loading...
Cover image for How Termius developers make self-serve data-driven product decisions

How Termius developers make self-serve data-driven product decisions

avohq profile image Avo Originally published at avo.app ・5 min read

“We believe in data-driven growth that doesn’t just come >top-down, so we build product thinking into our processes. >Avo’s integration with Mixpanel’s Lexicon has been integral >to realize this vision”.
– Kirill Yakovenko
‍Product Manager

Termius is the SSH client that works on desktop and mobile. It provides software engineers with a complete command-line interface (CLI) accessible on any device, allowing system administrators and network engineers to access their terminal from anywhere. It’s used by over 45,000 engineers daily, and is famously installed on Steve Wozniac’s iPhone.

Kirill Yakovenko is a product manager heading up analytics quality at Termius. “I think about what problems our users are solving day-to-day, while making sure our data is correct,” he says. The Termius culture empowers anyone on the team to develop their own analytics to better understand user behavior within the product.

“My role is to support this democratization of our product development, by peer reviewing suggested analytics updates and upholding our data quality standards,” Yakovenko explains. This includes assisting his team in designing and implementing high quality product analytics, while making sure they’re not already being tracked elsewhere.

Termius’s tech stack is JavaScript for web and desktop, Python on the backend, Java for Android and Swift for iOS. The team recently added Mixpanel to their product analytics stack.

Termius' challenge from day one was how to keep analytics consistent across the multiple platforms they serve. Each platform has numerous applications and different codebases, and analytics implementation is error-prone enough even on a single codebase.

How Termius Builds The Best Customer Experience Using Product Analytics

Termius is all-in on data-driven product growth. They’ve built a feedback loop between qualitative user feedback, verifying their learnings with product analytics – and vice versa. Yakovenko explains that the purpose of this approach fills in limitations of either approach.

First, the data can verify if users’ actual behaviour matches up with what they believe their behaviour to be, so insights from user interviews can be quantified with analytics. And for the reverse he adds, “it’s easy to get the numbers wrong.” “When we get insights from the data, we go back to the customer to verify that the data was right.” At the end of the day, it’s about reaching the best possible understanding of how Termius is used. “The source of the data doesn’t matter, we want the truth.”

Here’s an example that relates to Termius’s northstar metric, ‘Number of Connections to Servers.’ When the team discovered a seemingly mild issue in establishing such connections, they adjusted their analytics to measure what problem users were experiencing at a high level. “The data revealed a critical issue for our target audience,” states Yakovenko. This data-driven approach saved the team and their customers a lot of time and efforts otherwise spent on user feedback sessions.

“Before Avo, We Couldn’t Trust Our Data”

Keeping analytics synchronized across all of Termius’s platforms was an impossible feat before Avo. When writing analytics code, it’s easy to produce minor issues, such as different names or inconsistent casing for the same event, that escalate into a mountain of data-debt, rendering the data you do have noisy and unreliable. What seems like a trivial error, isn’t.

“Each mistake undermines our trust in our data”
Kirill Yakovenko.

“To combat this we needed a place for collaboration, where we could define all events with all possible properties, values, descriptions and other event meta-data”. Since using Avo,“ our trust in our data has increased.”

The problem wasn’t only that lack of trust. “The problem with tracking mistakes,” says Yakovenko,”is that each fix takes time. It might take a month to roll out a fix for a single issue to all our applications and users.” Fixing analytics issues didn’t just waste developer hours on the fix itself, it also delayed product decisions because insights are crucial to making the right product decisions.

“The human hours spent per analytics implementation has >reduced substantially, saving a huge number of hours while >fixing differences between platforms and commonplace >mistakes”.

“Avo’s biggest impact was in reduced overheads on analytics. We’re not only saving time for developers and product managers on designing, implementing and verifying analytics – but developers are also relieved that they no longer have to monitor the analytics after every release, to make sure the tracking didn’t break.”

Modern Developers Have To Have Product Skills, And Avo with Mixpanel Helps Unlock That##

Mixpanel is a leading product analytics platform that allows companies to answer questions about user behavior in seconds, without analytics expertise. This enables product development teams to measure what matters quicker, make decisions to improve experiences, and build better products through data. Over 26,000 companies including Uber, Expedia, Twitter, and Ancestry use Mixpanel to understand how their users convert, engage and retain.

Mixpanel and Avo partnered in 2020 and the integration built between the tools allows tracking descriptions, properties, and categories to be easily published to Mixpanel’s Lexicon with a push of a button. By customizing which events in Avo are published with their tracking plans, teams can understand more easily than ever what their data means. This organization and cleanliness has given Termius a restored trust in their data.

“To power their data-driven product growth, Termius needs to >fully trust their data“

“The value of having the Avo descriptions in Mixpanel is immense,” says Yakovenko. “Having the context where needed and removing any guesswork around the events saves time and effort for our developers and other product managers. Those who aren’t in the weeds of defining analytics can now see in Mixpanel what the events mean.”

“That is really important to us because we believe in data-driven growth that doesn’t just come from the product manager. Instead we build product thinking into our processes” Yakovenko explains. “The same goes for meetings; when we look at the data, the context is right there. Descriptions allow us to see immediately what is going on on our graphs.”

The long-term benefits of democratizing product thinking are clear. “We need everyone on the team to understand what we’re trying to achieve, not just with our code but what our customers need. When the developers ‘get it,’ they generate good ideas, challenge us as managers and challenge us on the product. Their suggestions are richer and provide better solutions for our customers when they understand the business and our goals.”

“Some developers just care about the technical work and don’t want to understand business, goals and customers. But in modern product development and for a modern start-up, you need all hands on deck to move your business forward: combine skills, combine ideas from everyone on the team” says Yakovenko, “and then: Prioritize. Analyze. Iterate.”

Recommendations

For anyone who’s looking to start standardizing their analytics and streamlining the implementation process, Yakovenko has the following insight: “Avo offers real benefits, especially if you are cross-platform or have a big team.”

Posted on by:

avohq profile

Avo

@avohq

Avo is a next-generation analytics governance tool improving how product managers, developers, and data teams plan, track, and govern product analytics across organizations. Learn more at avo.app

Discussion

markdown guide