At fourteen, Todd Olson arrived at a Delaware bank expecting the sort of summer work adults hand to teenagers when expectations are modest and the filing cabinets are full. He had been hired for data entry. Then someone noticed he could program. The assignment changed. Olson found himself working with Unix and databases, a young coder in the machinery of a large commercial institution because his mother had praised his computer skills to exactly the right person.
It is a neat origin story, but its useful detail is the reassignment. Olson keeps returning to the place where a system meets the person actually using it. A job description says data entry; observed ability says programmer. A software roadmap says a feature shipped; usage data may say nobody cared. He has spent a career distrusting the tidy label when behavior offers better evidence.
By his senior year at Carnegie Mellon, he had reduced his course load to build Cerebellum Software with a friend. The company made data-integration software and raised more than $20 million. It also supplied the first of two painful startup educations. A proposed acquisition fell apart after the board changed who was negotiating. The business did not become the ending its young founder imagined.
Failure leaves excellent notes
Olson moved through TogetherSoft, where he ran product development, and Borland, where he served as chief scientist for the Together business unit. Then came 6th Sense Analytics, a developer analytics company he founded in 2004. Rally Software acquired it in 2009. Olson eventually became Rally's vice president of products and helped the agile software company toward its 2013 public offering.
The résumé improved. The unresolved question did not. Olson had built two startups without finding the durable fit he wanted. Reading Steve Blank's The Four Steps to the Epiphany felt personal. The phrase product-market fit became less like venture vocabulary and more like a debt to be paid before growth could be trusted.
“You can't skip it. You can't hack it.”Todd Olson on product-market fit
At Rally, the next company announced itself as an irritation. Olson was responsible for products but could not easily answer elementary questions about them. Which features did customers use? Where did they struggle? Did they like the experience? The available choices were awkward: ask engineers to instrument individual events, stitch together analytics, or simply listen to whoever spoke loudest.
The metric before the money
In late 2013, Olson joined Erik Troan, Eric Boduch and Rahul Jain to start Pendo in Raleigh. The product combined software-use analytics with guides and feedback. Its technical wager was automatic tracking. Customers would install a snippet once instead of asking developers to tag every interaction by hand.
Raleigh was part of the choice, not a geographic apology. Olson has written about the extra work required to raise large venture rounds outside the Bay Area, while arguing that a company can turn its home market into an advantage. Pendo drew on product and engineering experience from Rally, Red Hat, Cisco and Google. It also built Pendomonium, a product conference that brought the market to its own front door. Community was not decoration around the sales motion. It helped teach a category that prospective buyers did not yet know how to search for.
That category required explanation. “Product analytics plus in-app guidance” was not a familiar shopping list when Pendo began. Search demand could not deliver buyers who lacked the words. Olson used introductions from venture investors and direct LinkedIn messages to reach product leaders. The work was manual by necessity: first explain the problem, then show the instrument, then ask for the install. Inbound marketing became useful only after the market learned the language.
For the first year, Olson cared more about installs than revenue. Installation required a customer to let young software into its own product, which made it a sterner signal than an encouraging meeting. Pendo reached roughly fifty installs before taking payment. Then Olson spotted one company using the product constantly, including screenshots of its charts in board material. He called and delivered an admirably plain sales pitch: if the charts were going to the board, perhaps the company should pay. The customer agreed to $500 a month.
Early signal strength · Olson's practical hierarchy
He remained the lead seller until Pendo reached about $500,000 in annual recurring revenue. That kept the distance between objection and code comically short. When a prospect wanted surveys and polls, the team built them and won what was then its largest account. Later, Pendo lifted its minimum monthly price from $99 to $1,000 in one move. The team feared demand would vanish. Deal volume held.
There is a discipline hiding inside this scrappiness. Olson declined to copy a common competitor feature, manual track events, for five years because he thought it created brittle work for customers. The refusal made automatic tracking distinctive. Saying no can be a product feature, provided the no is attached to a clear belief about the user's burden.
Founder mode, with phone calls
As Pendo grew, Olson tried to preserve contact with the raw evidence. He personally interviewed the first 500 employees, at one point spending 40 to 50 percent of his time hiring. He has said that he calls customers who give weak feedback scores. The practice turns an abstract number into an occasionally uncomfortable conversation, which is precisely why it is useful.
He also cooks for investors and customers at home. The domestic setting is more than hospitality. Enterprise software can acquire a ceremonial fog as organizations grow: prepared decks, rehearsed demos, polite nodding. Dinner makes the room smaller. A dissatisfied customer can still dodge a dashboard; it is harder to dodge the host passing a plate.
The company eventually acquired the scale that makes proximity difficult. Pendo crossed a $1 billion valuation after a 2019 financing and was valued at $2.6 billion in its 2021 Series F. The numbers placed a Raleigh software company on the same venture map Olson had once found harder to navigate. They also increased the number of layers between a CEO and an individual click. His habits make more sense against that pressure: interview broadly, call the low scorer, walk acquisition targets without bankers blocking the founder conversation, and keep finding reasons to touch the product.
Olson's public language is similarly practical. “Measure everything” is his starting instruction for product-led organizations. Teams, he argues, need shared data, shared definitions and a shared view of success. He favors compact, cross-functional “two-pizza teams” because consensus becomes expensive as groups swell. His 2020 book, The Product-Led Organization, expands the idea beyond a freemium funnel: the product should sit at the center of the customer experience, touching work from onboarding to support and retention.
The distinction matters. Product-led, in Olson's telling, does not mean the product department rules the building. It means every department looks for ways the product itself can carry part of the customer relationship. Guidance can replace a support ticket. Usage can sharpen a renewal conversation. Feedback can reach a team without waiting for the loudest account review. Software becomes both the service and the place where the company learns whether the service worked.
The coder returns
In 2025, Pendo acquired predictive analytics company Forwrd.ai, following its acquisition of feedback-analysis startup Zelta AI the prior year. A $100 million credit facility announced that September was earmarked for product innovation and expansion, with Olson describing a market that needed to understand how people use both conventional software and AI agents. In January 2026, he led a customer forum reviewing Pendo's agent analytics, prediction tools, model-context integration and agent mode.
The corporate language is large. Olson's personal update was smaller and more revealing: for the first time in a long while, he was building again. AI tools had made it easier for him to return to a blended product-and-engineering role. He described the work as energizing. After years of executive scale, the programmer who once escaped a data-entry assignment had found another route back to the terminal.
That return matters because product-market fit is not a trophy mounted in the lobby. Olson says it can be lost. AI changes how software is made, who makes it and what users expect from it. A company built to observe behavior now has to observe people working with systems that act. The old discipline still applies: watch what users do, decide what the evidence means, and resist mistaking a fashionable story for a durable habit.
“Ultimately, I'm a builder. That's what I like to do.”Todd Olson
Two early disappointments gave Olson his favorite sort of asset: a constraint. Pendo would not scale before signs of fit. It would measure installation, not enthusiasm. It would keep selling close to the builders until the objections repeated. Those choices sound obvious after they work. Before they work, they feel slow, awkward and faintly impolite.
The lesson available for borrowing is not to worship data or copy Pendo's pricing. It is to choose the customer behavior that costs enough effort to count. A compliment costs little. An install costs more. Repeated use costs time. Payment costs money. A board screenshot risks reputation. Each step says something stronger. Olson's career has been an extended attempt to listen at the strongest available volume.