was posted to Twitter and LinkedIn.
We brought on a team four talented engineers, enabling us to ship new features weekly.
In January 2023, we shipped an AI experience and repositioned our product to an AI-powered SQL editor.
Yet, despite all our efforts, Mason failed to find success.
The problems we set out to solve were primarily evident to larger teams already using a data tool. Unfortunately, Mason was neither 10x better than their current solutions nor seen as a complement.
For a data startup, it's crucial either to offer a product that complements existing tools or to optimize for other startups where there’s low switching cost. Interestingly, we learned that startups often seek a data tool that does it all, and it was hard to identify a single job Mason could excel at to become the preferred choice.
Collaboration was our unique value proposition. We experimented with multiple collaborative features, starting with "pull requests for data" during our alpha launch and progressing to a more Figma-like experience with a multiplayer editor, code comments, and a shared query library. However, these features failed to resonate with startups, where the number of people writing SQL is limited.
Code comments, together with our multiplayer editor, enabled teams to debug queries in realtime
We received great feedback from larger organizations, but it turned out that their pains were not big enough to get them to switch to a new tool. Selling to these teams would require significant investment in security, infrastructure, and sales, none of which would differentiate us from the competition.
Our “learn from every query” feature set got the most excitement from our users. With each query, Mason became smarter and guided users to relevant data and its usage. However, our recommendations seemed less impressive with AI's rapid advancements.
Based on these learnings, we were excited to introduce AI features in Mason to bring a single-player value and shorten the time to the aha-moment. While we had high hopes for our AI Assistant, it didn't meet user expectations. Existing users, proficient in SQL, weren't interested in a text-to-SQL solution, and AI currently lacks the precision to replace SQL for people who don’t know how to code. Our AI features ended up pleasing no one.
In summary, our unique features weren’t delivering enough value to our users. Mason was not differentiated enough to compete with the established tools in our category.
We received great feedback from larger organizations, but it turned out that their pains were not big enough to get them to switch to a new tool. Selling to these teams would require significant investment in security, infrastructure, and sales, none of which would differentiate us from the competition.
Our “learn from every query” feature set got the most excitement from our users. With each query, Mason became smarter and guided users to relevant data and its usage. However, our recommendations seemed less impressive with AI's rapid advancements.
Based on these learnings, we were excited to introduce AI features in Mason to bring a single-player value and shorten the time to the aha-moment. While we had high hopes for our AI Assistant, it didn't meet user expectations. Existing users, proficient in SQL, weren't interested in a text-to-SQL solution, and AI currently lacks the precision to replace SQL for people who don’t know how to code. Our AI features ended up pleasing no one.
In summary, our unique features weren’t delivering enough value to our users. Mason was not differentiated enough to compete with the established tools in our category.
The AI Assistant enabled users to generate and edit SQL using plain English
In closing
The decision to shut down Mason was not made lightly, especially when our team has put in countless hours, nights, and weekends into this product for almost two years. Throughout our journey, we made several big swing attempts to pivot our product and refine our positioning, but ultimately, we didn't find success. Our team, more than anyone, wanted to see Mason thrive.
To our investors, and everyone who supported and used Mason, your feedback and encouragement meant the world to us.
While our chapter is closed, we believe that the future of data is bright. AI will eventually be able to provide the step function change that will make data accessible to everyone. One can only hope it comes with an exceptional user experience.
In closing
The decision to shut down Mason was not made lightly, especially when our team has put in countless hours, nights, and weekends into this product for almost two years. Throughout our journey, we made several big swing attempts to pivot our product and refine our positioning, but ultimately, we didn't find success. Our team, more than anyone, wanted to see Mason thrive.
To our investors, and everyone who supported and used Mason, your feedback and encouragement meant the world to us.
While our chapter is closed, we believe that the future of data is bright. AI will eventually be able to provide the step function change that will make data accessible to everyone. One can only hope it comes with an exceptional user experience.
In closing
The decision to shut down Mason was not made lightly, especially when our team has put in countless hours, nights, and weekends into this product for almost two years. Throughout our journey, we made several big swing attempts to pivot our product and refine our positioning, but ultimately, we didn't find success. Our team, more than anyone, wanted to see Mason thrive.
To our investors, and everyone who supported and used Mason, your feedback and encouragement meant the world to us.
While our chapter is closed, we believe that the future of data is bright. AI will eventually be able to provide the step function change that will make data accessible to everyone. One can only hope it comes with an exceptional user experience.