How Ovoko eliminated its analytics bottleneck and achieved 10x faster time to insight

Ovoko is a two-sided marketplace for secondhand car parts, based in Vilnius, Lithuania and operating primarily across Europe. The platform connects dismantlers and scrap yards on the supply side with buyers ranging from bulk business purchasers to individual consumers looking for a specific spare part. With over 30 million parts listed on the marketplace, Ovoko has to make it easy for people to find the right part, trust the listing, and complete a transaction. 

As Ovoko scaled, its analytics team became the bottleneck for product decisions. Looker, the company’s BI tool, was too slow and technical for non-analysts, leaving product managers, designers, and engineers dependent on the data team for answers.

Ovoko adopted Mixpanel for self-serve product analytics, rolling out structured training and a certification program built around their own data. Today, teams analyze user behavior independently, time to insight has improved 10-15x, and Mixpanel insights have directly influenced product decisions—including reversing a major feature that negatively impacted the funnel.

Every product question waited in line behind the analytics team. Looker handled operational reporting well, but for deeper product analytics it was too slow and technical for most people to use:

• No self-service for non-analysts: Product managers, designers, and engineers couldn’t reliably get their own answers. Every question required an analyst.
• Slow query times: some Looker queries sometimes took 30 minutes or more, making real-time exploration impossible.
• Multiple sources of truth: Since Looker was cumbersome, people exported data to Google Sheets and worked from local copies. Metrics diverged across teams.
• Limited behavioral visibility: Understanding funnels, user journeys, and drop-off points in Looker was difficult enough that most teams simply didn’t try.

Ovoko adopted Mixpanel as its product analytics layer, keeping Looker for BI and operational reporting while using Mixpanel for all behavioral and product event data.

Key changes:

• A structured rollout with real data: The team started with a pilot covering one domain and a handful of product managers, first running a data cleaning exercise and establishing event definitions before opening access more broadly. Company-wide training was built around Ovoko’s own data, not generic demos.
• A two-tier certification program: Occasional users learn to read and explore; power users (mainly product managers) learn to build and own their dashboards. The data team compiled a whole presentation of unique Mixpanel use cases they’d discovered while learning the platform, now folded into internal training.
• Lexicon as a single source of truth: Standardizing metric definitions in Lexicon directly addressed the fragmented data problem that had built up over years of Looker workarounds.
• Mixpanel and Looker in tandem: Mixpanel answers “why it happened” behavioral questions; Looker answers “what happened” in aggregate. Both stay in the stack for different jobs.

1. Analysts are no longer the bottleneck

Product managers, designers, and developers now analyze data independently. The analytics team’s role has shifted: people reach out not for the numbers but “in case they found something that doesn’t make sense,” as Giedrė Oksaitė, Senior Data Analyst at Ovoko, puts it. Already, 23-25 users outside the data team are active in Mixpanel every week, even before the rollout is complete across all product teams.

“We’re seeing way more independent analysis and insights being generated without the bottleneck of an analyst. Stakeholders now find patterns and come to us to validate whether what they saw was correct, as opposed to coming with a question they need us to investigate,” Sabas says.

2. 10-15x faster time to insight

Queries that used to take 30 minutes in Looker now take seconds. For deeper trend analysis, what previously took a full day now takes an hour or less. “Previously it took a day to drill down into some weird trends by writing SQL,” says Nojus Esteris, Senior Product Data Analyst at Ovoko. “Now it’s an hour maximum.”

“I used to be a hardcore SQL user. Even if there was a dashboard, I’d prefer to write a query and control all the parameters. Now it’s much quicker and easier to just click a few things and get a shareable, reusable report,” Oksaitė says.

3. Bot attacks caught before they skewed the numbers

When unusual traffic patterns started appearing, the team used Mixpanel to identify and quantify bot attacks. Before, those intrusions would have inflated KPIs across the board, with no clean way to distinguish real user behavior from automated traffic. Now they can present reliable data to leadership with more confidence.

4. A product decision that surprised everyone

One of the clearest demonstrations of Mixpanel’s value came when a feature that had required significant investment turned out to have a serious downside in the funnel data. The finding was so counterintuitive the team checked the numbers six times before acting on it. They ultimately reversed the feature decision. “Without that data,” Sabas says, “we would have had no way to catch it.”

5. Data adoption in unexpected places

Perhaps the most telling sign of how far the culture has shifted: designers are using Mixpanel. They check feature adoption rates and track whether users are dropping off the flows they designed. “I noticed that designers are doing that themselves,” Esteris says. “Which makes me quite happy that I don’t need to help them every time.”

Building the culture, not just installing the tool

Ovoko didn’t just turn Mixpanel on and walk away. Giedrė, who spent much of the past seven months focused on the Mixpanel rollout, describes the work as a data democratization project. The team started with a pilot, cleaned up legacy events and data models first, then ran structured training built around Ovoko’s actual data.

They built a two-tier certification program: occasional users who need to read dashboards and explore independently, and power users who build and maintain them. Mixpanel University courses were woven in selectively, matched to each tier rather than assigned as a blanket requirement. “They’re definitely helpful,” says Giedrė. And Nojus built his own internal playbook: a presentation of unique Mixpanel use cases he’d discovered while learning the software, now folded into training for new users.

“Teams depend on analyst availability so much less now,” Oksaitė says. “With proper training, teams can do much more themselves than they used to.” The results of that investment are visible in the numbers. Without counting the data team, Ovoko already has 23-25 weekly active Mixpanel users. 

When the data catches what nobody expected

One story stood out from Ovoko’s first months with Mixpanel.

When unusual traffic spikes appeared in the data, Mixpanel made it possible to quickly isolate and quantify the automated traffic from real user behavior. That distinction matters: without it, inflated numbers would have made their way into KPI reviews and board meetings, distorting decisions. Now the team can identify the intrusion, quantify the impact, and present clean data with confidence.

Data adoption beyond the usual suspects

One of the clearest signs of a genuine data culture shift is who’s showing up in Mixpanel.

Ovoko designers are using Mixpanel. They check how much adoption their features receive, and whether users are dropping off the flows they designed. Developers use it to validate their own tracking implementations, checking that events are firing correctly without needing to ask the data team. One product manager from a SaaS product team used Mixpanel to decide whether to keep or shut down an underperforming product area. The insight took 30 seconds.

“Previously, all roles were dependent on analysts,” Oksaitė says. “Now product managers, designers, and developers reach out to us not to get numbers, but when something doesn’t make sense.”

Ovoko is still mid-rollout. Not all product teams have Mixpanel fully implemented yet, and the team is working toward complete event coverage across every domain. Ovoko has already connected marketing attribution data with product events, giving the team visibility into not just which channels drive traffic but which drive users who actually convert and transact.

Experimentation is the other big item on the roadmap. Ovoko currently runs experiments using in-house custom tooling, and the overhead creates reluctance to test smaller features. Nojus ran a first experiment in Mixpanel’s experimentation module in under an hour. “It didn’t take more than an hour to push all the extra data we need for experiment analysis and then just choose the metrics we want to see,” he says. “Even product managers could select multiple metrics and do the analysis themselves without needing an analyst, and ​​previously it took a lot of analysts time to do that”.

For Darius, the deeper appeal is context. Right now, comparing experiment results across separate tools means jumping between windows, re-checking whether the metrics match, and doing manual reconciliation that introduces errors. Having experimentation and product analytics in the same place changes that entirely. “You can be sure it’s using the same metrics, and you can drill deeper through the non-AB testing lens very easily,” he says. “You don’t have to jump contexts. You can go from the dashboard you’re familiar with and apply a filter that turns it into an AB test. Right now it’s jumping between toolsets and lots of human errors happen. Making decisions between two different windows is way different.”