Alright, if you’re also interested in building global web apps, let’s dive right in. Starting from this section, I’ll systematically break down everything I’ve experienced on this journey—covering common beginner questions, my own products, tips and tricks for driving traffic, pitfalls I’ve stumbled into, and my future roadmap. I’m laying it all out on the table, so feel free to jump straight to the sections most relevant to you.
1. “I Don’t Know What to Build”
As a beginner, it often feels like everyone else is shipping websites left and right while you’re still stuck wondering what to build. Don’t beat yourself up—you’re not alone. This is the most common hurdle for newcomers, and it’s completely fixable. Here are a few proven ways to break through it:
Study the Leaderboards
Get intimately familiar with the top-ranking products on major directories. In a livestream, mentor Liu Xiaopai once mentioned that he knew the monthly leaderboard on Toolify so well he could instantly tell you what every single product on that list did. That blew everyone away, and ever since, studying leaderboards has become a staple practice in our community. Another builder, Liangchenmei, even started dissecting leaderboard products one by one on his WeChat blog—talk about peak execution.
How to study them:
- Platforms: Toolify, TAAFT (There’s An AI For That), Traffic.cv, etc.
If you’re a beginner, I recommend analyzing leaderboard products across the following dimensions. Many people use AI to speed this up, but you’ll get the most out of it by doing the legwork yourself:
- 💡 What problem does this product solve?
- 👤 Who is the target user?
- 🤔 Why do users need it?
- 🗣️ What are users saying about it in the reviews?
- 🔍 How does it acquire users?
- 💰 Is it making money? How much?
- 🧠 What did I learn from this product?
- 🤔 What part of its execution is hard to replicate?
- 🤗 One-sentence elevator pitch:
- 💡 Alternative approaches:
- 🎉 Can I build this?
- 🧭 How would I find users for it?
- 🤔 Why am I the right person to build this?
- ❤️ Can I stick with it for the long haul?
Of course, if you’re a seasoned builder, you can selectively filter and study product categories that align with your target niche.
Keep a Daily Idea Log
Make it a daily habit to jot down user needs and product ideas whenever they strike. This method comes from Pieter Levels—the legendary indie hacker on X (Twitter) and author of MAKE, a book that’s practically a bible in our space.
How to do it in practice: Force yourself to log ideas every single day—say, at least three a day. Still drawing a blank? At its core, that just means you haven’t built the mental habit yet.
The truth is, once you start paying attention, you’ll have more ideas than you could ever build. Tune into everyday complaints and pain points—whether it’s a minor annoyance in daily life, friction at work, or something you notice while scrolling through short videos or playing games.
Take scrolling through short videos as an example: I intentionally like and bookmark AI-generated videos to train the algorithm to feed me more of them—especially viral AI reels on Instagram. The comments or captions often reveal which APIs were used behind the scenes, which gets me thinking: Can I chain different APIs together to build a product that mass-produces this exact style of viral video?
By the way, you’ve probably seen those crazy eye-catching video hooks—where a fail clip seamlessly transitions into someone doing a backflip. Well, someone actually went ahead and built a dedicated stock library website specifically for those universal viral video hooks. Yep, that someone was me.
Scrolling through videos is just one microcosm of daily life; I’ve simply conditioned myself to instinctively ask whether the hook or concept in a video can be productized. In reality, life is full of moments like this. As long as you stay observant and curious, you’ll soon realize you have an endless supply of product ideas. That holds true for building internet products, and it holds true for any other business.
The key is deliberate thinking and keeping an open mind.
By the way, using this exact method, I now have over 100 ideas sitting in my backlog—more than I’ll ever have time to build. Here’s a custom template I made that you can use for reference: [Idea Backlog Management – Copy].
Ride the Wave of “New Keywords”
The first two methods are fundamentally about cultivating long-term product intuition. But is there a faster shortcut?
Yes, there is—hunting for “new keywords.” In our space, new keywords usually refer to newly released AI model names or the ecosystem of wrapper tools and services built around those brand-new models.
This is a golden opportunity for beginners to leapfrog the competition. Catching a new keyword early often delivers fast positive feedback—and for a newcomer, early positive feedback is everything. Because of that, however, new keywords also attract fierce competition, which means speed is the name of the game. Which brings us to the next question: how do you move fast?
- Follow top AI influencers and accounts, browse X daily, and deliberately train your timeline algorithm to surface breaking AI news.
- Beyond manual scrolling, you can also set up AI monitoring tools to track breaking news and releases in your niche 24/7—I’m sure a quick search in the Shengcai community will turn up plenty of tutorials on how to set this up.
- Finally, speed comes down to how fast you can deploy a live site. Have a battle-tested Starter Kit ready to go—the most popular ones in our circle are ShipAny and MKSaas—and keep tweaking it until it fits your workflow like a glove. Turn every launch into a streamlined SOP. With the rise of tools like OpenClaw, you can even automate the entire pipeline—from spotting a product opportunity to coding the app and deploying the site.
That said, I want to add one important caveat: while chasing new keywords helps beginners get quick wins and taste early success, you shouldn’t rely on them forever. Experienced builders aren’t afraid to tackle “old keywords” (evergreen search terms with proven, long-standing demand) and excel at the second method we talked about earlier.
2. Real vs. Fake Demand: Validating Your Idea
Let’s start with a quick story:
Xiao Ming noticed that modern vision models are great at image recognition, which sparked an app idea: snap a photo of any ingredient at the grocery market, and the app instantly generates recipe combinations so you know what else to buy on the spot. Thrilled by the idea, Xiao Ming spent a whole month building it, launched it on Xiaohongshu (RED)—and heard nothing but crickets.
Full disclosure: Xiao Ming was me. The first half of that story is true, while the second half is fictional—because I ended up killing the project before writing a single line of code. I can’t say with 100% certainty that there’s zero market for it, but my analysis told me it was almost certainly “fake demand” (a tarpit idea).
Here’s why: the target audience is contradictory. Not many young people shop at traditional wet markets to begin with, so your narrow target user would have to be someone who goes to offline grocery markets yet doesn’t know how to cook—which is a contradiction in terms. On top of that, everyday home cooking revolves around the same dozen or so staple combinations; once someone uses the app two or three times, they don’t need it anymore, making retention a nightmare. Add in the fact that the pain point isn’t acute enough to make people open their wallets, and I decided to drop the idea.
I’m sharing this example to show that telling real demand from fake demand isn’t always black and white. Forget everyday folks—seasoned product managers and even tech CEOs misjudge the market all the time. History is littered with famous misses: when Musical.ly (the predecessor to TikTok) was raising money, a heavyweight shareholder pitched Pony Ma (CEO of Tencent) twice, and Tencent passed both times. Or look at Mark Zuckerberg burning billions on the Metaverse a few years ago, or Luo Yonghao’s ill-fated bet on AR.
So, is there a reliable way to avoid tarpit ideas as much as possible?
Yes. Here I have to credit another must-read classic for beginners by @Liu Xiaopai: How to Find Product Ideas (https://scys.com/articleDetail/xq_topic/212485241245121). It’s legendary in our community. Here is the cheat-sheet version—if you’re a beginner, just follow these four rules to the letter and you’ll start seeing results fast:
- Scratch your own itch (look up the concept of eating your own dog food).
- Look at proven products already on the market. This doesn’t mean blindly cloning them. Instead, ask yourself: Since this already exists, what is my angle of differentiation? Am I offering better value for money, a smoother user experience, or a specific niche focus?
- Use Xiaopai’s Golden Formula: Who, in what specific scenario, is willing to pay how much money to solve what exact problem? I literally printed this sentence out and taped it to my wall, haha—folks in our group chat even turned it into a meme sticker.
- The “Enemy Acid Test”: Imagine the product you’re about to build was created by someone you dislike—or even despise. Would you still begrudgingly pull out your credit card and pay for it anyway? If the answer is yes, you’re almost certainly looking at real demand.
Finally, let me add one more thought:
A lot of the time, it’s totally fine if you can’t be 100% sure whether an idea is real or fake upfront. Because the barrier to building an MVP today has dropped to the lowest point in human history—you can spin up and ship a working MVP in under an hour to let the market validate the idea for you.
So be bold! If an idea genuinely excites you, just ship it fast and play the numbers game. This brings me back to Liangchenmei again—shipping over a hundred websites a year, banking on sheer volume and relentless execution to hit winners.
3. Core Principles of Product Design
The Principle of Radical Simplicity
Growing up with domestic super-apps that are bloated and over-engineered, it’s easy to forget what a great product is actually supposed to feel like. In my book, a great product should be effortlessly simple to use. Note that “simple” doesn’t mean bare-bones—it can be powerful under the hood, but it must be intuitive for a first-time user, leaving them feeling delighted or even wowed after using it.
Think of classics like TinyPNG, remove.bg, or Anybox. Recently, the product that blew me away the most was Typeless—it’s so good that you almost feel guilty not paying for it.
As Xiaopai puts it: Keep the simplicity for the user, and absorb the complexity yourself—make it so intuitive that even your parents can figure it out at a glance. Next time you’re designing a product, keep asking yourself if you’ve hit that bar.
The MVP Principle
Anyone who has spent time in tech is familiar with the concept of an MVP (Minimum Viable Product). Compared to native mobile apps, web apps have the natural advantage of being lightweight and fast to deploy. Pair that with AI coding tools lowering the technical barrier and supercharging launch speed, and the MVP mindset becomes more powerful than ever.
I won’t bore you with that classic MVP skateboard-to-car illustration, because in practice, the real struggle is how to clearly define the boundary of your MVP. It sounds simple in theory, but in the trenches, most people either draw the line too wide by stuffing in “nice-to-have” features, or start with a tight scope and succumb to feature creep halfway through building. I even half-jokingly thought about building a tool just to help indie hackers ruthlessly scope their MVPs and hold them accountable until launch. What’s that? You think that’s actually a solid idea? Go build it then—ping me when you launch and I’ll be your very first user, haha!
All jokes aside, losing control of your MVP scope usually happens because you haven’t truly nailed down why you’re building the product in the first place—which loops right back to our earlier topic on how to find real demand. The two are deeply intertwined.
Here’s my practical takeaway: list out every feature you think your MVP needs, then run them against the demand-validation criteria we discussed earlier and cross them out one by one until you’re left with the single irreducible core. Compare that against your original reason for building the product. Better yet, if you can reach your target users, show them a rough prototype and talk to them directly—nothing snaps you out of your own echo chamber faster.
One crucial point I want to emphasize: an MVP can look rough around the edges, but its core feature cannot be half-baked. How do you know if the core feature hits the mark? When a target user tries it and immediately has that “Whoa, that’s neat!” Aha Moment—that’s when you know you’re onto something.
The One-Sentence Rule
The One-Sentence Rule asks a simple question: Can you explain what your product does in a single sentence—and does it focus on doing just one thing exceptionally well?
Or in today’s indie hacker lingo: stop trying to build an “All-in-One” platform on day one. It’s easy to nod along to this advice, yet remarkably easy to lose perspective and get derailed when you’re in the middle of building—yep, guilty as charged.
An “All-in-One” suite is almost always the result of evolution, never the starting point. Look closely at 90% of the comprehensive platforms on the market today, and you’ll see they were all iterated brick by brick. Every single one started as a sharp, elegant tool that solved one specific pain point extraordinarily well. Once they retained a loyal user base, they expanded their feature set organically based on real user feedback.
What stuck with me most was hearing Abiao share that Pollo AI didn’t start out as an all-in-one studio either. They noticed that after users generated images from text on their platform, they were downloading those images and taking them to other tools to turn them into videos—so the team naturally added an image-to-video feature. That expansion was pulled by real user behavior, not pushed by a founder daydreaming an “ultimate all-in-one suite” in a vacuum.
There are countless cautionary tales of the opposite—like Google Wave. If you’ve never heard of it, that’s because even Google fell into the exact same trap of trying to do everything at once (ask your favorite AI for the full post-mortem).
If you’ve read all the way to this point, props to you—you clearly have the grit and focus of someone who’s bound to succeed at whatever you take on. Yep, I’m talking about you.
Let’s take a quick breather with a little “demon” story. In the anime Demon Slayer, there’s a fan-favorite character named Zenitsu. Unlike his senior apprentice who mastered almost every form of Thunder Breathing, Zenitsu could only learn the very first form: Thunderclap and Flash. At first, he was deeply insecure and cried all the time. But through grueling training, he honed that single move to absolute perfection—pushing its speed, power, and mastery far beyond any other Thunder Breathing user, and ultimately using that one move to defeat his senior who had turned into a demon.
Greatness Is Iterated, Not Planned
By the way, beginners aren’t the only ones who fall into the “all-in-one” trap. From my years of working in the corporate trenches, I’ve seen plenty of companies—especially bosses flush with a bit of extra cash—kick off projects by grandly declaring, “We’re going to build this, and that, and create an entire XX ecosystem from day one!”
There’s a famous saying that greatness cannot be planned—and we should never forget the natural corollary that follows: greatness is iterated.
