Rebirth 10: I'm a Promoter at Tencent

4:00 AM, Xiaomi headquarters in Beijing.

The conference room lights were a cold, white, like the operating room lights. Lei Jun sat alone at the far end, with three freshly printed documents spread out in front of him. The edges of the pages were still slightly curled, and the paper was warm from the printer.

First report: "MiTalk Registration Process Optimization Team's First Day Progress Report"

Second document: "MiTalk 1.0 Feature List (Simplified Version)"

The third document: "Preliminary Investigation Report on 'Tencent Pioneer Lab' and its Person in Charge, Lin Shen"

He picked up the first report.

The header indicates a time of 3:40 AM, just after the team submitted it. The report is honest, even brutally honest:

Current bottleneck: SMS channel delay.

WeChat uses a financial-grade channel with an average latency of 1.8 seconds and a delivery rate of 99.7%.

We do not have Tencent Channel Level, but we can call Enterprise B Level, with an average latency of 8-15 seconds and a 92% arrival rate.

Directly replicating the WeChat user experience is technically not feasible at the moment.

Lei Jun's gaze lingered on the words "infeasible" for a full ten seconds.

The red pen twirled between his fingertips but didn't fall. He put down the pen, leaned back in his chair, and closed his eyes.

It's not about giving up, but about mobilizing a different mindset—not the "technical feasibility" mindset of an engineer, but the "user experience reconstruction" mindset of a product person.

Is 1.8 seconds really what users want?

Do they really care about the difference between 1.8 seconds and 6 seconds?

No.

Users want a feeling—a feeling of "speed," a feeling of "smoothness," and a feeling of "not causing delays." If it's technically impossible to achieve 1.8 seconds, then make it feel "very fast" in terms of perception.

He opened his eyes and wrote two lines in the margin of the report:

"Can six seconds of physical time be transformed into three seconds of psychological time? Waiting is not a blank space, but a space for content design. — Lei, 4:02 AM"

He turned to the next page, which showed the optimization results of MiTalk:

1. The installation package can theoretically be compressed from 12.7MB to 7.3MB without cutting core functions, but this is still a huge difference compared to WeChat's 4.3MB.

2. The registration page steps have been simplified from 5 steps to 3 steps, a huge improvement compared to WeChat's one-step registration mode that only requires a mobile phone number.

"It still needs optimization," Lei Jun thought. He was considering using a health checkup today; theoretically, a 7.3MB file would still take 25-30 seconds to download on a 3G network. But WeChat's 4.3MB file could be downloaded in under 15 seconds.

Just then, there was a knock on the conference room door.

"Enter."

The product manager of MiTalk pushed open the door and came in, carrying two cups of coffee. The bloodshot in his eyes was more pronounced than during the meeting that afternoon, but his mind was unusually focused, as if he had been forced to concentrate under high pressure.

"Mr. Lei, you're still here." Huang Jiang placed a cup of coffee in front of Lei Jun, the steam rising in the cold air.

"You're here too." Lei Jun took the coffee and gestured for him to sit down. "I've read the report. The bottleneck is clear, but we need to think outside the box."

Huang Jiang nodded and opened his notebook: "We tried three plans. First, we tried to learn from WeChat's cash-burning strategy, but the cost of acquiring a single user would have exceeded 7 yuan, which did not fit Xiaomi's hardware-driven traffic strategy, so we rejected it."

Lei Jun sipped his coffee without saying a word.

"Second, we will optimize the existing channels technically," Huang Jiang continued. "Our partner has agreed to give us the highest priority, but the limit is only an average of 6 seconds, and we cannot guarantee peak hours."

"Six seconds." Lei Jun repeated the number, "twice as fast as it is now, but still more than three times faster than WeChat."

"Yes, so the focus is on the third thing." Huang Jiang pulled up the design drawings. "Redesign and wait for the user experience."

On the screen, a cartoon character named "Mitu" starts running, with a progress bar above and prompts below.

"The animation lasts for 5 seconds," Huang Jiang explained. "Even if it actually takes 6-7 seconds, the user's perception time is only 3.8 seconds. We conducted an internal test with 20 people, and the data is very stable."

Lei Jun stared at the running rabbit for a long time.

"You're heading in the right direction," he finally said, "but it's not good enough."

Huang Jiang looked up.

"Mi Rabbit Running—this is still 'waiting'." Lei Jun put down his coffee cup. "The WeChat page says, 'The message has been sent and is en route to China Mobile Network...' They are turning waiting into a miniature story. Users are not waiting; they are watching the story unfold."

He leaned forward, tapping his fingers lightly on the table: "Can we take this a step further? For example, matching city landmarks based on phone number ranges—for Beijing users, Mi Rabbit runs over Tiananmen Square; for Shanghai users, it runs over the Bund; for Guangzhou users, it runs over the Canton Tower. Turning waiting into a 'virtual trip'?"

Huang Jiang's eyes lit up: "Technically feasible! We can build a simplified material library and match by number segment..."

"Not only that," Lei Jun continued, "While waiting, we can insert a little tidbit of trivia about the city. For example, we can show Beijing users: 'The text messages Beijing sends every day, if laid end to end, could circle the Earth twice.' This makes the wait more informative and memorable."

He paused for a moment: "The core logic is: users are not 'wasting time waiting for verification codes,' but 'spending a few seconds to watch an interesting little piece of content.' Once the perception changes, the experience changes."

Huang Jiang's fingers flew across the keyboard, recording at lightning speed.

"Installation package size." Lei Jun pointed to the report, "7.3MB is still too large. Continue with the existing compression scheme, but I'll add one more point: establish a '1KB task force'."

"1KB?"

"Yes, we meticulously scrutinize every detail," Lei Jun said. "How much do we save by reducing icon size by one level? How much do we save by compressing and merging CSS? How much do we save by deleting redundant logs? We cut corners KB by KB, and scrape through MB by MB. In this era where data is billed by MB, every 1 MB saved means 5% less user churn."

He opened the second document – ​​“MiTalk 1.0 Feature List (Simplified Version)”.

The list was cut down to just three lines:

Register/Log in with mobile number

One-on-one text messages

Voice messages (basic version)

"That's a pretty drastic cut," Lei Jun commented. "Can it be released in a week?"

"Yes," Huang Jiang nodded, but then hesitated, "but there are dissenting voices in the team who think that launching it this way is too rudimentary and will be ridiculed..."

"Laugh?" Lei Jun laughed, a laugh that showed he understood the true nature of the market. "Huang Jiang, do you remember how many people laughed when the Xiaomi 1 was released? They laughed at the design, at the lack of core technology, at me just being a 'phone seller'."

Without waiting for an answer, he continued, "And the result? Users voted with their wallets. Because they discovered that those aspects that were ridiculed as 'basic' didn't actually affect the core experience. And we invested all the cost savings into what users truly cared about—performance, battery life, and price."

He pointed to the streamlined list: "The same applies to MiTalk 1.0. Users don't want more features, they want a good core experience. As long as registration is fast, chatting is smooth, and voice calls are seamless—if these three points are done to the extreme, other features may be delayed by two weeks. Users may complain, but they won't leave, because their core needs have already been met."

Huang Jiang seemed thoughtful: "So what we need to create is a product that is 'functionally streamlined but offers an exceptional user experience'?"

"Yes, and it needs to be even more streamlined than WeChat," Lei Jun said. "But remember, it's not about blindly piling on features, it's about precise targeting. Why is WeChat so streamlined? Can't Tencent create complex functions? Just look at QQ."

He paused, his tone becoming serious: "WeChat targets three pain points: expensive data, expensive phone calls, and complex applications. MiTalk aims to target the same three, but more precisely and ruthlessly than WeChat. Because we have something they don't—"

"Hardware access," Huang Jiang replied.

"That's right." Lei Jun nodded. "Xiaomi phones are our home turf. In this home turf, MiTalk can better understand users—knowing the device model, the network status, and even the battery level. All this data can be used to make more precise experience optimizations."

Then he opened the third document.

The "Preliminary Investigation Report on Lin Shen's Team" is only two thin pages long, and the information is so sparse that it looks like it has been washed out:

Name: Lin Shen

Age: 22 (suspected)

Background: Recent graduate, Computer Science Department, Shenzhen University (Information from an outsourcing website)

Joined Tencent: July 2010 (only two months)

Previous project: Light project team

Current position: WeChat Project Manager (promoted ahead of schedule)

Team size: 9 people (confirmed)

Office location: A secluded room in Building C, Kexing Science Park

Other: No social media accounts, no recorded speeches, no published papers...

"Like a ghost," Huang Jiang commented.

Lei Jun stared at the words "22 years old (suspected)" for a very long time.

"22 years old." He repeated softly, "Zhang Xiaolong's willingness to entrust a strategic project to a 22-year-old newcomer who has only been with the company for two months speaks two things."

Huang Jiang raised his head.

"First, Lin Shen has something that transcends age and seniority," Lei Jun said. "Second, Zhang Xiaolong is gambling—gambling that the intuition of young people can outrun mature systems, and gambling that the agility of a small team can crush the cumbersome nature of a large group."

He closed the report: "Continue to collect information, but the focus should not only be on the people, but on their working methods. How are the nine people divided into roles? How are decisions made? Also, how can we ensure that nine people work weekly? These are the things we really need to learn."

Outside the window, the sky was beginning to lighten.

At 4:30 a.m. in Beijing, the morning light struggled to penetrate the smog, casting blurry patches of light on the conference room floor.

Lei Jun stood up, walked to the window, and suddenly said, "The WeChat team is probably waiting for the sunrise right now."

Huang Jiang hesitated for a moment before speaking: "According to reliable sources... the WeChat team already left work in the early hours of the morning."

"Off work?" Lei Jun paused slightly in his hand holding the coffee cup.

He turned to Huang Jiang and asked, "Do you know what I admire most about them?"

Huang Jiang thought for a moment: "Is it that kind of extreme simplicity? Or an obsession with user experience?"

"Neither," Lei Jun shook his head. "It's that they dared to sleep at crucial moments."

He walked back to the table and pointed to the optimization report: "And our team is still working all night modifying code. This isn't dedication; it's a flawed system! If the architecture is robust enough, the processes are clear enough, and the monitoring is intelligent enough, the team should be able to sleep soundly, because the system will run itself, problems will be reported automatically, and things that need to be handled will be handled automatically."

Huang Jiang was stunned.

"So what MiTalk needs to pursue isn't just the product experience," Lei Jun said, "but also the entire underlying operational system. Seven days after launch, I want to see not only a usable product, but also the prototype of a self-sustaining system. Understand?"

"Understood!" Huang Jiang stood up straight.

"Go ahead." Lei Jun glanced at the time. "Tell the optimization team to go home and get some sleep. They can come back at 10 a.m. tomorrow. This is my notification. I need their clear minds, not exhausted bodies."

Huang Jiang nodded, packed his things, and left.

As he reached the door, Lei Jun called out to him.

"Huangjiang".

"Mr. Lei?"

"It will launch in seven days. If the data is not as expected..." Lei Jun paused, "I will take the responsibility. But you must remember: this is not the end, it is the beginning. MiTalk will be updated every two weeks, continuously improving. This is a protracted war."

Huang Jiang nodded emphatically, pushed open the door, and left.

The meeting room fell silent again.

Lei Jun stood alone by the window, watching the city gradually awaken in the morning mist. In the distance, car lights began to flow along the streets, like rivers of light.

He picked up the report about Lin Shen again.

The paper is light, but the information on it carries great weight.

22 years old. Recent graduate. Two months on the job. Team of nine.

Every word says: This is unreasonable.

But the data of 40,000 new users on WeChat's first day of launch also suggests that this is reasonable.

Lei Jun looked at the name on the report and suddenly smiled.

It wasn't a mocking laugh, but the kind of laugh that comes from meeting a worthy opponent, a mix of wariness, appreciation, and fighting spirit.

"Lin Shen..."

He whispered the name.

"I'll remember this sneak attack!"

-----------------

At 10 a.m. the next day, the doors to the "garage" were opened one after another.

Cheng Xiangdong yawned, carrying coffee and bread he'd bought from the convenience store downstairs. Sun Hui's eyes were still a little red, but he seemed much better than he had been in the early hours of the previous night. Wu Feng was already sitting at his workstation, debugging something on a Symbian emulator.

Li Yue and Lu Chuan entered almost simultaneously, one carrying a laptop and the other holding a mobile phone, the screen displaying the latest discussions about WeChat on Weibo.

Lin Shen was the last to arrive, at exactly 10:00, not a second later.

He wasn't carrying any coffee or food; he just had a thin black notebook tucked under his arm.

He walked in, his gaze habitually sweeping over the monitoring screens on the wall—the number of users had quietly surpassed 45,000. After slowing down overnight, the growth curve began to rise again with the arrival of morning, but the overall picture remained healthy. Server load was stable, and the SMS channel latency monitoring was all green.

"Good morning." Lin Shen's voice was calm, showing no signs of staying up all night, as if he had just experienced an ordinary night.

"Good morning, Brother Shen!" Lu Chuan was the most energetic. He had gone to bed the earliest last night and was full of vitality. "The discussions on Weibo about our 'Garage Legend' are getting more and more bizarre. Some people have even figured out which takeout we order from!"

"None of that matters. What they talk about most is our promise to work weekly, saying, 'Let's see how many weeks the nine of us can keep it up!'"

"Let them guess slowly! Try to set the pace and see if we can get them to vote spontaneously."

Lin Shen briefly assigned Lu Chuan the information gathering task, then walked to the whiteboard...