TL;DR

My book summary app, Sumizeit, is profitable today — but the road there was a graveyard of expensive mistakes. I launched buggy code as a young engineer, burned through $100K I didn't manage well, poured money into paid ads that never paid back, and guessed wrong about what customers wanted. What turned it around: brutal honesty (eventually), feedback forms, Amplitude analytics, a 15-minute format, an Android app I built on a whim, onboarding flows borrowed from competitors, AI features when ChatGPT threatened to eat us alive, and — the real deciding factor — corporate partnerships I chased alone after my co-founder left.

The Bugs I Couldn't See

When we launched, the site was held together with hope and duct tape.

I was a young engineer, and I built it like one — confident, fast, and wrong in ways I couldn't yet recognize. Pages broke under load. Buttons did nothing. A user would sign up, hit a wall, and quietly disappear, and I'd never know why because my error logging was as immature as my code. I thought shipping fast was the same thing as shipping well. It is not.

The worst part wasn't the bugs. The worst part was the silence around them.

My business partner saw the problems. Of course he did. But he never told me. He was worried about hurting my feelings — worried that constructive criticism would land as an insult, that pointing at the cracks would crack our friendship. So he said nothing, and I learned the hard way that a co-founder's kindest silence can be the most expensive thing in the company.

Months later, when I finally understood how much had been quietly wrong, I asked him about it.

"Why didn't you just tell me?"

He looked genuinely pained. "I didn't want to discourage you. You were working so hard."

I understood the impulse. I even appreciated it. But I told him the thing I wish someone had told me earlier: "Discouraging me would have cost a week. Not telling me cost us a year."

That was the first real lesson of this company, and it had nothing to do with books. The most generous thing you can do for a founder is tell them the truth before it gets expensive.

Going Back to School (Mostly on YouTube)

Once I admitted how much I didn't know, I did the only thing that made sense: I went and learned it.

I took classes — real ones — on marketing, on product, on the parts of engineering I'd skipped on my way to looking competent. I read books the way a starving person eats. And honestly, some of the most useful education didn't cost a thing. YouTube taught me more about funnels, retention, and growth loops than any single course did. Late at night, headphones on, I'd watch a teardown of someone else's onboarding flow and feel the floor of my understanding rise an inch.

The classes did something concrete: they gave me the vocabulary and the frameworks to scale. Before, I was making decisions on instinct and vibes. After, I could name what I was doing — and naming it meant I could measure it, repeat it, and fix it when it broke. You can't optimize a thing you don't have words for.

The $100K Money Pit

Here is the most uncomfortable part of the story, the one founders rarely put in their highlight reels.

I invested over $100,000 into this company. And for a long stretch, we managed that money badly.

Most of it disappeared into paid ads. We experimented with them early, the way everyone does, treating ad spend like a faucet you turn on to make customers pour out. The faucet ran. The customers trickled. Apple Search Ads were the worst of it — a beautiful, well-designed money pit. We poured budget into keywords, watched the impressions climb, and got back installs that cost more than the customers would ever be worth. The dashboards looked busy. The bank account looked emptier every week.

I remember staring at a spend report one night and feeling the specific nausea of watching real money convert into nothing.

"It has to start working soon," I told myself, which is exactly the sentence that keeps founders feeding a money pit. It did not start working soon. It did not start working at all.

The lesson took longer to learn than it should have: spending money is not a strategy. Acquiring a customer for more than they're worth isn't growth — it's a slow, well-funded way to go out of business.

We Built the Wrong Product (Twice)

While the marketing was bleeding money, the product itself was built on a guess.

We started by hiring freelancers to write the summaries. Some were fine. Many were flat and generic, missing the spark that makes a book worth summarizing in the first place. And we made a second, bigger assumption baked right into the format: we built 5-minute summaries, certain that busy people wanted the shortest possible version. Faster is better, we reasoned. Just give them the gist.

We were wrong, and we only found out because we finally stopped guessing and started asking.

We built feedback forms and actually read them. We instrumented everything with Amplitude — and I can't overstate how much that mattered. Amplitude's interface is clean and intuitive, making analytics feel less like forensics and more like a conversation. We tracked everything: where people dropped off, which summaries they finished, what they opened twice. For the first time, the customer became a reality.

And the customer told us, clearly, that 5 minutes was too thin.

"It feels like a tweet, not a takeaway," one piece of feedback read. People didn't want the gist. They wanted to actually learn something — but they didn't want to commit to the whole book either. They wanted the sweet spot.

We found it in 15 minutes. Long enough to deliver real substance, short enough to finish on a commute. The moment we landed there, completion rates climbed, and the complaints quieted. Fifteen minutes wasn't a compromise. It was the product.

We paid customers $100 to interview them. It was hard to get free feedback.

We read competitor reviews to see how we could differentiate ourselves.

When customers canceled, we sent them all an email saying that the first person to respond with the reason for their cancellation would win a $100 Amazon gift card. A lot of them responded.

What Actually Brought Users In

Once the product was right, the question became how to grow without lighting money on fire again.

Paid ads were out. What actually worked was almost embarrassingly organic: Reddit, where real conversations in the right communities sent us genuinely interested readers. SEO, the slow compounding kind that keeps paying long after you've done the work. ChatGPT, which — to my surprise — started recommending us when people asked it for ways to read more. And the oldest channel there is: word of mouth, the quiet engine that only turns when the product is actually good.

We also tried lifetime-deal platforms. We partnered with both AppSumo and StackSocial, and I'll be honest about the tradeoff. Lifetime memberships were not sustainable for our business — you can't fund a recurring-cost company on one-time payments forever. But those partnerships flooded our app and website with users, and that visibility was worth something real. They were a spotlight, not a salary.

The lifetime partnerships also brought a ton of backlinks, which helped with SEO.

Then came the decision that embarrasses me a little, because it was so obvious in hindsight. For a long time, we believed an iOS app was enough. iPhone users, premium audience, why bother with the rest?

On a whim, we built an Android app.

It made us a lot of money. An enormous, previously invisible audience had been standing right outside the door the entire time, and we'd never opened it. Sometimes the highest-leverage move isn't clever. It's just finally doing the obvious thing you'd been talking yourself out of.

Borrowing the Onboarding Flow That Printed Money

I spent a season doing nothing but studying competitors. Not to copy them blindly, but to understand what was working in their machines that wasn't working in mine.

One pattern jumped out: the apps that were winning had serious onboarding flows. Not a quick "welcome" screen, but a guided sequence that asked you about yourself before showing you anything.

So we built one. During onboarding, we asked users what kinds of books they loved, what they wanted to get better at, and — crucially — how they'd heard about us. Then the app reshaped itself around their answers.

The effect was immediate, and it taught me something about psychology I hadn't appreciated. People started paying not because the app had more content, but because it felt like theirs. The questions created a sense of personalization, and a personalized experience feels worth paying for.

The onboarding did double duty, too. That "how did you hear about us?" question quietly became our best attribution tool. The dashboards told us numbers, but the users told us the truth about which channels actually worked. We finally knew where our best people were coming from.

When ChatGPT Came for Our Lunch

Then the ground shifted under the entire category.

The market got saturated — new book summary apps kept launching, each one a little cheaper, a little louder. And worse, something more fundamental happened: people started getting their book summaries directly from ChatGPT. Why pay for a summary when you can ask a free chatbot to produce one in seconds?

It would have been easy to panic. For a few weeks, I did.

But you don't beat AI by pretending it isn't there. You beat it by giving people something a blank chat box can't. So we leaned in. We integrated AI features directly into the app — Ask a Book, where you can interrogate a title and get answers grounded in its ideas, and an AI Personal Reading Plan that builds a path through our library based on your goals.

People loved it. The very technology that threatened to make us obsolete became a reason to stay.

We pushed differentiation everywhere we could. Text, audio, and video summaries, so you could read on the train, listen on a run, or watch over coffee. Quizzes to make the learning stick. And we found new ways to make money beyond subscriptions: we charged for infographics, added personality and intelligence quizzes people happily paid for, and monetized the AI integrations directly. The lesson there was simple — when one revenue stream gets commoditized, you don't fold, you diversify.

Going It Alone, and Finding the Real Business

Through all of this, I lost my co-founder.

He decided to focus on his venture-backed company instead — a fair, even sensible choice, but it meant the company I'd been building shoulder-to-shoulder with someone was suddenly mine to carry alone. There's a particular loneliness to running a startup solo that nobody warns you about. The Slack messages you used to send a partner, you now just think to yourself.

But being alone also clarified things. With no one to debate, I could move on conviction. And my conviction pointed somewhere we'd barely explored: corporate partnerships.

Instead of fighting for individual subscribers one Reddit post at a time, I started selling Sumizeit to organizations — companies that wanted to give their teams a real, structured way to keep learning. The economics were everything consumer acquisition wasn't: larger deals, stickier relationships, revenue that didn't evaporate the moment an ad campaign ended.

That was the deciding factor. Not the onboarding, not the AI, not even the 15-minute format, as important as all of those were. The thing that finally tipped Sumizeit into durable profitability was deciding to stop chasing pennies and start building partnerships.

What the Whole Mess Taught Me

Today, sumizeit.com is a profitable company. I can write that sentence plainly now, but I never want to forget what it cost.

It cost a launch full of bugs because I didn't know what I didn't know. It cost a year of silence because criticism felt scarier than failure. It cost more than $100,000, a chunk of it set on fire in ad accounts. It cost two wrong guesses about the product before we found 15 minutes. It cost a co-founder.

But every one of those expensive mistakes paid for a lesson I now consider an asset. Tell founders the truth early. Spending isn't strategy. Stop guessing and ask your users — then actually instrument the answers. Do the obvious thing (build the Android app). Make people feel the product is theirs. When the technology that threatens you arrives, absorb it instead of denying it. And when the consumer game gets brutal, look up — sometimes the real business is selling to the companies, not the crowd.

If you're in the painful middle of your own version of this, here's the only reassurance I can honestly offer: profitability rarely arrives as a breakthrough. It arrives as a long accumulation of corrected mistakes. Mine just happened to be about books.