He built a $12.7 Billion Company, Got Publicly Fired, and Then Accidentally Invented One of the Best Startup Ideas of the Decade.
This is the story of Andrew Mason, the man who went from Groupon's billion-dollar CEO to a publicly humiliated founder to the quiet architect of Descript, the AI-powered video and podcast editing platform that's now doing $55 million a year and changing how the world makes content.
A Kid Who Thought in Products, Not Paychecks
Dreams of working in Silicon Valley didn't come to Andrew Mason at birth. Instead, growing up in Mount Lebanon, a suburb of Pittsburgh, Pennsylvania, Andrew came from a family that was all about producing products, selling, and making money. His father sold diamonds, while his mother ran a photography firm.
At the age of 15, Andrew started Bagel Express - a delivery service that delivered bagels to people on Saturdays. Not any old-fashioned lemonade stand or mowing lawns, but a bagel delivery service - an efficient, periodic and local-level operation that Mason would run in decades ahead.
In 2003, Andrew graduated from the prestigious Northwestern University, where he majored in music. Music, not computers or entrepreneurship - yet another detail to keep in mind.
After graduating from college, Mason joined the ranks of a web design company owned by Eric Lefkofsky, who would later become an important figure in Groupon's history. After that, he was offered to pursue his education in public policy at the University of Chicago. But he never got that far. There were more important things to do.
The Point, the Pivot, and the Birth of Groupon
What Mason started building was known as The Point. It was a community site based on Malcolm Gladwell's book 'The Tipping Point' where people came together around an idea and took action when enough like-minded people were willing to join them—a fascinating concept, difficult to market, too conceptual for the online audience of 2007.
One day, however, someone from their team realised that people used The Point to organise bulk purchases. As long as a certain number of individuals agreed to buy one item, the seller would have to offer them a discount on it. This was their insight. This was their pivot.
It was reduced to just one behaviour and branded as Groupon, derived from "group" and "coupon". It was a very basic concept, perhaps even crude. But the concept was brilliant.
Groupon launched on the first business day in November 2008, at the height of the global financial crisis – a time when consumers sought discounts and businesses sought clients. Timing couldn't have been better. Expansion was quick. Groupon launched in many cities worldwide within a couple of months. Within two years, Groupon had internationalised itself. Its earnings are said to have jumped up to $800 million per year by the end of 2010.
Andrew Mason turned down a $6 billion bid from Google.
Groupon went public in November 2011 at a $12.7 billion market valuation; it is considered the largest IPO of the internet age, following Google's.
At this point, Andrew Mason had turned thirty.
The Fastest Rise. The Most Public Fall.
What followed was one of the most painful stories in the history of technology entrepreneurship, not because it was dramatic or scandalous, but because it was so devastatingly ordinary.
Groupon's growth, it turned out, was easier to copy than anyone had anticipated—daily deal competitors launched by the hundreds. Merchants began reporting that Groupon customers were discount-chasers who didn't return after the deal expired. The economics that looked so compelling in 2009 looked fragile by 2012. Revenue growth slowed. Profits proved elusive. The stock, which had IPO'd around $20, cratered.
Mason had never really been a traditional CEO. He was quirky, direct, and often funnier than was comfortable in a boardroom. He sent unusually honest all-staff memos. He reduced his own salary to $756.72 a year at his own suggestion. He was not a man who performed the role of CEO; he just tried to do it, visibly, imperfectly, in public.
On February 28, 2013, the day after Groupon missed analysts' expectations for the second consecutive quarter, Mason was fired. He had been CEO of Groupon for 4.5 years.
His farewell letter to employees, sent immediately after the board's decision, became an instant hit in startup circles. It began with a joke, "After four and a half intense and wonderful years as CEO of Groupon, I've decided that I'd like to spend more time with my family. Just kidding, I was fired today", and then turned into one of the most honest, self-aware exit statements a founder has ever written. He didn't blame the board. He didn't blame the market. He acknowledged, clearly and directly, that the company's struggles happened on his watch.
His severance, based on his contract and publicly available salary information, was calculated at approximately $378.36. For the CEO of a company that had once been worth $12.7 billion.
Mason later admitted that being so transparent about being fired was something he came to regret ,not because it wasn't true, but because every article written about him for the next decade would begin with the words "ousted Groupon CEO."
The Healing Phase: Fat Camps, Rock Albums, and Learning to Breathe Again
What Mason did next is the part of the story that very few people know.
He lived for a week at a health spa, which he good-naturedly called a "fat camp", in Miami. He was taking six months off completely to be with his wife, singer-songwriter Jenny Gillespie, and their new baby. He hadn't pitched investors. He did not announce any new companies. He just stopped. "
Then, in late 2013, he cut a seven-song motivational rock album titled Hardly Workin', a seriously tongue-in-cheek business guidance wrapped in catchy tunes, pitched to young people heading into the workforce. It was absurd. That was hilarious. It was a man who had been humiliated in public and chose to confront that humiliation by being himself, not hiding. He was also beginning to think seriously about what was next. He entertained the idea of building an "Uber-for-expertise" bazaar where people could talk to domain experts on demand. He spent time on the concept. Then he realised, with the clarity that only comes from having made it through something difficult, that he wasn't excited by it. He was not going to waste the next decade of his life doing something he didn't really want to do.
So he followed his passion.
Music. Storytelling. Cities. Sound. Detour: The Company Nobody Remembers That Made Everything Possible
Detour was established in early 2015 by Andrew Mason and other co-founders. This was a mobile application that provided audio walking tours in different cities across the globe via GPS navigation. This was going to be a very interesting platform where one would experience the city as a story, with music varying by location.
They did take pains in creating these tours. For example, they collaborated with Radiolab of NPR. Tours are also done in collaboration with filmmaker Ken Burns. Creating these tours costs Detour tens of thousands of dollars per unit. But as Andrew said himself, it was a long-term commitment that would show the user what 'good' looks like.
He funded the company himself, deliberately avoiding venture capital. After Groupon ,where explosive funding had contributed to explosive expectations had contributed to explosive failure ,he wanted to build slowly and carefully. He took a year to launch.
Detour launched with San Francisco tours and later expanded to New York, Chicago, and Los Angeles. At $5 per tour or $20 per year for a subscription, it was never going to be a billion-dollar company. Mason knew that. He also knew, as he has said publicly, that getting people to go on guided audio tours for fun would always be an uphill battle.
But something interesting was happening inside Detour's studio.
The Accidental Invention of Descript
Detour was producing hundreds of hours of spoken-word audio content. That meant hiring voice talent, booking studio time, conducting recording sessions—and then editing it all. Lots and lots of editing.
And editing audio, Mason discovered, was a uniquely miserable experience.
"We would be doing voice-over sessions in our studio and noticing that it took probably twice as long to edit the sessions as it took actually to record them, and you needed a professional engineer to do that," he said.
Traditionally, audio editing involved editing waveforms, which were these mountain-range images of sound and required quite a bit of technical knowledge to understand. To edit one word, you have to figure out precisely when it started and stopped and then cut it out, hoping not to leave any audible gaps. This was difficult for someone like Mason, who knew all about the craftsmanship of sound. But for his non-engineer employees at Detour, it was practically impossible.
It was for that reason that Mason tried to find an app that could edit sound. There was no such tool.
So they built one themselves.
They found a PhD student at UC Berkeley working on AI-powered transcription, the ability to convert speech to text with high accuracy automatically. They got him. And then they developed a piece of technology that had never existed before, something that would create a text file from your recorded audio, then connect each word in the text file to the relevant point in the audio file.
Implications were massive. Editing an audio file amounted to editing a text file. Delete a word from the text, and it disappears from the audio. Reordering of words affected the audio. No more worries about waveforms or engineer mentality. There was nothing like spending twice as long editing as recording. All that was left to do was read a text and make adjustments.
"This sort of thrilled our visitors to the point that we sort of realised maybe this is a better business," said Mason while demonstrating his creation to audio producers in the industry.
What Mason was trying to develop was audio tours. What he developed was audio editing software for the future.
Spinning Out: $5 Million and a New Beginning
In December 2017, Mason spun out Detour's editing technology and launched it as a standalone company called Descript. Andreessen Horowitz general partner Alex Rampell wrote the first cheque, a $5 million seed funding round.
Detour was acquired by Bose in 2018, which had been planning to use the technology for an augmented reality audio platform. Mason's attention turned entirely to Descript.
The initial product was simple by today's standards: a Mac app and web service that transcribed spoken audio into a text document, allowed you to edit the audio by editing the text, and let you export the result. Pricing was equally simple: $10 per month for the full audio editing experience.
It was a big moment for podcasters.
At this point, podcasting was blowing up. The success of Serial had made it an accepted part of popular culture. Joe Rogan was creating the world's biggest podcast, and millions more people were trying to create their own, but each one was hitting an immovable wall. That wall was editing; it was difficult, cumbersome,
and technical. Descript took down that wall.
Adopters became fans. Adoption rates increased. Mason, now building his third startup, was not building it the way he did with Groupon. This time around, he was taking things slow. He was building with intention. He was focused on building something that would be loved.
Building the Machine: Funding Rounds and a Product That Kept Getting Better
Over the next five years, Descript secured $101 million in four funding rounds. Its investor list is impressive – consisting of the biggest names of Silicon Valley, including Andreessen Horowitz, Redpoint Ventures, Spark Capital.
However, the biggest funding round came at the end of 2022, in November, when Descript closed a $50 million funding round in its Series C led by the OpenAI Startup Fund, i.e., the vehicle established by the developers of ChatGPT. For Mason, this validation from the world's leading AI research organisation meant something beyond the capital. It confirmed what he had believed since the beginning: that AI was not a feature for Descript. AI was the entire foundation.
The timing was no accident. In 2022, Mason had seen an early version of ChatGPT in private preview. The experience, he has said publicly, gave him complete confidence that Descript's foundational vision, a world where AI handles the tedious mechanical work of media production so humans can focus on the creative work, was not just possible but inevitable.
Around this time, Descript hired Laura Burkhauser as VP of Product. A former Director of Product Management at Twitter with an MBA from UC Berkeley's Haas School of Business, Burkhauser brought a rigour and product judgment that complemented Mason's creative instincts perfectly. Within weeks of her joining, Mason had already decided privately that she was his successor.
The product evolved dramatically during this period. What had started as an audio editing tool became a full-stack content creation platform:
- Overdub: A voice-cloning capability that enables users to fix errors in their recordings by typing words in the software and producing audio with their own voices
- Video editing: Video can be edited using the same text-based editing methodology, so people can edit their interviews, YouTube videos, and social clips by editing their transcripts.
- Studio Sound: AI-powered noise reduction that cleaned up audio recorded in imperfect environments, bedrooms, cars, and noisy offices
- Underlord: An AI video editing agent, launched mid-2025, that allowed users to create or edit a video simply by describing what they wanted in natural language, Cursor for video, Mason called it
- Screen recording, captions, social clips, AI avatars: A full suite of tools that positioned Descript not just as an editing tool but as a complete content production environment
The core user base, podcasters, YouTubers, marketers, corporate learning and development teams, and sales enablement teams grew steadily. Apple named Descript a finalist for App of the Year. Enterprise customers began signing on alongside individual creators.
The Failure He Learned to Name
But Descript did not lack tough periods either. Mason openly discussed the challenges of developing a single product that would serve multiple user groups – professional creators, amateur users, and business teams. The product had moments when it became too much of everything, thus failing at being the best solution for all its target users.
Yet another topic that has arisen during the creation process is balancing AI capabilities with users' trust. The emergence of a new feature called Overdub, which enables the creation of a clone of a person's voice from an audio sample, raised an ethical question. Could some criminals take advantage of the invention and produce false voices? Descript took care of preventing that.
His reaction in such times was always the same: face the truth quickly and without denying it. Just as the candid approach that made him famous with his Groupon farewell letter, the same product philosophy is applied in Descript.
He was also very straightforward when it came to explaining what the position of a CEO meant, as well as recognising when you were no longer suitable for that position.
"As I've experienced before, this is not a surprise. I have been transparent that at some point it wouldn't be my place anymore. While I can work myself into an excitement about the role… why should I when I have someone ready to devour it?"
Stepping Back: The Graceful Exit That Groupon Never Got
Andrew Mason left his post as CEO at Descript in August 2025. He became Executive Chairman, a move which he made with a trademark self-deprecating sense of humour, saying, "The title that entrepreneurs give themselves when they want to keep hanging around the office undermining their new CEO until they get bored and take back the reins."
Burkhauser ascended as CEO. Upon the transfer, Descript had achieved revenues of $55 million per year at a growth rate of 75%, while Mason observed that the rate was increasing. The product had grown from what he built at a startup for audio tours into one of the fastest-growing AI productivity products in the creator space.
Groupon was nothing like that. It was in this situation that Mason was forced to quit publicly because the firm failed to meet analysts' expectations. This time around, he chose to quit, chose his successor, wrote the press statement himself, and even made a funny video with affectionate greetings for his successor.
He understood the pain of letting go.
Current Status: What Descript Looks Like in 2026
At present, Descript is an established AI-powered platform for creating videos and podcasts that serves creative, marketing, learning & development, and enterprise customers across over 150 countries worldwide.
Important aspects relating to where the platform is right now:
The product has evolved far beyond its origins in text-based audio editing. Underlord, the AI editing agent, is now the centrepiece of Descript's product strategy, enabling users to go from raw footage to polished, captioned, and distributed video through natural-language instructions rather than manual editing. AI avatars allow users to present content without ever being on camera. Voice cloning, now used responsibly with speaker consent workflows, allows creators to correct mistakes without re-recording.
The revenue model includes subscription plans at three levels: Creator ($144 annually), Pro ($288 annually), and Enterprise ($600 per seat annually). In addition, the company uses AI credit models based on usage. In doing so, the company can generate revenue in two ways.
The team is now led by CEO Laura Burkhauser, whose track record of shipping beloved AI features and whose deep product instincts, Mason publicly credited, have driven the company's recent acceleration,gives Descript continuity without stagnation.
Andrew Mason remains Executive Chairman ,involved, available, but no longer responsible for the day-to-day. And now he's gotten into another venture, creating a board game club called Tabletop Library in Berkeley and testing how artificial intelligence can assist a physical location-based business. And of course he did.
What Andrew Mason's Story Teaches Every Founder
It would be unfair and unrealistic for us not to discuss Descript's history while ignoring Andrew Mason's, since one cannot exist without the other. The problem with all this is that there's always a message behind each discussion, which makes it very uncomfortable.
The lesson of Groupon is not that Mason failed. Groupon made billions of dollars. It genuinely changed how millions of people discovered local businesses. Its failure was the failure of a company that grew faster than it could manage, in a market that proved easier to copy than anyone anticipated. Mason wasn't a bad CEO; he was a first-time CEO of a rocket ship, and rocket ships are hard to fly.
The lesson of Detour is not that it was a detour. Detour is what formed Descript. There was no learning experience in enduring numerous hours of listening to audio tours before they were eventually edited. Nor was there any insight from this that could not have led to Descript had there been none, because without such an insight, Descript would not have existed. Hence, the alleged failure became the R&D phase of its successor.
The lesson of Descript is not just a product. This story shows how a founder learned, through experience, to use both integrity and humour to build much more wisely than before.
This is someone who learned to hire people who were much better at certain aspects of running a company than he was himself. This is someone who knew when to resign as the company's chief executive officer before it became necessary for others to do the job for him. This is someone who learned to talk openly about his problems without letting his team or his investors down.
Final Thought: The Music Degree That Made Everything Make Sense
Andrew Mason majored in Music at Northwestern University. This tidbit is generally regarded as amusing background trivia in whatever bio you read. Now, think about what Andrew was now allowed to build.
Descript is fundamentally a way to take the raw materials of your human voice, recordings, and stories and turn them into something cohesive, finished, and fun to listen to. In some respects, Descript embodies software audio production: Raw sound is shaped and refined through meticulous sculpting.
Despite his musical background, Mason did not build Descript on his own. He founded Descript because all the music education he ever got was really about listening to the world, being curious about the relationship between sound and story, and trusting that the tools you use to tell those stories can always be made better.
Following a bagel delivery in Pittsburgh for $12. And from a $7 billion IPO to a fired CEO's farewell letter to a rock album recorded in a Miami hotel room to an accidental audio tool in a San Francisco studio to $55 million in annual revenue and an AI agent that lets anyone edit video by talking to it, Andrew Mason's story shows that the best careers are rarely straight lines.
The question was never whether he would build something great again. The question was always just: what problem would make him angry enough to solve it?
Audio editing, as it turns out, made him very, very angry.
And the rest of us are better for it.