CEO post-mortem: How Medium went from losing $2.6 million a month to profitability by cutting its losses?
TechFlow Selected TechFlow Selected
CEO post-mortem: How Medium went from losing $2.6 million a month to profitability by cutting its losses?
How to save a startup that even investors don't want?
Navigating Web3 tides with focused insightsOriginal Author: Founder Park
01 Introduction: Falling into the Hole, Then Climbing Out
In 2022, Medium was losing as much as $2.6 million per month, and paid subscribers were steadily leaving—meaning this massive spending couldn’t even be justified as investment in growth. Internally, we felt embarrassed by the content we promoted and touted as successful. Our users were less polite, calling the platform full of “get rich quick” garbage—and worse.
Then, the entire venture capital funding pipeline dried up. There was no more investor money to refill our rapidly depleting bank account (and frankly, we didn’t deserve it in our then-current state). No buyers were interested in taking over a complex, shrinking, high-cost business. But this actually simplified our decision: either make Medium profitable, or shut it down.
The challenges went far beyond that, but fortunately, there remained a group of people who genuinely wanted Medium to succeed. The arc of this story follows the classic Kurt Vonnegut narrative structure he called “Man in Hole”: we were doing well, fell into a deep pit, and then fought our way back out.
02 Past Glory: Minimalist Design and a New Business Model
The “doing well” phase owes much to former CEO Ev Williams. He founded the company after previously launching Blogger and Twitter. Now stepping back as chairman and largest shareholder, he still enthusiastically sends me countless messages as current CEO (that’s me).
Ev led two distinct eras here. The first was the “design era,” where the team redefined what a writing platform could look like, making every aspect of the user experience both simple and beautiful. In the second era, he pioneered a new business model—moving away from the toxic incentives of advertising toward a unique bundled subscription service that shared revenue with all creators.
The problem lay precisely in this business model. It turned out to be incredibly difficult to operate successfully while simultaneously fulfilling our grand vision of “building a better internet,” serving readers and writers well, operating as a healthy business, and resisting speculators, spam, and trolls.
03 Medium’s Content Quality Crisis
In July 2022, I took over as the second CEO, tasked with two missions: save content quality and fix the finances. As mentioned, the financial situation was dire, and the quality of content we were paying to promote was equally bad.
By the time I arrived, we had already cycled through multiple failed experiments in content quality—much like the “Goldilocks” story: we tried approaches that were too expensive, which failed; then ones that seemed cheap (but were actually costly), which also failed. We desperately needed to find the “just right” balance.
Note: The “Goldilocks” reference comes from the classic British fairy tale “Goldilocks and the Three Bears.” In this context, it means that in any system, there exists an optimal, “just right” balance point—avoiding the pitfalls of both extremes.
Fairly speaking, Medium has had moments of high-quality content. The first was from 2012 to 2017, before we even had subscriptions—a time when Medium was the purest, most thoughtful home for writers on the internet, attracting people with genuine ideas and insights to share. The second was from 2017 to 2021, when we hired a professional team of seasoned media executives and editors to commission thousands of professionally produced articles.
The end of that professional editorial era was driven by both financial and mission-related reasons. As an active Medium user and publisher at the time, I felt the mission conflict most acutely. Strategically, paying professionals to create high-quality content to attract subscribers seemed logical.
But as a community member, I felt these outside “pros” were competing with native community members for space. Professionals were taking the stage from amateur creators—yet it was those amateur creators who formed the foundation of the platform and were most likely to share unique, commercially valuable personal experiences. Medium works best when it gives a voice to those who don’t want to be professional content creators. We believe these voices (your voices) often hold the most valuable stories. The internet cannot belong only to professional media, influencers, speculators, and content creators. There must be a place that values user-generated content (UGC), that appreciates the sharing of expertise or academic insight, and understands the value of lessons drawn from interesting lives lived and written about.
When I joined as CEO, I increasingly felt the heavy cost of that “professional editor” era. While that team brought Medium over 760,000 paid members, it also burned through huge sums of money. One of my core tasks upon joining was to help fill the financial hole left behind.
After the professional editorial era came an 18-month period of low-quality content. As investor Bryce Roberts put it: “When your product is giving money away, you’ll always find a so-called ‘product-market fit.’”
We were giving money away, naively believing we were incentivizing loyal users. Predictably, this attracted a wave of new authors with questionable motives.
By mid-2022, readers complained that Medium was flooded with endless “get rich quick” scams. Founder Ev lamented the prevalence of clickbait and shallow summaries of others’ work. A common “viral” formula emerged: take a Wikipedia article, attach a sensational headline, rewrite it in flashy, emotional “broetry” style, and wait for the money to roll in. One such article could earn you $20,000.
I fully agreed with Ev: Medium’s foundation is being a mission-driven company. Our mission is “deepening understanding.” Much of the content we were paying for lacked real insight and violated our mission. This made us ask: What are we even showing up for?
To improve quality, we introduced Boost, adding human curation and expert judgment into the recommendation system. We revised the Partner Program’s incentives to reward thoughtful, substantive work. We added Featuring, allowing publications to directly promote content they endorse—the core idea being that readers should have the final say in whom they trust.
No one claims the process was easy, nor do we claim these systems are flawless. But the articles that now rise to prominence on Medium are worlds apart from the past. This allowed us last year to confidently declare we’re building a better internet—one that values deep thinking and authentic connection, not misinformation and division. No one accuses us of empty rhetoric—that’s one of many signs we’ve succeeded.
04 Chaotic Internal Governance, Lost External Confidence
At the bottom of the hole, two groups were tied to Medium’s fate: investors and employees. (Of course, readers, writers, and editors too!)
Investors had long lost faith and showed no interest in helping us climb out. That’s normal for them. They expect some investments to fail and walk away when they do. We were just another failed portfolio company.
But our team unrealistically wanted to climb out anyway. I think it was Medium’s core spirit that inspired every employee during those darkest days. And I haven’t even finished describing how dire things were.
Medium’s future depended entirely on years of hard work by this team (and future hires). Yet all decision-making power and economic upside were heavily tilted toward investors who had already disengaged.
We owed investors $37 million in overdue loans. Folks, that means we were technically bankrupt on paper.
Additionally, investors held $225 million in liquidation preferences. This is a common startup investment term—essentially meaning that in a company liquidation, investors get their full investment back before employees get anything. In a booming market with inflated valuations, this isn’t an issue.
But in tough times, with a company already bankrupt on paper, this clause effectively tells employees: every bit of your future hard work will go 100% to investors who’ve already vanished. This was devastating for morale.
In short, the debt and liquidation preference represented the brutal financial cost of our fall. It was even worse—so to give the full picture, I must lay it all bare.
Accepting such massive investment created an extremely messy governance structure. You might assume that as CEO, I’m in charge. But on major decisions, I needed investor approval. At Medium, this meant getting majority consent from five separate tranches of investors (who, again, had long stopped caring). Worse, following standard VC fund practices, these funds might eventually package and sell their stakes to secondary buyers—shifting our governance from indifferent but predictable investors to completely unknown, unpredictable ones.
Oh, and to make things more complicated, we owned and operated three other companies.
This was rock bottom. The best advice I received then was: “Don’t be a hero.” It came from one of our investors. He rightly pointed out a common founder flaw: thinking cleverness can solve any problem. But all these issues meant that no matter how brilliant our turnaround plan, we’d be blocked from hiring talent or making key decisions. Even heroic execution could be undermined by any single investor pulling the rug.
05 The Only Way Out: Profitability and Recapitalization
We ultimately didn’t run out of cash, get sold to private equity, or file for bankruptcy. On the contrary, since August 2024, Medium has been profitable.
We successfully renegotiated loans with creditors, eliminated liquidation preferences entirely, simplified our governance to a single investor tranche, sold two acquired companies, and shut down the third.
Looking back, it feels like we accomplished an incredibly difficult task. Because of content quality concerns, we couldn’t simply cut costs. Doing so would have made us profitable—but selling content we were ashamed of. That might be commercial success, but to us, it would be mission failure and a waste of life.
So we had to keep enough team to innovate on quality (as described), while aggressively cutting costs, finding growth paths, and renegotiating with investors.
The team performed brilliantly. I’d like to think I did okay too. Before joining Medium, I spent 15 years running small companies as CEO, and I’ve always taken pride in making my own ventures profitable. To me, that’s the essence of entrepreneurship.
But I did have two “superpowers.” First, running small companies gave me intimate knowledge of every operational detail—often because I had to do everything myself. Second, in the social media space, there may be no more hardcore Medium superuser than me. I’ve deeply used the platform in nearly every role: amateur writer, thought leader, business promoter, daily newsletter author, and creator of three of its largest publications. Nearly 2% of Medium’s pageviews come from my publications and articles.
06 Achieving Profitability: Growing Subscribers, Cutting Costs, Right-Sizing the Team
Recapitalization required a delicate timing: the company had to look good enough to be worth saving, but not so good that investors had other options.
So alongside fixing content quality, we first tackled the finances. This was basic financial risk management: we were burning cash, our bank balance was dropping monthly, and without intervention, we’d face insolvency.
We closed a gap—from a $2.6 million monthly loss in July 2022 to a $7,000 profit in August 2024. Since then, we’ve remained profitable. We’ve saved some profits as reserves, but mostly reinvested them into improving Medium itself.
Conceptually (not strict accounting), we split this financial turnaround into three parts: growing subscribers, cutting costs, and right-sizing the team.
-
Growing subscribers. This is undoubtedly our proudest achievement. When I started, our subscriber base was shrinking. When I mentioned terrible content quality, I also meant users hated it and canceled at alarming rates. Now, things are different. By restoring quality standards, we proved people will pay for thoughtful, insightful writing. For a better internet, this is a powerful vote of confidence.
-
Cutting costs. This too fills us with professional pride. Most savings came from cloud infrastructure, reduced from $1.5 million to $900,000 per month. This resulted from extensive engineering optimization and disciplined cost control. Inside Medium, we have a “step” slogan—each step up saves or creates $10,000 in value. We don’t care if it’s from growth or savings, so we celebrate wins on both fronts.
-
Right-sizing the team. This is important but sensitive to discuss. Medium once had 250 employees; today we have 77. I don’t know the full history of past layoffs, but I led one round. Two points: First, no leadership decision on layoffs was ever made lightly. Second, for a struggling company, layoffs are a harsh reality. A 77-person Medium is healthy; a 250-person one would be bankrupt.
In cutting costs, we also learned a painful lesson from office leases. Leases often last years longer than needed—it’s common. But sometimes, to secure office space, you must sign a lease longer than your cash runway. In normal times, subleasing markets exist—if you can’t afford rent, someone else might take over.
Yet we paid $145,000 monthly for a 120-desk office in San Francisco that we no longer needed. Like many companies during the pandemic, we went remote. Our staff was scattered across the U.S., so we didn’t use the office during or after the pandemic. We’re now committed to being fully remote—the office is obsolete.
Unfortunately, nearly all companies faced this, collapsing the sublease market. Our landlord was extremely rigid in renegotiations. We suspected they needed to falsely report occupancy to their own investors, counting our paid-but-empty floor as “occupied.” In that 7-story, ~800-desk building, fewer than 20 people were typically present on any day—none from our team. We tried everything to exit—offering to pay remaining rent upfront just to recover some fees. But the landlord refused for a long time, likely needing paying tenants to prop up their own loan negotiations. Only after their talks concluded did they allow us to pay an exit fee.
07 Recapitalization Was Hard—but Necessary
This section covers renegotiating with investors.
Honestly, despite lacking relevant experience, I found the process oddly enjoyable. Medium attracts curious people—I’m one of them. This mess was exactly the kind of scenario you hear about theoretically but rarely get to live through.
An ironic twist: the frozen VC market helped us. It left only two choices: shut down or become profitable. In a healthier market, our lenders might have forced a sale. But in that depressed climate, the Medium team held the real leverage: either give us motivation to keep working, or we all quit and you lose everything.
This type of negotiation is industry slang for a “recap”—a restructuring of the “cap table” (equity ownership). Initially, I resisted the idea, as it clashed with my view of business ethics: if you take someone’s money, you owe them returns.
So I had to shift my mindset. Every founder may need to realize this: sometimes, unless you clean up the cap table, the company has no future.
The day before I started, my investor friend Ross Fubini met me. He seemed genuinely excited about my turnaround plan—then mentioned recapitalization. It was the first time anyone suggested it. I firmly believed I’d never do something so “rude” to prior shareholders. But he was blunt: unless I renegotiated with investors, all my work would be wasted. About a year later, I admitted he was right.
Then came the question of how. Typically, recaps are led by an external “white knight” investor when a company faces existential threat. All existing shareholders must choose: accept the new terms or let the company die.
But we had no external investor option—both because the VC market was frozen and because, even in good times, we weren’t attractive due to lack of “VC-scale” growth.
So I learned about two alternative forms this “death threat” could take. The first came fittingly from a Medium article titled “Clean Up Your Own Shite First.”
The investor author, Mark Suster, noted that for relationship reasons, new investors avoid being the “bad guy” forcing a recap. They prefer management to act first—by threatening mass resignation.
Side note: this perfectly illustrates what I call “the commercial value of amateur writing.” An amateur’s insider insight created millions in value for all of us. If you’re a paid author on Medium today, your income owes everything to Mark’s article. Long live UGC!
Frankly, the “management resignation threat” strategy was far outside my comfort zone. It wasn’t just that I’d never seen a recap—I wasn’t the type to issue ultimatums over $200M+ in investor rights. But the logic was clear: without this recap, Medium would eventually fail, and my efforts would be pointless.
So I began arguing that without recap, there’d be no employee incentive—but then realized we had $37M in loans maturing across multiple investors. That was an additional, clearer “death threat.”
I proposed to loan holders: convert your debt to equity, or management walks, and we’ll structure a recap offering enough equity to new investors.
Essentially, a recap involves two things: investors give up special rights like liquidation preference and governance roles. They also usually accept heavy dilution—so if they once owned 10%, they might now own 1%. Hence, recaps are sometimes called “cram down rounds,” compressing existing investors into a smaller pool to make room for future teams and investors.
This recap was structured as a new funding round. But Medium had so many rounds—XX Round, Z Round—that our lawyers called it “A prime,” symbolizing a fresh start.
A fairness-preserving element allowed prior investors to participate in the new round. Though terms were aggressive, participation rights theoretically prevent reducing old investors to zero. In our case, only 6 of ~113 investors participated. I see this low uptake as proof we didn’t offer unfairly favorable terms.
Beyond negotiating terms, a recap requires extensive relationship work: with investors, ex-employees, and current staff.
Investors were surprisingly easy. This confirmed truths about top-tier startup investors (we had several big names): they’re reliable partners focused on “home runs,” not nickel-and-diming. They won’t squeeze extra cents from such deals. They care about reputation, so they avoid drama. I never thought I’d praise VCs, but post-experience, I can easily commend Ross at XYZ, Mark at Upfront, Greylock, Spark, a16z.
A recurring theme: old companies get “written off” by the market. This applied to many ex-Medium employees. I’m friends with many, having rented space in various Medium offices over the years. Their equity was also heavily diluted. I called some to explain their shares were likely worthless, and the recap might make them slightly more than zero—but ensures their work building Medium wasn’t in vain. For a few, I even offered a return to work if they wanted meaningful equity. If you work at a startup and your equity is “not worth the paper it’s printed on,” here’s a perfect case study on why that’s often true. These ex-employees had no power to block it—I called out of responsibility.
Then came current employees, some holding equity from the earliest days. They too were diluted. I wrestled with this, but recap logic prevailed. To justify the recap, you must show you’re clearing incentives for the future team—meaning past efforts are diluted, only future ones rewarded. So we issued new equity grants with new vesting schedules, but no replacements for old grants. This included me. Explaining their prior equity was likely worthless was straightforward: high exercise costs, buried under massive liquidation preferences, in a company unable to repay overdue loans. I use “likely” because we never truly know a company’s value until someone tries to buy it. But post-recap, current liquidation preferences are below annual revenue—so equity now has a real chance of value.
All this marks our emergence from the hole. We have clean finances, profits, a product we’re proud of, and a simple corporate structure. I increasingly appreciate our lawyer’s point: we’re making a brand-new start.
We now often remind ourselves why we do this work. I can’t give a rational reason beyond loving reading and writing, falling for this company years ago because it did too, and now wanting to see what we can build on solid ground. I’ve been CEO for three years, but in love with Medium for 13. So to me, regardless of business impracticality, saving this company feels worthwhile.
08 Appendix
Just noting a few things that didn’t fit naturally into the story—it’s already long.
-
To bridge the period between deciding on and executing the recap, we launched a “Change in Control (CIC)” plan to benefit employees. It’s like a contractual form of equity—a rare tool. I’m happy to share our template with any founder who might need it. It was ultimately replaced by the recap but filled a gap, ensuring employees benefited if the recap failed.
-
I completely ignored team psychology. This was crucial to the turnaround, starting with a team battered by repeated failures, unsure why the new guy (me) wouldn’t just be the next failure. I leaned on “Find Your Allies” from The First 90 Days, and the broader need to build confidence in the plan. I think I read the latter part in an HBR article, but can’t recall which.
-
There was another $12 million loan that wasn’t overdue. It was converted to equity as part of the recap.
-
Startup valuations sometimes carry vanity. Our peak was $600M. I have no vanity about our current valuation. But I won’t disclose it, as I don’t want it used to compare with other startups. We’re profitable; they’re not. That comparison favors us!
Join TechFlow official community to stay tuned
Telegram:https://t.me/TechFlowDaily
X (Twitter):https://x.com/TechFlowPost
X (Twitter) EN:https://x.com/BlockFlow_News
