The Story of Illumex: Building Data Observability From the Trenches

The 3am pages never stopped coming. Fix a pipeline. Debug a query. Explain to a panicked business stakeholder why their dashboard showed zeros. Repeat.

For Ina Sella, working on data infrastructure at Meta meant living this cycle constantly. “I was building a lot of pipelines for business teams and data science teams,” she recalls. “And very often I would get paged in the middle of the night because something broke in production and I need to go and fix it.”

In a recent episode of Category Visionaries, Ina Sella, CEO & Founder of Illumex, shared how those sleepless nights at one of the world’s most sophisticated tech companies became the foundation for a business that’s now helping enterprise data teams prevent their own 3am disasters.

The Decision to Build

Leaving Meta to start a company wasn’t an impulsive decision. Ina had seen the same problems play out repeatedly across different teams and projects. The existing monitoring tools weren’t solving the real issue—they’d tell you something broke, but not why or how to fix it quickly.

But data observability was already a crowded space. Multiple well-funded competitors had staked out territory. The logical advice would be to find a different opportunity. Ina saw it differently. “There are already a lot of companies in the space,” she acknowledges. “But I felt like if I had this problem and I worked at one of the best companies in the world that built all of these technologies, I’m assuming that a lot of other data teams are also experiencing this.”

The insight was critical: if Meta, with all its resources and talent, struggled with data pipeline reliability, smaller companies with fewer resources must be struggling even more. The market wasn’t too crowded—it was underserved by solutions that actually solved the problem.

Building What Was Missing

From the start, Illumex took a different approach than existing players. Rather than focusing on monitoring and alerting, Ina focused on what happened after the alert fired. “We started working on building our data observability solution that is more focused on helping data teams to troubleshoot and resolve issues as fast as possible,” she explains.

This meant building capabilities others had overlooked. Full column-level lineage showing exactly how data flows through systems. Actionable recommendations for fixes instead of just error messages. Tools that helped data engineers actually resolve problems rather than just know about them.

The differentiation wasn’t about having more features—it was about going deeper on the features that mattered most to the people getting paged at 3am.

Learning to Sell

Like many technical founders, Ina had to learn sales from scratch. Her first call was a masterclass in what not to do. After sending the recording to an advisor, the feedback was blunt: “you asked like maybe one question during the entire call. You just talked the entire time. You need to ask questions, you need to learn about them.”

That moment became a pivot point. Instead of treating sales calls as pitching opportunities, Ina reframed them as learning sessions. “Even now when I go to sales calls, I always try to learn something from the customer that I didn’t know before the call,” she says.

This approach transformed both her sales effectiveness and Illumex’s product development. Every customer conversation revealed new insights about how data teams worked, what they struggled with, and what they needed from an observability platform. The founder who started knowing nothing about sales became the company’s most effective salesperson—not despite her technical background, but because of it.

The PLG Experiment

As a technical product serving technical users, Illumex seemed like a perfect candidate for product-led growth. The team built out self-serve capabilities and waited for signups to convert. They didn’t. “We thought about doing like a free trial or freemium, but it didn’t really work,” Ina admits.

The problem wasn’t interest—it was implementation. Getting value from a data observability platform required connecting data sources, configuring integrations, and understanding the insights. Prospects would sign up but wouldn’t complete the setup without guidance.

Rather than abandon self-serve entirely, Illumex adapted. “What we ended up doing is keeping the ability to self-serve and sign up, but we will contact you right away,” Ina explains. This hybrid model preserved low-friction entry while ensuring customers actually got to value.

The lesson reinforced something Ina was learning across every aspect of building the company: conventional wisdom about GTM strategy matters less than understanding what your specific customers actually need.

Competing Through Care

As Illumex grew, the competitive pressure intensified. Larger, better-funded competitors had bigger sales teams, more marketing budget, and established brands. The traditional playbook would suggest either finding a niche or trying to outspend competitors on brand and demand generation.

Ina chose a different path: outcare the competition. “We try to be extremely responsive,” she says. “We try to be very hands-on with our customers. I’m very involved myself. I join sales calls, I join customer calls.”

This wasn’t a temporary founder-mode strategy to be shed as the company scaled. It was a deliberate positioning: be the company where customers can actually talk to the CEO, where problems get solved in hours instead of days, where feature requests get real consideration instead of disappearing into a roadmap.

Larger competitors couldn’t match this even if they wanted to. Their organizational structure and processes prevented it. Illumex turned scale disadvantage into a competitive moat.

Moving Upmarket

The customer profile evolved significantly over time. “In the beginning we had a lot of kind of like initial customers that were more like mid-market companies,” Ina notes. “Today a lot of our customers are enterprise.”

This shift required more than just product changes. Enterprise customers brought longer sales cycles, more stakeholders, stricter procurement processes, and different proof point requirements. Illumex had to build out case studies, invest in security certifications, and develop more sophisticated ROI frameworks.

But the fundamental value proposition remained unchanged. Whether serving a 50-person startup or a Fortune 500 company, the problem was the same: data teams spending too much time firefighting and not enough time delivering value.

What’s Next

For Ina, the vision extends well beyond monitoring and alerting. “I want it to be the place where people go when they want to understand anything about their data,” she explains.

This means evolving from a troubleshooting tool into a comprehensive data intelligence platform. Deeper AI capabilities to predict issues before they occur. More sophisticated analytics to understand patterns across data systems. Broader coverage across the entire data stack.

The roadmap is ambitious, but it’s grounded in the same principle that started the company: solve real problems that data teams actually face, not the problems that look good in a pitch deck.

The journey from those 3am pages at Meta to building Illumex hasn’t followed a conventional path. There was no obvious white space in the market. No clear differentiation on day one. No magical growth hack that unlocked scale. Just a founder who deeply understood a painful problem, built something that actually solved it, and learned to sell it by staying obsessively close to customers.

That might not make for a tidy narrative, but it’s increasingly what building successful B2B companies actually looks like. The winners aren’t always the ones with the most funding or the slickest positioning—they’re the ones who understand their customers’ problems better than anyone else and refuse to stop until they’ve actually solved them.