Companies Are Equations: Tom Conrad on Pandora, Quibi and the Math Behind Product Success

Conrad was 15 when the Macintosh launched and spent his teenage years wanting to work for Apple, at a time when that was not an obvious ambition. He studied computer engineering at the University of Michigan, which happened to be an Apple recruiting school, wrote a letter to land an internship, and then talked to the Finder manager Ron Lichty on his last day. Months later Lichty called to say he had a college hiring slot that would expire the next day, and Conrad got the job.

He joined in 1992 with ideas for making the Finder more usable. One was spring-loaded folders: hover over a folder while dragging an icon, drill down through the file system, and have the opened folders close when you drop. Late one night he needed an animation to show a folder was about to open, so he wrote code that blinked the folder five times, planning to ask a designer to replace it.

The feature shipped four years later and the folder still blinked five times. After the Finder was rewritten years afterward and spring-loaded folders were missing, users complained and it was re-implemented with the same blink, and iOS later brought a similar blink to folders. Conrad says the lesson is that designers and product people take shortcuts they plan to clean up later, and some of those details linger for 30 years.

Asked how to pick where to work, Conrad says young Tom would have said to find a product you are passionate about. Experience taught him that day-to-day satisfaction comes mostly from collaborators: whether you learn from them, like them, are challenged by them and are given latitude. His advice for years was to find something you love with people you think are amazing.

Looking back across hits and notable disasters, he says his professional satisfaction was not well correlated with external metrics such as big audiences or financial return. It correlated closely with whether he loved what the team was building and the people he worked with. Lenny notes that most people only get an interview or two before deciding.

Conrad thinks people are good at sensing a values mismatch. Every time he took a job with people whose values or working styles differed from his, he knew, but he told himself a story about why his suspicion was wrong. It is like meeting strangers at dinner, he says: sometimes you go home certain the person was the worst and sometimes you feel you made a lifelong connection, so pay attention to that instinct.

Conrad says the biggest lesson from both of his famous failures is that they made him better and opened doors. Even Pets.com's notoriety helped in later interviews, since few people had heard of the thousand other e-commerce sites of 1999 but everyone knew Pets.com. It also became a reference point later in his career, when the Pandora founders were impressed that he had run engineering there.

He describes three overfunded pet e-commerce sites that each raised more than $50 million and treated the market as zero-sum. As one spent on promotion and national television advertising, the others followed, creating an unwinnable arms race. He argues that excess investment can become an albatross that leads to unwise decisions.

Timing mattered too. Chewy, which PetSmart bought for $3 billion, executed well for a decade, while Pets.com launched when 80% of internet users were still on dial-up. Conrad rejects the critique that shipping dog food was inherently a stupid business, calling it simply very early. The company went from nothing to public to out of business in about 19 months, and rather than go bankrupt the leadership shut down and returned the remaining cash to investors, which he says no public company had done before.

After leaving Snap in December 2018, Conrad thought he might be done with software and listed unrelated careers such as pastry chef and landscape photographer. Meg Whitman and Jeffrey Katzenberg called about their venture, then called NewTV, to merge Silicon Valley mobile expertise with Hollywood production. He declined for seven months until Whitman said two hires had left within weeks and asked for a couple of days a week as a personal favor. He ended up immersed and fell in love with building again.

Conrad says he used to believe that a delightful product solving a real problem would earn word of mouth and the rest would follow. What he now appreciates is that a company is a math problem that pours investment in and returns out. Product details influence variables in the equation, but if the equation is fundamentally broken, iteration cannot fix it. Quibi bet that a bespoke library could convince people to subscribe and stay for a couple of billion dollars.

The company made 70 shows in 18 months, more than all the major broadcast networks combined in a year. About a third of the content spend went to daily shows made day-of in purpose-built studios, and Quibi launched two weeks into Covid, so hosts shot from their garages and the content looked like YouTube. Conrad says the real flaw was that the library needed to be much larger and the format needed time to iterate. The math moved from $2 billion to $6, $8 or $10 billion, a risk nobody could stomach.

Conrad says Quibi planned to use Hollywood's apparatus to market, not just make, its content. Katzenberg recruited a Rolodex of famous celebrities who would appear in culture to promote their shows. The theory was that Hollywood techniques that fill theaters with 20 or 30 million people on opening weekend could drive app downloads and free trials.

He calls that audacious, noting that Uber has enormous budgets for paid acquisition and is not consistently a top-10 app. Quibi's math required landing in the top 10 on day one and staying there. Conrad's self-critique is that he should have beaten the drum harder about how unlikely that distribution was.

It was not structurally impossible, only highly improbable, which made it hard to argue. Quibi did get millions of downloads and decent retention, but not record-setting retention. Conrad assumed they would iterate on the funnel as young companies do, but with about a billion dollars a year going to content, you cannot afford to not be a hit. Lenny adds that these are bets investors know they might lose, and Conrad accepts that risk and reward are linked but feels a responsibility to do better for investors.

Conrad worked at Apple in the gap between Steve Jobs's two stints. The company shipped hardware and moved from Motorola 68000 CPUs to PowerPC but struggled to ship consumer software, and his own folder features shipped only after he left. The culture rewarded people who contributed broadly across engineering, design and marketing, and he left after four years feeling he had been six miles wide and a centimeter deep.

At Berkeley Systems, which made You Don't Know Jack and the flying toaster screensaver, he became technical director with a clear job: ship software on time with high quality. Feedback on gameplay or packaging would not help his career. He says being forced to go deep is when he truly became a software engineer, and the franchise is still releasing new versions nearly 30 years later.

His takeaway is that it pays to get really good at something rather than remain a generalist. Culturally he says swim lanes, where people excel in their own lane, can be valuable. He calls himself lucky that he stumbled into the game when he did.

After Pets.com and a stint in enterprise software, Conrad wanted to build a destination for music discovery, drawing on a recommendation engine he had built in the enterprise world. A chance meeting at Coachella in 2004 with a high-school friend led to Tim Westergren of Savage Beast, which licensed a music recommendation engine to companies such as Best Buy and AOL. He declined the VP of engineering offer first, because he wanted a consumer product, then accepted when they called again.

Within 90 days the founding head of technology and product had resigned and new CEO Joe Kennedy proposed a 1-Click Personal Radio. Kennedy gave Conrad the keys as long as it fit that brief, and they spent 10 years building Pandora together. Conrad says everything he is as a product and engineering leader comes from that period.

At launch, support@pandora.com was an alias to all@pandora.com, so every employee saw each request and whoever saw it first replied. There was no support team in the first year, no macros, no FAQ links and no rules about what staff could say, and employees could agree that a feature was bad. Conrad says they could not out-Apple Apple on brand during the iPod era, so being themselves fueled word of mouth. Pandora spent no money on paid acquisition in his time there, and Lenny notes it grew from zero to 80 million users and about 3,000 employees.

Conrad says Pandora still reaches tens of millions of US listeners monthly, but he sees it as bittersweet because the team misplayed an opportunity. When it started, digital music was a category investors avoided because of lawsuits, label relationships and no big exits, since the largest outcome had been Yahoo buying Musicmatch for around $400 million. Investors who did believe put in hundreds of millions of dollars, but never matched the enthusiasm Spotify and Snap later enjoyed.

By Spotify's rise the comparables had grown from Musicmatch to Pandora at $8 billion and even Netflix at $100 billion. Being public meant market investors were still figuring Pandora out, which limited its access to capital and appetite for risk. Conrad adds that the team was inspired by disrupting terrestrial radio, a $30 billion advertising category against $8 billion for recorded music.

The team assumed rivals would chase the smaller owned-music market while Pandora was left alone in the bigger one. Conrad says they got that wrong, since all music was inevitably going to be delivered as a stream and the record labels would set terms in their preferred Spotify structure. Pandora operated under a government statutory license that the labels disliked, and moving to direct deals was risky but a requirement. The company eventually did, but too late.

Lenny observes that Pets.com and Pandora were both early, which suggests a Quibi-like idea might come back. Conrad notes a headline from that very day asking whether TikTok is the new TV, which was Katzenberg and Whitman's own framing. He says purpose-built mobile video clearly looks a lot like TikTok.

At Pandora, Conrad ran both product and engineering and had a seat at the table because Kennedy let his team do their jobs. After a year off he was introduced to Evan Spiegel, and a morning of coffee and interviews in Los Angeles turned into a written offer to be VP of product by one o'clock. He asked for another week to meet more of the team, then accepted knowing he would be Spiegel's right hand, executing Spiegel's vision at high level and in the details.

Conrad calls Spiegel brilliant and a kind person with very high standards, but the job was exhausting and less soul-satisfying than he wanted. The big lesson was about risk: with strong investor support, Spiegel could acquire a technology he believed was game changing and build speculative features that other roadmap processes would prune. Conrad says he now takes bigger risks and encourages bigger swings because of that.

On how much a product-oriented CEO should dictate, Conrad says it depends on company stage. Apple's talented people talked themselves out of every imperfect solution, and he jokes that a vote of a thousand to one was a tie. Jobs's return ended the voting. In a small organization, he argues, a lighter touch works better, and since nine times out of ten an employee will start working on whatever the CEO casually mentions, leaders must be careful how they engage with details.

After Quibi, Conrad wanted to shape a company's culture and values directly and to work on something with more weight than media. Independently, eight months into the pandemic, his health was at an all-time low: he was 50 or 60 pounds overweight with cholesterol in the red zone. He studied metabolic health deeply, used Zero for his own health, and when founder Mike Maser said he wanted to step up to executive chair and find a CEO, it ticked every box.

He describes Zero as a 27-person company with more than a million monthly users, over 100,000 of whom pay, with tens of millions of dollars in revenue and double-digit growth. He says 75% of users lose weight, typically 5 to 10% of body weight in the first 60 days, and growth is entirely organic with no paid acquisition. The focus has been unit economics: what value the company derives from each install and how efficiently it converts, engages and retains people.

Conrad says that at 53 the scales fell from his eyes: companies are equations. He began with the leaf nodes, such as conversion from install to registration to trial, but ended up building a spreadsheet model of the whole business from BI data and subscription transactions, which he uses to predict scenarios. It also explained why board members' questions had seemed unrelated to the product, because they were looking at the equation.

Lenny points out that consumer subscription apps rarely work, and calls Zero's scale rare. Conrad credits a decision to optimize the value of organic traffic and expand lifetime value instead of top-of-funnel spending, when competitors were raising huge sums to buy users in 2021 and 2022. He says that approach, unfashionable then, fits today's demand for capital efficiency.

Asked how he avoids burnout, Conrad cites the Japanese concept of Ikigai, the overlap of what you love, what you are good at, what the world needs and what you can be paid for. He says that for him all four have always overlapped, because he loves building, writing and using software, and even his hobbies are things like teaching himself After Effects. He notes that the intersections with only three of four can be a trap, because they feel close to something satisfying.

He says he is not a hustle-culture leader. At 23 he worked from 4 p.m. until 2 or 3 a.m. at Apple and got attention for it, while a colleague who arrived at 8:30, wrote a lot of great software and left each night contributed more but got less recognition. At Pandora he made a point of celebrating contribution over performance, though he accepts that some people, like his younger self, will want to work through the night.

After Pandora, he left with options and planned to live in London and Mexico City collecting experiences. After about six months he realized he cared more about relationships, and because people go to work on Monday morning, he wanted to collaborate again. That pulled him to Snap and later Quibi. Lenny compares it to Mark Manson's argument that freedom is useful mainly for discovering problems we want to keep solving.

Conrad's contrarian view is that the industry would be better with fewer founders. Many smart, creative, hardworking people could raise a $2 million seed and build a company that goes nowhere, when they would be better off joining a team that needs their skills and has a winning formula and a right-sized market. Whatever you want, acclaim, money or cultural impact, he says you can achieve it in collaboration with others without raising the seed round.

Lenny agrees and says he tells people not to start companies because it is painful and rarely works. Conrad adds that, though he is now a CEO, he has never been a founder and rejects it when founders at Pandora and Zero say he is like one: they took all the risk, and when he arrived there was a team, momentum and money in the bank.

In the lightning round he recommends Elad Gil's High Growth Handbook and the science fiction classic Hyperion. On interviewing, he advises that questions should start with "Tell me about a time in your career when...", so candidates describe something they have actually done rather than speculating about your business. Near the end he asks candidates to describe a really great day at work, to learn what naturally rewards them.

His life motto comes from Charles Eames: "The details are not the details. They make the design." He also shares that he once worked at CompuServe with Steve Wilhite, the creator of the GIF, who insisted on a pronunciation that Conrad firmly disagrees with.