Empowering Brazil’s Creator Economy

User demand for insights

When Caio Pedroso joined the Hubla team as the sole Data Engineer in May 2021, he was tasked with reigning in all of the data Hubla collected on its creators so that the Hubla team could build new features to best support them. It didn’t take long for him to realize the insights he was gleaning would be valuable to the creators themselves.

Initially, Caio and team tried sharing their users’ data with them via Excel-style spreadsheets and then by embedding a Metabase dashboard behind a feature flag. Caio quickly learned two things:

1. Their user base would require a much higher quality UX
2. Their user base needed to be able to drill down & filter without waiting on charts to reload

While the load performance and user experience didn’t meet the Hubla team’s standards, Caio got the early validation that the insights were valuable to creators. Hubla had opened their eyes to what kind of information was available and now they wanted even more. The Hubla team received messages from creators like,

In his quest to find the perfect solution, a friend mentioned Patch to Caio. Looking into it, he saw Patch checked all of the team’s initial requirements:

• Design and customization: Hubla’s founder is a Product designer by background so the creator experience is always a top priority for the team. This meant the dashboard had to be native to the Hubla platform and design system.

• Actionable insights: Many of Hubla’s creators are full-time content creators and they trust Hubla to help them manage their business. They take the ability to increase the amount of money they earn month over month seriously, so insights needed to be interactive with filters, segmentation, drill downs, etc.‍
• Fast: A major part of the user experience is page load performance, especially in the face of interactive elements like filters or dropdowns that change the data that is displayed.
• Integration and developer workflow: Hubla’s data resides in BigQuery, where it’s modeled with dbt. They also used TypeScript on their backend and React frontend components. It was vital that any solution could allow engineers across teams to collaborate effectively.
• Expertise and support: Hubla would not be able to dedicate a large internal team to this project, so it was vital they’d have the expertise and support of the Patch team though to its completion.

Boosted by strong customer demand for their data and given the green light to start an initial project, Caio and his team of four — one Technical Lead on the backend and three Frontend Designers — got started.

Delivering fast, interactive analytics with zero infrastructure setup

The dashboard they created is highly interactive and customizable, mixing various charts with tables of metrics. After all, creators seek insights to help them drive their business forward.

The dataset contains categorical data on creators and their product offerings as well as large time series data on video engagement and product sales. Every visual element has filters, segmentation, and adjustable time grains. Users can even select different metrics to display. When a user changes a filter or time grain, queries still load with sub-second latency from Patch’s highly scalable, distributed memory cache.

Caio uses dbt to model their data in BigQuery. The final models are loaded into Patch, which keeps the data fresh by detecting changes in the underlying source tables. Then, their TypeScript application queries Patch. Hubla extended their UX design system for the new dashboard, so the final step was to plug the Patch query response data into their new React components.

Once they were in production, the team was able to easily update source table schemas without risking downtime by using Patch’s versioning system. This workflow enables a data engineer like Caio to update dbt models on his own timeline. Meanwhile, the application continues to query the original version until the application developers have properly tested and promoted the latest to production.

Reflecting back on the time they spent implementing Patch, Caio says, “I took 20 days off and I was nervous because I knew that if there was any problem with the project, it might become stalled if my team didn’t reach out to Patch for help — despite your team’s support, I knew they might feel too shy simply due to the language barrier and because I had led on it so far. However, while I was away, there wasn’t a single problem.”

He goes on, “It was really amazing to go from the beginning of the project with my technical lead asking me, ‘Okay, I understand Patch, but is it really necessary?’ to the end and him saying, ‘We would not have done anything without Patch.’ It was truly a great experience from a product standpoint and a team support standpoint.”

Creators bringing more business to the Hubla platform

Approximately three months after beginning the project, Hubla fully released its dashboards to all of its creators. Immediately, they saw a 50% engagement rate, much higher than anticipated. It came as an even bigger surprise to see that the monthly retention rate of its creators engaging with their dashboards has held steady at close to 60%.

“People are really using this feature,” says Caio. “In fact, one of our most active creators reached out to tell me that he’d taken a weekend to go through his dashboard and find insights on how to improve his business. He’s spent many years working in digital products and has experience on all of the creator platforms we have here in Brazil. He said he’s never seen a dashboard with the amount of information we are able to show him.”

“It can be really difficult to show Return On Investment (ROI) with data but this (Patch) dashboard shows that if we continue to develop and show our users their data, they will bring more product which will bring more money to us. It’s basically everyone’s dream to show ROI this easily.”
– Caio Pedroso, Hubla Data Engineer

What’s next

This is just the beginning of the Hubla + Patch story. The team is now all in with Patch and planning another dashboard focused on engagement to help their users determine which videos their followers are engaging most and least with and what topics they’re discussing in various chats.

Despite the fact that the upcoming engagement dashboard they’re planning is, in Caio’s words, “a monstrosity — very big”, he expects it will take a team of three (himself, a Technical Lead, and a Product Manager) just two months to complete it thanks to the charts they built for their first dashboard project. “We just need to do the integrations — and Patch provides us a lot of those integrations — so we’ll just need to normalize a little bit of data. The speed to development time is very important to us.”

As for other plans they have to use Patch, stay tuned because they’re keeping them close to the vest for now.

Get started with Data Packages

What started as a simple mission to bring data to their creators now has the Hubla team planning various ways they can empower them to grow their communities in ways they’d never imagined possible. It’s this out of the box thinking that we are inspired by and why we’re passionate about bringing Data Packages to teams everywhere — there are limitless ways it can transform the work you do.

If you’d like to try Data Packages, or if you’re interested in Patch’s managed infrastructure for versioning, access controls, and query performance, let’s chat.