/
.webp)
Generated by Master Biographer | Source for LinkedIn Content
San Francisco. Sometime in early 2012.
Three MIT graduates are sitting in a cramped apartment, staring at a dashboard they built for themselves. The dashboard is not their product. Their product is a voice recognition app called Sonalight — a hands-free texting tool for Android that lets you dictate messages while driving. They have 500,000 downloads. They have a spot in Y Combinator's Winter 2012 batch. They have seed funding. They have, by every reasonable measure, a functioning startup.
And none of them are looking at Sonalight.
They are looking at the analytics system they hacked together to understand what their users were doing inside the app. Where people got stuck. Which features people actually used versus which ones they assumed people used. The gap between what they thought was happening and what was actually happening — that gap was enormous, and the only reason they could see it at all was this internal tool they'd built almost as an afterthought.
Spenser Skates would later describe the moment with the brutal clarity of someone who has had years to process a failure: the analytics were more interesting than the product.
Not slightly more interesting. Fundamentally more interesting. Sonalight was a voice recognition app competing in a market that Apple's Siri was about to swallow whole. The app worked. It was, in Skates' words, "pretty good." But "pretty good" is a death sentence in consumer mobile. Pretty good means you'll never build a moat. Pretty good means a trillion-dollar company will ship your exact feature set as a system-level capability and your users will never open your app again.
The analytics tool, though — the thing they'd built to diagnose their own product's problems — that solved a pain they felt every single day. And when they started talking to other founders, they discovered something that changed everything: every other startup was feeling the same pain, and nobody had a good answer.
Google Analytics could tell you how many people visited a page. It could not tell you what those people did after they got there. It could not tell you that users who completed three actions in their first session retained at 40% while users who completed two retained at 12%. It could not connect behavior to outcomes. It operated in a world of pageviews and sessions — a model designed for the web of 2005, not the product-led world of 2012 where every click, every swipe, every hesitation was data that meant something.
The existing tools measured traffic. What Skates and his cofounders had built, almost by accident, measured behavior.
They shut down Sonalight. A working product with half a million downloads. They told their investors — the same investors who had backed the voice recognition play — that they wanted to pivot. The stigma of a pivot in 2012 was real. It meant admitting the original thesis was wrong. It meant burning months of work. It meant going back to the beginning.
Their investors said yes.
Not because the analytics tool was polished. It wasn't. But because the founders had done something that most pivoting companies never do: they had discovered the new idea by being their own first customer. They hadn't read a market report and decided analytics sounded lucrative. They had felt the absence of the tool in their own hands, built it out of necessity, and recognized that the necessity was universal.
On June 26, 2014, Amplitude officially launched. The voice recognition app was dead. The internal tool was the company.
Spenser Skates grew up in the kind of household where intellectual ambition was atmospheric. He enrolled at MIT to study computer science, where he met Curtis Liu during their freshman year. The two would become inseparable collaborators — the kind of cofounding partnership that's forged not in a conference room but over hundreds of late nights debugging code together, building things that didn't need to exist, just to see if they could.
Before Sonalight, Skates had worked as an algorithmic trader and researcher at DRW Trading Group in Chicago. The job was quantitative, precise, data-obsessed — a world where the difference between a good decision and a catastrophic one lived in the data, and your survival depended on reading it correctly. He was not a product person by training. He was a data person by instinct. That instinct would define everything Amplitude became.
Curtis Liu, the quieter half of the founding pair, was the engineering backbone. Also MIT, also computer science, also the kind of programmer who thought in systems rather than features. If Skates was the one who could articulate why behavioral analytics mattered, Liu was the one who could architect the infrastructure to capture and query billions of events without the whole thing collapsing under its own weight.
The third cofounder, Jeffrey Wang, came aboard as the company crystallized around the analytics vision. Wang's background was different — he had held engineering roles at Palantir Technologies, the data analytics firm cofounded by Peter Thiel that had built its reputation on making sense of massive, chaotic datasets for government and intelligence agencies. At Palantir, Wang had seen firsthand what happened when you gave smart people the ability to query complex data without writing SQL. The interface mattered as much as the infrastructure. The question you could ask mattered as much as the answer you could compute.
Wang brought that sensibility to Amplitude. The product would not be a tool for data engineers. It would be a tool for product managers, for growth teams, for designers — for the people who needed to understand user behavior but who would never write a database query. The insight from Palantir was not technical. It was philosophical: analytics is only as valuable as the number of people in your organization who can actually use it.
The three of them — the trader-turned-CEO, the systems architect, the Palantir veteran — formed a founding team whose collective obsession was not analytics software in the abstract but a very specific belief: that most companies were making product decisions based on intuition when they could be making them based on evidence, and the reason they weren't was that the evidence was locked inside tools that only engineers could operate.
They applied to Y Combinator first in 2011 with a different idea entirely — a new take on Amazon's Mechanical Turk. YC rejected them. They came back for the Winter 2012 batch with Sonalight and got in. The $120,000 seed check from YC was modest, but the network was invaluable. Sam Altman, then president of YC, would later point to Amplitude as one of the batch's standout outcomes — though at the time, nobody knew what the company would become, because the company didn't know yet either.
The early years of Amplitude were a masterclass in content-led growth before the phrase "content-led growth" existed.
The product launched into a market that already had an incumbent: Mixpanel. Founded in 2009 by Suhail Doshi, a 20-year-old college dropout who had been mentored by PayPal mafia member Max Levchin, Mixpanel had a three-year head start, serious venture capital backing, and a product that developers already knew. If you were a startup in 2014 and you wanted event-based analytics, you used Mixpanel. That was the default.
Amplitude's challenge was not just to build a better product. It was to convince a market that already had a solution that the solution was insufficient. And the way they chose to do this was not through enterprise sales or aggressive advertising. It was through ideas.
The North Star Metric framework became Amplitude's Trojan horse.
The concept was simple, almost deceptively so: every product should have one metric that captures the core value customers get from it. Not revenue. Not DAU. Not signups. The metric that, if it goes up, means you are genuinely delivering more value to more people. For Facebook, it was daily active users. For Airbnb, it was nights booked. For Spotify, it was time spent listening. The North Star Metric was the number that aligned every team in your company around what actually mattered.
Amplitude didn't invent the idea from nothing — the concept of a single guiding metric had floated around product circles for years. But Amplitude did something nobody else had: they productized the framework. They published a free playbook. They ran workshops — hundreds of them, eventually — helping product teams at companies of all sizes define their North Star. They created templates, blog posts, conference talks. They built an entire intellectual ecosystem around the idea, and at the center of every conversation was Amplitude's product, the tool you needed to actually measure and track your North Star Metric.
It was content marketing as category creation. You didn't just buy Amplitude's software. You bought into Amplitude's worldview. And if you believed that product decisions should be driven by behavioral data rather than gut instinct, then Amplitude was the obvious choice — because Amplitude had taught you to think that way in the first place.
The strategy worked. By 2016, Amplitude had raised a $15 million Series B led by Battery Ventures. By 2018, Sequoia Capital led an $80 million Series D. The customer list grew from scrappy startups to enterprise logos: Walmart, NBC, Burger King, PayPal, Atlassian. Revenue climbed from almost nothing in 2015 to $68 million in 2019 to $102 million in 2020 — a 50% year-over-year jump that put the company squarely in the crosshairs of Wall Street.
The Mixpanel rivalry intensified through this period, but it evolved in unexpected ways. Mixpanel's founder Suhail Doshi stepped down as CEO in 2018, handing the reins to Amir Movafaghi, a former Twitter executive. Under Movafaghi, Mixpanel made a deliberate choice: stay private, stay lean, stay profitable. The company had raised $77 million in total venture funding, and $60 million of it was still in the bank when Movafaghi took over. Mixpanel didn't need to go public. Mixpanel didn't need to prove anything to Wall Street. Mixpanel could play the long game.
Amplitude chose a different path.
And then there was John Cutler.
Cutler joined Amplitude as a Product Evangelist, and his impact on the company's brand was disproportionate to his title. He was already one of the most widely read voices in product management — a prolific writer who had published close to 800 posts on product development, organizational dysfunction, and the patterns that separated high-performing product teams from everyone else. His newsletter, The Beautiful Mess, had a devoted following. He had worked at Zendesk, Pendo, AppFolio. He understood the product management world not as a vendor but as a practitioner.
At Amplitude, Cutler became the intellectual conscience of the brand. His writing didn't sell the product. It sold the philosophy — that product teams were drowning in feature factories, that most organizations were optimizing for output instead of outcomes, that the gap between what companies said they cared about (customers) and what they actually measured (shipping velocity) was vast and destructive. His thought leadership gave Amplitude credibility in rooms where marketing decks would have been thrown out. Product managers trusted Cutler because Cutler wasn't trying to sell them anything. He was trying to make them better at their jobs. That Amplitude happened to be the tool that enabled the approach he advocated — well, that was just convenient.
The combination of the North Star framework, Cutler's thought leadership, and a product that genuinely worked created a flywheel that carried Amplitude from scrappy challenger to category leader. By the time the company was ready to go public, it had 1,600 customers, $150 million in annual recurring revenue, and the confidence to do something unusual.
On September 28, 2021, Amplitude began trading on the Nasdaq under the ticker AMPL. But it didn't do an IPO.
It did a direct listing.
The difference matters. In a traditional IPO, a company sells new shares to raise capital. Investment banks underwrite the offering, set the price, and take their cut. The company walks away with fresh cash. In a direct listing, no new shares are created. No new capital is raised. Existing shareholders — founders, employees, early investors — simply make their shares available for public trading. The company gets liquidity for its people without diluting its ownership or paying Wall Street's toll.
Amplitude was the seventh company to choose a direct listing. Spotify, Slack, Coinbase, Roblox had gone before. It was still an unusual move, especially for a company Amplitude's size. The signal was clear: Amplitude didn't need the money. The Series F, a $150 million round led by Sequoia just months earlier, had valued the company at $4.15 billion. The balance sheet was healthy. The direct listing was a statement of independence.
Shares opened at $50. By market close, AMPL was trading at $54.80, giving the company a fully diluted market cap of approximately $7.1 billion. Spenser Skates, the MIT kid who had shut down a "pretty good" voice recognition app nine years earlier, was worth roughly $450 million on paper.
For six weeks, everything went up. The stock touched $87.98 on November 4, 2021 — the all-time high, achieved in the fever dream of late-2021 SaaS valuations, when every cloud company was priced as if growth would compound forever and interest rates would stay at zero until the heat death of the universe.
The fever broke.
What happened to Amplitude after November 2021 was not unique to Amplitude. It happened to every SaaS company that went public at peak multiples. But the severity of Amplitude's decline made it a case study in what happens when market gravity reasserts itself.
The Federal Reserve raised interest rates. The discount rate on future cash flows went up. Every high-growth, high-burn SaaS stock got repriced overnight. But Amplitude had a specific problem layered on top of the macro: its growth was slowing.
In its fourth-quarter 2021 earnings report, released in February 2022, the company trimmed its full-year 2022 revenue guidance. Where it had previously projected at least 40% growth, it now called for 35% to 40%, expecting $226 million to $234 million. A five-percentage-point haircut. In normal times, this would have been a footnote. In the environment of early 2022, it was a death sentence. AMPL dropped 45% in February alone.
The descent was relentless. From the November high of nearly $88, the stock fell through $50, through $30, through $20, through $10. By February 2026, AMPL touched $6.00 — its all-time low. A decline of 93% from peak to trough.
The human cost arrived in April 2023, when Skates announced a 13% workforce reduction — 99 employees laid off globally. His statement acknowledged what everyone already knew: "We are not immune to the headwinds being faced by our customers or to the macroeconomic environment at large." The layoffs were not an anomaly. They were part of a wave that swept through tech in 2022 and 2023, claiming 93,000 workers in 2022 and over 200,000 in 2023. But knowing you're in good company doesn't make the conference call easier.
The financial reality was sobering. Revenue growth decelerated from 50% in 2020 to 16% in fiscal 2023 ($276 million) to just 8% in fiscal 2024 ($299 million). The company that had once been priced for hypergrowth was now growing at a rate barely above inflation. ARR — annual recurring revenue, the SaaS metric that matters most — crawled from $281 million in 2023 to $312 million in 2024. Not shrinking. Not dying. But not growing at a pace that justified anything close to its public market debut.
Meanwhile, Mixpanel — the rival that had stayed private, stayed lean, stayed quiet — was doing just fine. No stock price to defend. No quarterly earnings calls. No activist investors. No public scrutiny of its growth rate. Mixpanel had chosen the unsexy path: build a profitable business, don't over-raise, don't chase the public markets. By 2024, both companies were building remarkably similar products — session replay, heatmaps, experiments, feature flags. The technical differentiation had narrowed to near-zero. The strategic differentiation was enormous: one company answered to Wall Street, and the other answered to itself.
There were signs of recovery. Fiscal 2025 brought acceleration: ARR reached $366 million, up 17% year-over-year. Q4 2025 revenue hit $91.4 million with $11.2 million in free cash flow — the first time in a while the company was generating real cash. The bleeding had stopped. The patient was stabilizing.
But the scar remained. Amplitude is, as of early 2026, one of the worst-performing public SaaS stocks since the 2021 peak — down roughly 84% from its all-time high. The company that chose a direct listing because it didn't need Wall Street's money has spent four years learning what happens when Wall Street decides it doesn't need you, either.
The Mixpanel comparison haunts every earnings call. Not because anyone says Mixpanel's name. But because Mixpanel's existence is proof that there was another path — a path where you don't go public at the top, don't promise 40% growth to people holding your stock, don't spend four years trying to earn back a valuation you should never have had.
Skates is still CEO. The product is still used by thousands of companies. The North Star framework is still taught in product management courses. John Cutler has moved on — he went to Toast as Senior Director of Product Enablement — but the intellectual infrastructure he helped build at Amplitude endures.
The question that hangs over the company now is not whether Amplitude will survive. It will. The question is whether the public market pressure made it a better company or just a more anxious one. Whether the layoffs and the guidance cuts and the stock collapse produced discipline, or just exhaustion. Whether the accelerating growth in 2025 is the beginning of a genuine second act, or just the kind of base-rate recovery that happens when expectations get reset low enough.
Amplitude's story is the story of every SaaS company that mistook a market cycle for a mandate. The product was real. The technology was real. The insight — that behavioral analytics would become essential to every digital product team — was prophetic. What was not real was the idea that a company growing at 50% deserved a $7 billion valuation, or that the growth would last forever, or that going public at the peak of a bubble was a reward rather than a trap.
The voice recognition app is long forgotten. The internal analytics tool became a public company. The public company became a cautionary tale. The cautionary tale is now, maybe, becoming a comeback story.
We'll see.
1. The YC rejection that preceded the acceptance.
Skates and Liu first applied to Y Combinator in 2011 with an idea for a new version of Amazon's Mechanical Turk. YC said no. They came back months later with Sonalight, the voice recognition app, and got into the Winter 2012 batch. Neither the rejected idea nor the accepted one became the company. The third idea — the one they didn't apply with — did.
2. Jeffrey Wang's Palantir DNA runs through the product.
Wang, the third cofounder, spent formative years at Palantir Technologies — the same company where Heap Analytics founder Ravi Parikh also worked. The Palantir philosophy of making complex data queryable by non-engineers became a foundational design principle at Amplitude. The product was always built for the product manager, not the data engineer. That choice came directly from watching what Palantir did for intelligence analysts.
3. The North Star Metric was a content strategy disguised as a framework.
Amplitude didn't just popularize the North Star concept — they turned it into a category-creation engine. The free playbook, the workshops (hundreds conducted), the blog ecosystem — all of it was designed to embed Amplitude's worldview into how product teams thought about measurement. By the time a team decided they needed a tool to track their North Star, Amplitude was the only obvious answer. It was the most sophisticated content marketing play in SaaS, and most of the people it worked on never realized it was marketing at all.
4. Sonalight had 500,000 downloads when they killed it.
Most pivot stories involve a failing product. Sonalight wasn't failing. It had half a million downloads and a functional product. The founders didn't pivot because the app didn't work — they pivoted because they recognized that "working" and "winning" were different things. The voice recognition market was about to be swallowed by platform-level capabilities from Apple and Google. Skates made the decision to abandon a viable product for an unproven one. That required the specific kind of courage that comes from caring more about the destination than the sunk cost.
5. The direct listing was a philosophical statement that aged badly.
Choosing a direct listing over an IPO in September 2021 was supposed to signal that Amplitude was different — mature enough not to need the capital, confident enough not to need the underwriter's safety net. The irony is that by not raising capital at the peak, the company went into the downturn with one fewer financial cushion. Every other SaaS company that IPO'd in 2021 pocketed hundreds of millions in fresh cash before the crash. Amplitude chose to be principled. The market did not reward the principle.
