Certainly. Humblytics is a primary analytics tool for Webflow, co-founded by Amir and me, who previously ran a full-service B2C agency. Our challenges were twofold: integrating effective analytics for performance tracking and identifying areas for improvement.
Google Analytics was the traditional choice, but it posed difficulties due to overwhelming data and complex setup via Google Tag Manager. With the rise of no-code tools like Webflow, analytics remained a pain point. To address this, we created Humblytics, offering effortless setup by authenticating SSL through Webflow for automatic script installation. This verifies website ownership and tracks events like form submissions and clicks instantly.
I had some experience with basic SQL and MongoDB setups, but dealing with something at this scale was new to me. Amir and I realized we didn’t have the budget to hire data engineers or experts in the field.
Then, I recalled my time at Interspace, a company specializing in indoor intelligence. They tracked millions, even billions, of events with just a small team of 10, so I reached out to them for advice. They recommended Timescale for their database needs.
After Googling Timescale, I was shocked by how incredibly user-friendly and cost-effective it is to get started.
Our top priority was ensuring platform scalability during popularity surges and client traffic spikes. We wanted a managed service to avoid manual AWS maintenance, while ensuring high reliability.
Our users demand Google Analytics-level reliability, even during rapid growth, expecting uninterrupted performance and unlimited storage. Disruptions or system downtime for paying users could lead not only to data loss, but also erode trust in our platform.
The initial test involved writing insert statements, followed by creating a basic proof of concept. We then funneled real web analytics data into the database and assessed service utilization for future vertical scaling.
Unlike many other managed tools that add unnecessary complexity, Timescale made the process smooth and intuitive, particularly for those familiar with SQL. Our testing began in April, and we launched publicly in May or June.
Signing up with Timescale provided immediate access to a connection string containing the necessary detail—username, password, and database name. It was straightforward, requiring no extra steps. Creating my first table, a hypertable, was another smooth experience with no unexpected surprises or steep learning curves. The dashboard, like the Explorer for table navigation and metrics/logs just made sense. It was all very intuitive.
The 30-day free trial allowed us to thoroughly test the platform to ensure it met our needs. The managed aspect of Timescale saved us from opening an AWS account solely for database setup.
Timescale provides us with scaling insurance, so we don’t have to worry about performance issues, storage limitations, or data loss negatively impacting our users, which has been crucial for maintaining growth and user trust. Features like read scaling, high availability, and data tiering, addressed any scalability concerns and helped us make the decision to build Humblytics on Timescale.
Our current setup isn't too complex—we run a Node.js server on Heroku, which passes through Cloudflare to reach our Node.js server. This server then sends data to Timescale.
Two notable features that greatly benefited our workflow were compression and aggregates. Compression proved valuable in reducing storage costs, while aggregates streamlined data management by summarizing it. Instead of storing multiple instances of the same event, like button clicks, aggregates facilitate quicker querying for these counts and streamline our data-handling process.
Currently, we ave around 1 GB of data in Timescale, with about 50 users and 100 websites connected. Our goal is to grow significantly within the next year, aiming to scale to thousands of users and tens of thousands of websites. It’s hard to say if the data is going to scale linearly, especially if a few websites generate a lot of traffic. We still haven’t made use of all of the different tools you offer to scale the performance.
Timescale is perfect for applications that deal with high data loads and need efficient ways to ingest, query, and manage time-based information. With Timescale, you can tackle complex engineering challenges related to time series data without the headache
Maxim Antipin, Humblytics, Co-Founder and CEO