3 a.m. and No One Can Tell You If You’re Right

0

3 a.m. The office was empty except for him. The blue glow of the screen fell across his face. The backend data curve lay flat, motionless—like a dead electrocardiogram line. He had just paid this month's salaries—swiping his own credit card. Not the first time.

He scrolled through his contacts, then put the phone face-down on the desk.

No one could give him the answer.

Last month, an old friend who worked in product management had texted him: "Your direction is fine, just hold on a little longer." Nice words, but useless. That friend clocked out on time every day, and his salary hit his account every month on schedule. He had no idea how many times a person could ask themselves the question: Where is today's money coming from?

He stood up and walked to the window. The convenience store downstairs was still lit. The cashier was leaning on the counter, scrolling through a phone, motionless. He used to think that kind of life was dull. Now he envied it a little—that person knew that at 8 a.m. tomorrow, the register would still be there, the shift would still be there.

Three years into his startup. He'd lost the down payment on an apartment, two relationships, and most of his friendships. Over time, he came to understand one thing: the worst part wasn't the lack of money. It was waking up every morning to the same question, and that question never had a standard answer.

Had he chosen the right path? Should he keep pouring money in? Had the direction been wrong from the start?

No one warned you. The world didn't pop up a window saying, "You have deviated from the route, please make a U-turn." It just stayed silent—and by the time you realized you were wrong, the price had already been paid.

1

Zhang Qiang, 29, held a master's in software engineering. He used to be a product manager at a big tech company. Three years ago, he quit to start a knowledge-payment platform—not the kind that sold courses, but a "deep Q&A community connecting experts and users."

He felt the market had no truly good deep-knowledge services—only clickbait, quick-fix courses, and scams. He wanted to build something that "brought knowledge back to its true value."

The team had seven people—all former colleagues and old classmates he'd pulled in one by one. They took salaries at 60–70% of industry standard in exchange for equity. They often worked until 1 or 2 a.m. The product launched: clean UI, smooth flow, dozens of experts signed. Beta users said, "Amazing," "Finally someone's doing this," "I'll definitely pay."

And then? Three months after launch. 20,000 registered users. Conversion rate—0.7%. One hundred and forty people had paid.

It wasn't that the product was bad. Retention was decent, and repurchase among paying users was above 30%. The problem was: most people thought it was good, but they couldn't cross the threshold where "good" became "worth pulling out my wallet."

He tried discounts, annual plans, and promotional pricing for the first thousand users. Nothing worked. Paying users grew from 140 to 180—over two months. Monthly fixed costs were over 200,000 yuan. The company had about 900,000 left in the bank—at this rate, they could last just over four months. And that assumed no one asked for a raise.

He pitched to seven or eight investors. The feedback was eerily uniform: "We like the direction. Run the numbers a bit longer." Translation: your user base is too small—we're not betting yet.

He started losing sleep. Lying in bed, staring at the ceiling, his mind churned: wrong pricing? wrong target users? did this market even exist? Each question made him more anxious, because none of them could be verified immediately. Verification took time—and time was what he was running out of. The hardest part wasn't not having answers—it was not even being able to pinpoint where the problem was. Like reaching out in a pitch-black room, touching nothing, not knowing whether you were facing a wall or a cliff.

2

He remembered a story he'd read.

QQ had been around for two years. Users had exploded—millions chatting online, servers crashing every day, Ma Huateng worrying daily about hosting fees. How worried? He tried to sell the company for 3 million yuan—and no one wanted it.

Back then, QQ was exactly where he was now: people used it, the numbers looked good, but no one knew where the money would come from.

They tried ads. In 2000, advertisers didn't understand the internet. You say you have millions of users? Where? How do I know they even see the ads? They tried an enterprise version—no one bought it. They tried charging for registration—users fled to free ICQ.

How did they survive? Not through some grand innovation. It was Monternet—China Mobile's SMS platform. Tencent tied itself to the carriers' SMS billing system: 5 yuan a month, users chatted via SMS, and Tencent took a cut.

Parasitism. The most "undignified" decision Ma Huateng ever made. He turned Tencent into a cog in the carrier ecosystem, catching a breath through someone else's channel.

Zhang Qiang knew the story. He'd even wondered: did Ma Huateng lose sleep in those years? Did he doubt whether what he was building would ever make money?

He picked up his phone, then put it down again. 3:30 a.m. He remembered that in the second month after launch, the youngest developer on the team—a kid who'd graduated two years earlier—had casually asked over lunch: "Qiang, when are we going to get full pay?" The tone was the same as asking "Are we having a meeting this afternoon?"

He'd smiled and said: "Soon, soon."

That kid had left last month. Before leaving, he said: "Qiang, what you're doing is really great. Honestly. But I have a mortgage to pay."

Zhang Qiang couldn't hold him back. He could barely pay his own mortgage.

3

Among the investors, only one—Old Zhou—told him the truth.

After forty minutes of conversation, Old Zhou leaned back in his chair and said: "Zhang Qiang, I've looked at your product. It's good. But let me ask you something: when a user pays today, does that money go into your pocket, or are you just collecting it for someone else? What exactly is your role in this transaction? Why does anyone have to go through you?"

He froze.

Old Zhou continued: "You've signed all these experts. Why would they stay with you? Because you split revenue with them? Then what if another platform offers two points more? Do you have something that's only available here? I don't mean content—I mean relationships—your relationship with users, your relationship with experts, your relationship with money. Have you thought about that?"

He hadn't thought about it. Or rather, he'd been avoiding it. He was building a "matching platform"—users find experts, experts answer, platform takes a cut. Sounded logical. But when you looked closely, there was no indispensable reason for this structure to exist. Users could reach experts directly. Experts could sell consultations directly. Why did they need him in the middle?

Unless—unless the platform created some value that neither side could achieve alone. Trustworthiness, for example. Or curation. Or a knowledge base that became more valuable with use. But all of that took time, scale, and money to burn before that "valuable thing" was built.

He'd circled back to the same deadlock: he needed money to prove he could make money.

You knew where the exit was, but you might starve to death before you got there.

4

He started having insomnia—night after night.

One night at 2 a.m., he got up again to check the backend. A user had just posted a question about parenting education. The expert had replied with over 400 characters, very detailed. The user responded: "Thank you, teacher. Very helpful."

Just that one line. He stared at it for a long time.

"Very helpful"—the user saw value. The expert was willing to answer. Both sides were right—and he was the one in the middle trying to collect the fee. He felt redundant.

He wanted to quit. Not because he was afraid of failure—he'd already made peace with that possibility. It was the thought that this whole thing might have been wrong from the start. If that was the case, continuing was just wasting everyone's time.

But he couldn't bring himself to say it. Seven people had followed him for three years, taking low pay. If he now said, "Guys, I was wrong, this isn't going to work," it wouldn't just be a company shutting down—it would be him wasting their youth.

That feeling was worse than the money.

The next morning, he still showed up at the office at 9 a.m. He ran the stand-up, talked about new feature plans, kept his tone steady, even cracked a couple of jokes. Everyone laughed. No one knew he'd been up until 4 a.m., no one knew he was thinking maybe I should just call it quits.

At noon, he went downstairs alone to buy lunch. He stood in front of the chilled-food section, hesitating for three seconds over which box to pick. In those three seconds, he suddenly realized he hadn't hesitated over something that trivial in a long time. His mind was full of big things: next month's payroll, investor meetings, whether to give up the office. No room left for "what do I feel like eating today."

He took his bento to the checkout. The person behind the counter was the same one he'd seen from the window at 3 a.m.—still leaning on the counter, scrolling a phone. Now it was noon, and that person was still there. Zhang Qiang wondered: how long had this person been doing this job? Standing by that cooler every day, watching people come and go. Did he ever lose sleep too?

5

Another month passed. The account had about 500,000 left. The team had lost two—five remained, and he didn't hire replacements.

He made a decision.

No more "deep knowledge community." No more chasing the glossy vision. He pivoted the platform to a simpler model: an online booking and payment tool for expert consultations. Experts set their own prices and brought their own traffic. The platform only handled transaction security and workflow management. The cut dropped from 30% to 8%. He was no longer in the content business—he was in the tools business.

This was a universe away from his original ideal. "Bringing knowledge back to value"? Now he was just clearing payment hurdles for experts and users. Like an accountant. But he'd run the numbers: as long as experts and users were willing to complete transactions through his tool, even at an 8% cut, volume would eventually cover costs. Thin margins—but real money.

He called his five core experts one by one: "I've changed the model. Open a booking page on my platform—users book your consultations through my channel, and I take 8%. I handle contracts, invoices, and guarantees. Want to give it a try?"

The first said yes, let's test it. The second said: "8% is okay, lower than other platforms." The third hesitated, then said: "I was going to set up WeChat bookings myself, but it's too much hassle—you do it for me, fine." The fourth hung up on him. The fifth said: "You should have done this a long time ago."

Two days. Three orders. Real, tangible orders. The platform took less than 200 yuan in fees—but it was money. He stared at the deposit notification for a long time. It wasn't excitement—it was a feeling he hadn't had in a while: he finally knew what made money flow in. The path was narrow, crude—compared to his original dream of "transforming the knowledge service industry," it looked like a street stall.

But a street stall was still a business.

6

Another month passed. Daily transaction volume grew from a few hundred yuan to two or three thousand. Experts started bringing other experts, and users slowly multiplied. Small, but at least it kept breathing. He moved the office from the CBD to a co-working space in the suburbs, saving 8,000 yuan a month in rent. The team was down to three: himself, one full-stack developer, and one operations person. The other two had left—one went to a big tech firm. Before leaving, he shook Zhang Qiang's hand and said: "Take care, Qiang."

Zhang Qiang smiled: "Go make your money first. When I get bigger, I'll hire you back."

He said it without much conviction. But he said it. What else could he do? If he collapsed here, who would be left for the others?

Now, every morning, his first thought was still where is today's money coming from? But the answer was more concrete than before—from the consultations completed through his platform, from that 8% cut. Thin, but there. He could roughly estimate how many orders today might bring in, what the day's fees might be.

The uncertainty remained. But he was learning to live with it. Not by finding the answer—by learning to keep searching when there was no answer.

One night he worked late again—not to tweak requirements (the tool was simple, and users weren't asking for much). He was going through transaction records one by one. One user had booked three consultations: the first on career planning, the second on resume revision, the third asking, "Should I start my own business?"

The expert had replied: "If you're still thinking about it at 3 a.m. every night, go try it. Just don't use your parents' money."

Zhang Qiang stared at that line for a long time.

3 a.m. He rarely stayed up till 3 anymore. His body couldn't take it. He'd learned one thing: some things you couldn't power through—you had to find a way around them.

He shut down the computer and walked to the window. The convenience store lights were still on. The cashier was someone new—he didn't recognize him. He thought: maybe that person had started a business once, maybe not. Who knows.

Entrepreneurship wasn't "perseverance." Perseverance was gritting your teeth and waiting for an outcome. Entrepreneurship meant you never really got that "outcome"—you just kept making choices. If you chose wrong, you adjusted. If you couldn't adjust, you changed course. Changing course wasn't shameful. What was shameful was stubbornly walking into a wall.

He glanced at the backend again. That user had asked "should I start my own business," and the expert had said "give it a try." He remembered the day three years ago when he'd quit his job. His manager had asked: "Are you sure?" He'd said: "Yes, I'm sure."

He hadn't been sure at all. He just knew that if he didn't try, he'd regret it for the rest of his life.

He turned off the lights, locked the door, and went downstairs. Passing the convenience store, he stopped to buy a bottle of water. The cashier glanced up at him and said nothing.

The street was quiet at this hour. He twisted the cap open, took a sip, and walked toward the subway station. Tomorrow he still had to go to work—his own company, his own shift. It actually didn't feel bad. Even though he still didn't know whether the company would be there this time next year.

But he did know that tomorrow at 9 a.m., he'd be sitting back in that co-working space, opening his laptop, checking data, replying to messages, making calls. Doing all those small, unglamorous, uncool things. Over and over.