Before Greatness, Learn to Sell

Last month, I served as a marketing mentor at the Apple × Zhejiang University Mobile App Incubation Camp, teaching a day and a half to over 20 startup teams.

All the participants had products. Some had already launched their apps, others were still polishing their MVPs. Almost everyone came with the same question: “I’ve built the product — how do I get people to know about it?”

I covered a lot in the lecture. Here are the points I find easiest to overlook — sharing them with you.

1. The App Store is not a mall.

Users open Taobao because they want to buy something. When they open the App Store, that’s no longer the case.

The current path is: users scroll through Xiaohongshu, Douyin, or Bilibili, come across content that resonates, remember the app’s name, and then go to the App Store to search and download.

The App Store is a warehouse; content platforms are the shelves. You need to appear on the shelves first, and then users will go to the warehouse to pick up your product.

This is not an exaggeration. By 2026, there are almost only two growth paths left in the domestic market: App Store editorial recommendations and content platform seeding. The former is uncontrollable, but the latter is something we can take into our own hands.

2. The essence of marketing is translation.

Tech teams have their own language: “I implemented a four-quadrant task manager, supports iCloud sync, based on the Eisenhower Matrix.”

Nothing is wrong, but users don’t get it.

Users only care about themselves. They don’t care how complete your features are — they only care, “What does this have to do with me?”

Content translation works like this:

• Feature layer: Supports iCloud sync.
  Value layer: Your records won’t be lost. Available on your phone and iPad anytime.

• Feature layer: Based on the Eisenhower Matrix.
  Value layer: Always being pushed around by urgent tasks? This tool helps you see what’s truly important.

Translation means making everything tangible. Every sentence should make users feel, “This is about me.”

There’s a basic positioning formula: We help (who), (when), solve (what problem).

Not “a note-taking app,” but “help people who easily get emotional become happier by scientifically recording good things.”

3. Find specific people, not an imagined demographic.

Not “efficiency tool users,” but: administrative teachers at schools who need to write papers, teach classes, and handle various chores — they need to see priorities clearly in a four-quadrant view.

This person is real, findable, and may even be right beside you.

Understand users across three dimensions:

• Scenic: When would they use it? Place the feature into a user’s day — you need to be able to describe the exact moment they use your product.
• Emotional: What feeling do they want to resolve? What users pay for, in the end, is to feel better.
• Identitarian: What kind of person does using your product make them? A user of Hao Shi Fa Sheng (Good Things Happen) is “someone who loves life.” A user of Flomo is “someone who loves learning and thinking.” If your product strengthens a user’s self-identity and makes them willing to share that identity, you have the foundation for organic word-of-mouth.

4. The cover image is the line between life and death.

If your notes have low exposure, 100% of the time it’s a content problem, and 90% of content problems are related to the cover image.

If the cover doesn’t make users want to click, everything else is useless.

Users scroll through dual-column feeds extremely fast. The cover’s job is to spark the impulse in less than a second: “This relates to me, I need to click in.”

Here’s an exercise: browse Xiaohongshu with a “shelf mindset.” Which post did you click on? Why did you click? Go back and look at that cover again — deconstruct what it did to you. Do this enough, and you’ll develop an intuition for covers.

5. Content type doesn’t matter — alignment with your product’s tone does.

Popular posts generally fall into a few categories: highly structured information aggregation (dense, informative covers), personality-driven problem targeting (direct talking-head videos with subtitles highlighting pain points), aesthetic (visuals themselves are the selling point), or suspense/hook types that clickbait.

None is inherently better. The key is matching your product’s tone, then doing it consistently and validating with data.

Viral hits are probabilistic, but content is scientific. Your goal isn’t to place a single bet, but to build a mechanism that can repeatedly produce viral elements. Once you find an effective element, don’t use it just once. Break it apart and recombine it — same scene × different emotional angles, same pain point × different audience descriptions. Produce by multiplication.

6. Search keywords are not tags — they are clues buried in content.

According to official public data, 70% of Xiaohongshu’s monthly active users engage in search behavior. Users often search for need-based terms before brand terms. If you can appear in the results at the moment a user searches, the customer acquisition cost is extremely low.

Embedding search keywords means naturally and repeatedly mentioning the words users would search for — in the cover, title, and body text — not just stacking hashtags at the end.

Keywords to cover roughly fall into these categories:

• Audience terms: office workers, students, 30-somethings.
• Scene terms: overtime work, travel, bedtime, early morning.
• Product terms: your app name and its aliases.
• Competitor terms: so users can find you when searching for competitors (provided your product is worth recommending — otherwise you’re just sending traffic to competitors).
• Category terms: productivity tools, mental wellness, task management.

You don’t need to include all of them in every post, but at least a subset. The more posts you have, the broader the keyword coverage, the higher the probability of being found.

7. Paid promotion is an amplifier, not a lifeline.

99% of failed paid campaigns are due to poor content.

When many developers see poor data, their first reaction is to adjust the promotion strategy. That’s the wrong order. Paid promotion is just an amplifier — good content gets amplified better, bad content gets amplified worse.

Invest 75 RMB on Xiaohongshu or 50 RMB on Douyin. What you’re buying with that money isn’t traffic — it’s a conclusion: “Should I continue spending on this content?”

If ROI is positive, keep spending on this piece, and start producing similar content, because it contains viral elements. When promotion isn’t working, fix the content first, then think about promotion.

8. Understand above-the-board vs. below-the-board collaborations.

Above-the-board notes are transactions with influencers through official platforms like Pgyer (Pugongying), with the platform taking about 10% commission, and the content is labeled as an ad. Below-the-board collaborations are private direct deals — influencers charge lower rates, and the content appears more like authentic sharing.

For app promotion, hard ads often perform worse than soft, genuine sharing. On a community like Xiaohongshu where authenticity matters, content that looks like an ad is less likely to be trusted. Go below-the-board when possible, but ensure the content quality itself is good.

Regarding blue V certification (business accounts): For app businesses, proactively certifying is not recommended. Content posted from a certified account is recognized as corporate ad traffic, limiting organic reach. Additionally, the conversion path for apps doesn’t require the private message traffic channel. Unless a platform salesperson actively reaches out to you, certification is more loss than gain.

9. Write a brief before contacting influencers.

Influencers are not just amplifiers — they are discoverers of product highlights. Some of the best product selling points are not summarized by you, but translated by influencers.

You must write a brief before reaching out: your product’s core concept keywords, design intent, key highlights, and what you don’t want them to say. The brief is not a script — it’s a framework for them to work within.

Find influencers of different sizes at different stages. Early on, prioritize micro-influencers and KOCs with fewer than 1,000 followers — lower cost, more authentic content. Once you have proven content that scales, consider mid-tier influencers for amplification.

10. The logic of seeding with real users.

Relying solely on the founder’s account and the official account creates too few channels. Higher leverage comes from: getting more people, from various identities, to speak for your product.

The logic of seeding with real users is straightforward: the weight of an individual account doesn’t matter. When you have enough accounts with average health, the overall health emerges.

To find seed users, you can post “recruitment posts” to attract creators willing to exchange. Exchange collaborations offer the most cost-effective way to scale: give away an annual membership in exchange for a note, find real users to share real experiences. The first batch of UGC comes from this. We’ve built several successful cases using this method.

11. Pricing should be logical, not based on gut feeling.

Many developers price by looking at competitors and setting it slightly lower.

That logic is flawed. Pricing is part of your product positioning — it determines what value users perceive you to have.

You don’t have to charge the same price for domestic users and US users. The same feature can have completely different pricing strategies for different markets and user groups.

Be clear about whom you’re serving, then decide how much to charge. Don’t set a price first and then look for users.

12. Private domain is not a WeChat group — it’s your true user asset.

Platform traffic is borrowed. Today it’s Xiaohongshu; tomorrow it could be something else. What truly belongs to you is the user asset you can manage continuously.

For an app, private domain doesn’t have to be a WeChat group — it can be: subscription usage data from App Store users, core user groups willing to participate in testing, early users willing to leave authentic reviews.

After-sales service is also part of this. If designed well, the refund rate can be very low. If not designed, users will go straight to Apple to request a refund if they’re unhappy, and you’ll have no chance to intervene. After-sales is not the final stage — it’s the last line of defense to retain users.

13. Do the math.

Whether content is good or promotion is worth it is not judged by gut feeling — it’s calculated.

Every budget should have a clear question it needs to answer: Is this content worth continuing to spend on? Are there reusable viral elements in this piece? Can ROI turn positive within two months?

For OPC developers, turning positive within two months is a rough but practical benchmark for health. If you invest money and see no positive ROI signal after two months, either the content is the problem or the product doesn’t solve a real need. Both situations require serious attention.

Good content operations don’t start from scratch every time. It should be a documented, reusable mechanism: a playbook of scripts, brief templates, content element tables, A/B test records. Once built, this system can be handed off, outsourced, or even run by AI.

14. Growth should be front-loaded, not something you think about after launch.

In class, I asked every team: “When did you start thinking about growth?”

Most answered: “After launch.”

You should find users before writing the first line of code. Not test users — real people who will pay for it.

The belief that “if the features are strong enough, users will naturally discover you” is an easy trap for developers. I’ve been down that road myself.

🍎

The above applies beyond apps. If you’re in e-commerce or lead generation, the same thinking can be reused.

That’s all for now. Hope it helps.