---
title: "The Chinese Robot Dog Mark Zuckerberg Bought via Private Message: Where Does It Come From?"
author: "Unique Research"
sourcePublication: "Unique Research Substack"
originalPublishedAt: "2026-09-02T04:01:36+00:00"
canonical: "https://ffcap.cn/en/research/src-20260902-01html"
source: "https://uniqueresearch.substack.com/p/src-20260902-01html"
language: "en"
---

# The Chinese Robot Dog Mark Zuckerberg Bought via Private Message: Where Does It Come From?

_Original · Unique Research · 2026-09-02_

_Editor's not_

[![](https://substackcdn.com/image/fetch/$s_!b0Rq!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fa88b1f2f-fd8b-4ebf-a9c0-fdb01d320664_2048x1152.jpeg)](https://substackcdn.com/image/fetch/$s_!b0Rq!,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fa88b1f2f-fd8b-4ebf-a9c0-fdb01d320664_2048x1152.jpeg)

_e: The first-person report and its judgments belong to the original Chinese author. This_

[![](https://substackcdn.com/image/fetch/$s_!sCZA!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F9a42084a-7457-4436-a0ce-1e99f634a5b3_2048x1152.png)](https://substackcdn.com/image/fetch/$s_!sCZA!,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F9a42084a-7457-4436-a0ce-1e99f634a5b3_2048x1152.png)

_English rendition retains the opening essay, six insight sections, closing reflection, and the full roundtable transcript. All named companies, products, and people are preserved. Crowdfunding amounts, user statistics, and product claims are speaker claims, not independently verified findings._

Original · Unique Research · 2026-09-02 12:00 Shanghai

AI Industry Observer

Crowdfunding Hits $1M, Time to Celebrate? The Second Half of the Pitfalls Has Only Just Begun

"What you think is the finish line is actually the starting point; the number you think matters most is precisely the one that matters least."

Mark Zuckerberg once used his personal email to buy a robot dog made by a Chinese team for his daughter.

This didn't appear in any press release. The person who discovered it was founder Li Rongzhong (李荣仲) himself. He still watches all after-sales emails, personally replying to about a third, and one day he saw this name in his inbox.

You might ask: a hardware company founder going to customer service — is that really necessary?

Li Rongzhong's answer: absolutely. He says after years doing Support, he's received大量特别 direct and sincere user feedback, plus some意外 surprises, like this Zuckerberg order. His original words: "This kind of thing, you really only encounter when you're personally serving users on the front lines."

At a Shenzhen Physical AI Summit roundtable on AI hardware going global, themed "After the Crowdfunding Hit: How to Truly Reach the World's Shelves." Four people on stage: Li Rongzhong, consumer robots; Wu Danni (吴丹妮), scalp care brand; Gu Mingjun (顾明君), Indiegogo China head; and Tang Chang (唐畅), who's putting his AI ring on Kickstarter next month.

After the whole roundtable, the most valuable part is how they dissected "crowdfunding success" from start to finish, then told you: what you think is the finish line is actually the starting point; the number you think matters most is precisely the one that matters least.

Cutting a Hundred-Thousand-Dollar Robot Down to $200

Start with product.

Li Rongzhong is a Nanjing University physics undergrad, did a physics PhD and CS master's in the US — originally not in robots. After graduating he taught himself hardware programming; one day he put sensors on a robot and saw the "eyes" move, and suddenly had a thought: could this joint also be made into legs?

Back then, robots that could do this kind of agile movement were basically only at labs and companies the level of Boston Dynamics and MIT, costing hundreds of thousands to millions of dollars each.

Li Rongzhong directly cut the price to two or three hundred dollars.

When the product came out, nobody believed it; nobody could believe this price could produce something like this — which instead brought the first wave of massive traffic. He put the little robot cat on Indiegogo, spent zero on ads, and raised $140K. Second campaign, still barely any ads, raised $600K.

But what really impressed me wasn't these numbers, it was the reason he made the product.

He's kept cats, dogs, turtles, fish; petting a kitten, he can't help feeling its bones. He says, for a human and a creation to truly form an emotional connection, a hard box and a screen aren't enough; it needs some biological attributes, ideally also able to move autonomously in 3D space. Only then do you feel it shares the same space and the same time with you.

He figured out he wanted to make pet robots back in 2018. Nearly a decade later, everyone suddenly discovers this has truly become a trend.

"Some judgments are just like this: ten years ahead is called paranoia; two years ahead is called a trend. The only difference is whether you lasted until that day."

90% of Users Give Up Before It Works

Wu Danni comes from marketing, did medical PR, rode the O2O and VR waves. She started a scalp care brand for the most direct reason — simple and blunt: she herself was losing hair, for many years, still a grade-2 hair loss person.

As a "fellow hair-loss friend," she discovered this category has a classic death loop: brainwashed by ads, buy product, come back and don't use it, don't use it and no effect, no effect and depressed, depressed and buy again.

Where's the problem? The time gap. A hair growth product truly taking effect usually takes 3-6 months, but most users can't last that long. She gave a number: maybe 90% of people drop off before they truly see results.

So her entry point is counterintuitive. Everyone else stacks features — massage, red light, drug导入 — three-in-one wanting to be ten-in-one. She cut first, reducing features to the barest experience, solving one thing first: making users willing to keep using it. In industry terms, compliance.

She doesn't even make efficacy claims. Her words: at this stage, first make it a good tool, make the experience good enough and compliance high enough; efficacy, later.

"A brand selling hair growth products, first not talking about hair growth. Sounds like self-harm, but actually it's clear-eyed: in this category, making users persist for 90 days is a hundred times harder than making the spec sheet look pretty."

How to Tell a Real Hit From a Traffic Bubble

Gu Mingjun has been Indiegogo's China head for nearly 8 years; she's seen more crowdfunding projects than most people have scrolled short videos.

I directly asked her the most practical question: landing pages can all be made beautiful; what signals should ordinary users look at to judge whether a project is worth backing?

Her answer is note-worthy: don't look at the page, look at the comments section.

Pages and videos — anyone can make them fancy now; post-production capability is too strong. But the comments section can't be faked. Don't look at templated praise; find real user questions. Like in a storage product's comments, real users ask: I live where it's 30+ below zero, how long does it really last at night? If a bunch of people keep asking around these real usage scenarios, the demand is real.

The second question is more critical: how does the platform judge whether a project is a real new category or a traffic-stacked bubble?

Gu Mingjun's answer surprised me: before crowdfunding ends, you simply can't tell.

Why? Because crowdfunding long ago stopped being the early game of relying on organic traffic and luck. Anker, Ugreen and these giants are all doing crowdfunding; marketing will polish the data beautifully. The real answer is two things to look at later.

First, is there a long tail after crowdfunding ends? Crowdfunding solves 0 to 1; an ideal project by the end should already have a brand, tone, and first batch of users; the later independent site, Amazon should keep growing. If crowdfunding is the historical peak and then declines all the way, be wary.

Second, community vitality. Active during crowdfunding doesn't count; after delivery, after new users come in, whether people still chat in Facebook groups and Reddit communities — that counts.

She gave Plaud as an example. This AI recording device raised over $2M combined on Indiegogo and Kickstarter back then, but the point isn't there. The point is after crowdfunding, it kept surging on Amazon, TikTok, independent sites, and spawned a batch of competitors.

"What Plaud validated for everyone is a category. The crowdfunding amount isn't the most important; what's important is whether the whole market keeps growing after."

Crowdfunding Hits $1M, the Second Half of Pitfalls Is Just Beginning

Next is the most piercing part: delivery.

Gu Mingjun says many people only stare at the amount on the page. Today hit $1M, think it's a success, start the celebration banquet.

"Actually not. The second half of the pitfalls has only just begun."

She's seen too many companies stuck on delivery. The most typical way to die: crowdfunding amount is beautiful, marketing ratio is healthy, then enter delivery and the supply chain is completely different from what was negotiated. Factory boss doesn't honor the earlier quote, wants to raise prices. Crowdfunding profit is already thin; raise price and the whole project has no profit. In the end crowdfunding looks "successful," but goods can't be delivered.

So her advice is very specific: before officially launching, use the service provider's estimated sales to negotiate with supply chain — delivery cycle, capacity, cost, all locked in advance. Product should at least be in mold opening stage, or very close, before going on crowdfunding.

"Why can't you just bring a hand-built prototype now?"

"Because users have been scammed too much." She says, before there were projects claiming "world's smallest projector with highest lumens," raising $2M+, delivering something completely different. Enough of these and users learn; now before buying they frantically search for third-party review videos; without reviews, many won't order at all.

Wu Danni's体会 of this passage was bought with half her life.

In 2024 she squatted in a factory for half a year, watching the production line daily. Her words: "I used to be a pretty gentle person, then I was forcefully turned into a shrew, arguing with the factory every day." At the time the team had no one; after arguing with the factory boss during the day, she had to calmly come back and reply to overseas user emails, often until 3 or 4 AM.

"I used to be a pretty gentle person, then I was forcefully turned into a shrew, arguing with the factory every day."

She also contributed my favorite supply chain insight of the session. Supply chain bosses are all very "jianghu"; her initial mistake was that one boss immediately loved her philosophy and said he really liked the product — her trust instantly maxed out, and later pitfalls all came from this.

"So now when I meet supply chain bosses, I first evaluate from a slightly 'evil' angle, then slowly relax as trust truly builds later."

Host Kang Zhengzhong summarized for everyone: first the小人, then the gentleman.

Doesn't sound nice, but anyone who's done hardware knows this is unvarnished truth bought with tuition.

On Li Rongzhong's side, the delivery keyword is different: unboxing success rate.

He said something I think every hardware founder should copy down. The maker market is comfortable because makers will check docs and solve problems themselves — "we serve the smartest batch of people in the world." But once truly entering delivery, the thinking must completely flip:

"You must assume you're serving the person in the world least able to use this thing."

And you can't call the user stupid. "Users hand you their hard-earned money for your product — how can you call them stupid? When users encounter problems, it's ultimately our own failure to design well."

"Users hand you their hard-earned money for your product — how can you call them stupid? When users encounter problems, it's ultimately our own failure to design well."

So he still personally reads after-sales emails, still remotely helps kids abroad solve problems, still ran into Zuckerberg in the inbox.

Talking here I suddenly understood the opening story. That's not sentimentality; that's their quality inspection system.

Europe Isn't One Market, It's a Dozen

Gu Mingjun gave an example. A domestic entrepreneur made a small washing machine specifically for underwear, naturally thinking this must be global — Chinese users have the habit of washing underwear and socks separately. Result: a colleague just went to the US and saw her husband throw socks and shoes all into one washer, and collapsed on the spot. After ten years in the US, she started doing it herself.

"Behind many demands is actually culture and lifestyle. Sitting in a Shenzhen office, you can't imagine someone else's washing machine."

So what to do? Gu Mingjun's answer: this is exactly crowdfunding's greatest value. Take a prototype to the global market; during crowdfunding users' emails and addresses go to the brand, and you can see exactly which country, which region, what kind of person is willing to pay for your product, and why.

"One of crowdfunding's greatest values is letting companies wanting global markets quickly validate: is globalization actually feasible? If not fully globalized, where exactly is your small piece of the global market?"

Wu Danni validated this with real money. She crowdfunded in 2023, at an awkward time — November, when annual traffic is most expensive. The team at first, like all Chinese companies, poured into big markets like US, Canada, UK, Germany; no matter how they spent, ROI wouldn't work, miserable.

Instead, some countries they'd never considered had surprisingly good data.

"Europe isn't one market; Europe is a dozen countries, a dozen markets."

Li Rongzhong's principle is cruder: economic base determines superstructure. Whichever region pays most, features get adapted to that region first. US users buy most, so first make English docs and interface good. Japanese, French — luckily the product is open source, some users will help do it themselves. A startup's resources are limited; sentiment doesn't rank in priorities.

Three people, three versions, pointing to the same thing: globalization has no uniform template. The market truly yours is often not the first one that pops into your head.

More Conversation Details

Guests

Tang Chang, Founder, Nuoyi Innovation (诺一创新)

Li Rongzhong, Founder, PaiTuoYi (派拓艺)

Wu Danni, Founder, Xingyang Tech (幸氧科技)

Gu Mingjun, China Head, Indiegogo

Host

Kang Zhengzhong, Partner, GreenSeed Global / Shenzhen Head, Unique Research

After the Crowdfunding Hit: How to Truly Reach the World's Shelves

Kang Zhengzhong: I'm today's host Kang Zhengzhong; we mainly help enterprises enter North American offline shelves. This session's theme: after the crowdfunding hit, how to truly reach the world's shelves.

I want to discuss in three stages: product, merchandise, hit product.

The last session discussed more how to be a good product, how to find essential demand, how to make an AI hardware go from "novel at first sight" to "used every day."

This session we move one step further toward commercialization.

Now many AI hardware going through roughly these stages: first crowdfunding, like Kickstarter, Indiegogo; second entering Amazon, independent site, DTC and other online channels; third truly entering North American, European offline retail and supermarket shelves.

So we'll chat along this path.

Li Rongzhong: Hello everyone, I'm Li Rongzhong. I did my undergrad in physics at Nanjing University, then spent 9 years in the US doing a physics PhD and CS master's. Actually I wasn't originally doing robots; after graduation, by a very accidental opportunity, I started doing robots, posted it online, suddenly blew up in tech circles, and started full-time entrepreneurship.

My first bucket of gold actually came from Indiegogo.

Back then I put a little robot cat on Indiegogo, with zero ads, raised $140K, shipped over 1,000 units locally in the US. Later found production is actually hard, so returned to China.

Second campaign also barely any ads, raised about $600K — at that point considered relatively successful.

In recent years we've been iterating small quadruped robots. When we started, there weren't many such products on the market. We're probably one of the earlier companies to validate servo motors as a high-performance quadruped robot solution; in 2018 we also open-sourced the code.

Recently we'll be launching a new product, right here in my pocket, can show everyone later.

Wu Danni: Hello everyone, I'm Wu Danni, founder of Xingyang Tech; our brand mainly does scalp care direction, closer to home beauty and personal care medical.

Coincidentally, our first bucket of gold also came from Indiegogo.

But our category isn't particularly hot in crowdfunding, relatively niche. But at the time our Indiegogo crowdfunding performance was decent.

Now overseas is mainly our own independent site; we're not focusing on platform e-commerce for now, choosing some precise professional channels.

Gu Mingjun: Hello everyone, I'm Gu Mingjun, Indiegogo China head; everyone calls me Mingjun.

Just hearing the two of you, you're actually both quite familiar with Indiegogo; we've exchanged many times before.

Indiegogo, including other crowdfunding platforms, now has very deep ties with the entire DTC going-global. Indiegogo is also one of the world's larger smart hardware crowdfunding platforms; my side mainly supports Chinese brands in China, through crowdfunding to complete the first step of DTC going-global.

I think taking this opportunity, I can clarify crowdfunding a bit.

Why can crowdfunding have deep ties with DTC?

Because crowdfunding's platform mechanism is very special.

If you want to do crowdfunding in the future, you actually don't need to mass-produce and stock from the start. Generally, when you have a functional prototype, you can already start crowdfunding.

Users pay first, but actually getting the product may take a long time.

Based on this, many companies can very early take a prototype, through real user orders to validate whether the new product actually has demand, whether the market actually has feedback.

Also, because users need to pay in advance and wait, their requirement for crowdfunding product innovation and differentiation is particularly high.

This is also why many new hardware innovations first emerge on crowdfunding platforms.

At the same time, overseas channel partners long watch crowdfunding platforms. If your project amount is relatively high, channel partners will proactively contact you, hoping to be agents and distributors — equivalent to sales paths also having a chance to be directly opened.

Another important point: crowdfunding users and ordinary retail channel users are actually very different.

Take Indiegogo as an example, main users are in North America and Europe, with quite a portion earning $100K+ annually. They have money, have time, and are especially willing to share.

So if brands operate the first batch of crowdfunding users well, they easily become your fans, actively helping you spread the word.

Simply put, a high-quality crowdfunding project can be a very good springboard for brand going-global.

But crowdfunding at most helps you complete 0 to 1.

The later true 1 to 10 still needs independent sites, Amazon, and later retail channels to take over.

Let's Start With Product

Kang Zhengzhong: Let's start with product.

Mr. Li does consumer robots; everyone can see some robot dogs and quadruped robots outside at the booths.

Now AI develops faster and faster, consumer robots themselves are getting smarter.

I'm curious, when you started, why did you choose this product form?

Because in the past people understanding robots might first think of large robots, humanoid robots. Why did you choose this small, programmable robot?

Another question: users buying your product, are they buying the "programmable" capability, or as the product iterates, will it slowly form companionship, even a long-term relationship?

Li Rongzhong: Actually as just said, my entrepreneurship was very accidental.

After graduating, in the principle of "more skills don't pressure the body" — because I originally did theory and software — I started learning hardware programming.

One day, I put sensors on a robot, saw this "eye" could move, and I wondered, could this joint also be made into legs?

Back then robots that could do this kind of agile movement were basically only Boston Dynamics, MIT these big labs or companies, each costing maybe hundreds of thousands to millions of dollars.

I directly cut the cost to two or three hundred dollars.

When the product came out, nobody believed it, so it brought the first wave of huge traffic. That's why I started entrepreneurship.

Later you see it has two little eyes, and you feel it's especially agile.

I personally like pets, kept cats, dogs, turtles, fish since childhood. I've always been interested in "interaction between a human and another existence."

Including I personally like bionics. Sometimes petting a kitten, I can't help feeling its bone texture — these things later gave our product design a lot of inspiration.

I've always felt, between a human and a creation, to truly form an emotional connection, a hard box and a screen aren't enough.

It needs some biological attributes.

And ideally it can move autonomously in 3D space.

Only then do you feel it and you are sharing the same space, the same time, you're spending a period of time together.

So around 2018, I was already quite clear I wanted to make pet robots.

Now almost 10 years later, everyone suddenly discovers this has truly started becoming a trend.

Kang Zhengzhong: So first step, take the original hundred-thousand, million-dollar robot, first make it a price C-end users can accept.

Second is also very related to the founder's interest: you really like pets, really studied human-animal interaction, so you put this kind of thing into the product.

Danni's side is a different direction.

Now beauty tech products are more and more, and from my understanding, different countries' acceptance of beauty tech hardware actually differs a lot.

Europe is relatively conservative, maybe more used to hair tonic-type products; North America, including China, may have higher acceptance of devices.

So I want to ask two questions.

First, why did you originally choose "hardware" this way to do scalp care?

Second, users buying a hair loss prevention or growth product, actually don't care what specs you have or how many patents; what they care most about might be: which step did you actually save me? Did it actually work in the end?

How do you define this question?

Wu Danni: Let me first talk about my background. I'm from marketing, first did medical PR, then participated in several big waves like O2O, VR, always doing Marketing.

Later when preparing to start a business, I actually didn't first think "I want to make a hardware."

The most direct reason was: I myself was losing hair.

And losing it for many years, I'm still a grade-2 hair loss person now.

So these years I've personally used a lot of products.

From a "fellow hair-loss friend" angle, I found a very typical problem: I'd often get brainwashed by market ads, then buy various products.

Come back and don't use them.

Don't use and no effect.

No effect and very depressed.

Depressed and continue buying things.

Always in this vicious cycle.

So I thought, in this chain, is there something we can solve?

Making hair tonic isn't actually my specialty, because I don't have a Medical background. Including now many people tell me, you should make consumables, but this isn't our priority right now.

So if I don't touch hair tonic, can I cut in from the user experience itself?

Because hair care is actually a very chronic process.

I don't know if everyone's used minoxidil or other hair tonics. A hair growth product truly seeing results usually takes 3 to 6 months.

But most users, before 3 to 6 months truly work, already stopped using.

Maybe 90% of people drop off before truly seeing results.

So when we made this brand, what we initially wanted to solve was actually: can we through product design and technology, improve user compliance?

We started from this dimension first.

And when first making the product, we actually did a lot of subtraction.

Now on the market there are many three-in-one products, massage, red light, drug导入, all functions stacked. But we initially cut very many functions, only keeping the most basic experience.

And our product actually isn't sold cheaply.

But I think for people truly with pain points, they're not necessarily that price-sensitive.

What they really care about is: can you actually help them solve a clear problem?

The problem we saw was the original dosing and care experience was very bad, so user compliance was especially low.

We cut in from here first.

Maybe we have one advantage, which is the founder herself is the user, and also the company's product manager.

Kang Zhengzhong: This actually returns to a very important point: the founder herself is the first batch of deepest users, and also the biggest product manager.

Mingjun, I want you to blind everyone from the crowdfunding platform angle.

Now AI develops fast, a product's concept images, videos, seeding pages can all be made very beautiful.

But a beautiful page doesn't mean the market truly needs it.

So if I'm an ordinary user, before backing a crowdfunding project, what signals should I actually look at?

Also, from the platform angle, how do you judge whether a project truly has a new category, or short-term traffic-manufactured heat?

Gu Mingjun: First one.

Just now at opening I mentioned, crowdfunding projects must have very strong innovation and uniqueness for users to be willing to back.

So brands to do crowdfunding well, really rack their brains thinking of various innovations and uniqueness.

But the problem is: what the brand thinks is innovative isn't necessarily what users truly need.

If from the user angle, I think it splits into "inward" and "outward" two parts.

Inward is first ask yourself.

When you see this product, are you excited?

Do you have代入感?

If this product truly got in your hand, what help would it actually bring to your life?

Does its function truly make you feel: "I want this."

That's first.

But one person's judgment definitely has single-sidedness, so second step is outward finding universality.

You can go look at crowdfunding pages, videos — of course these will all be made fancy, because post-production capability is strong now.

So I instead suggest everyone go look at the comments section.

Crowdfunding project comments sections are usually very lively.

Don't just look at templated comments, find real user questions, see how the brand answers.

Like storage products, real users ask: I live in a 30+ below zero area, how long does it really last at night?

If many users keep asking around similar real usage questions, it actually shows the product's underlying demand is real.

This is from the ordinary user angle.

Second question, you ask how we judge whether a project is traffic product or truly new category.

I think before crowdfunding ends, it's actually very hard to judge.

Because today's crowdfunding and early crowdfunding are already very different.

Early crowdfunding really might rely on organic traffic and luck, didn't need lots of marketing to make data look good.

But now crowdfunding and brand connection are deeper and deeper; many Chinese giants, including Anker, Ugreen these companies are all doing crowdfunding, everyone will find ways to do traffic.

So during crowdfunding, marketing will whitewash lots of data.

To truly judge, I think you should look at two things later.

First: after crowdfunding ends, is there long-tail sales growth.

Just said, crowdfunding solves 0 to 1.

An ideal crowdfunding project, after crowdfunding ends, should already have a brand, have tone, have first batch of user base.

Then the later independent site, Amazon, in theory should keep growing.

If crowdfunding is the peak, later sales get lower and lower, then it's worth being wary.

Second look: how long community vitality truly lasts.

During crowdfunding everyone is definitely active, because there are various project progress, marketing activities.

But what's truly worth looking at is after crowdfunding ends.

Like how's delivery, how are new orders, after new users come in do people still discuss.

If your Facebook community, Reddit community, after project ends still long-term stays active, people keep exchanging, it actually shows it's not quite a pure traffic product.

If crowdfunding end is the highest point, later nothing, I think to some extent it might be a traffic product.

Conversely, if crowdfunding is just a starting point, later sales, channels, user spread all keep growing, it's more like the start of a truly new category.

Like Plaud.

Plaud in our view is a relatively early AI + hardware success case, also started through crowdfunding.

At the time Indiegogo and Kickstarter two platforms combined about $2M+.

But crowdfunding was just its start.

Later everyone looks again at Plaud's sales on Amazon, TikTok, independent sites — growth is very strong, Amazon market share is also very high, user discussion heat has never come down.

And after Plaud, a batch of competitors and友商 appeared, starting to make similar products in various scenarios.

So actually, what Plaud validated for everyone is a category.

At this point you look back and find: its crowdfunding amount isn't the most important; what's important is whether the whole market keeps growing after crowdfunding.

Kang Zhengzhong: So crowdfunding amount is of course one metric, but truly judging whether a project is good, you still have to look later at scaled delivery and channel progress.

Tang Chang: I'm Tang Chang, founder of Nuoyi Innovation.

We're now making a voice intelligent interaction AI ring.

We define it as a "minimum cross-section" of human-AI interaction.

We're built on a large model, mainly solving the thoughts, feelings, Todos that users have at any moment within 3 to 10 seconds.

It's actually quite different from Plaud or meeting notes-type devices.

Our main usage scenarios are during commuting, after meetings, and some informal communications.

Many times you suddenly have a fleeting piece of information, but didn't have time to record.

We want to make a very convenient, seamless, while not interrupting your original workflow device.

Now Spark Ring already has a warm-up page on Kickstarter, and we've built our own community on Facebook.

First Version Delivery: Which Experience to Defend First?

Kang Zhengzhong: Just now we discussed how to judge whether a crowdfunding project succeeds, and one very important point is delivery.

Mr. Tang you're officially crowdfunding next month, so let's follow this topic.

As ordinary users backing crowdfunding, often encounter a big pit: many features and experiences written on the page, between the truly received product there's a very big gap.

If clearly telling users this is a second-gen product, features will upgrade later, everyone might understand.

But some products put many future capabilities on the page, and when truly delivered it's completely different, then users will be very confused.

So I want to ask, when Spark Ring truly delivers for the first time, what experience do you most need to defend first?

Which capabilities must be implemented in v1?

Which capabilities can be clearly told to users are done through later upgrades?

Tang Chang: This question is very good.

Recently actually launched many AI rings, including various AI hardware.

A relatively common characteristic is: spoken about very well, but delivery may not be that good, or not that fast.

What's lost in the end is user trust.

So we take this question very seriously.

If you're in Shenzhen wanting to long-term build a brand, definitely put users first, put interaction experience first, while balancing AI experience.

Our product is a voice ring. In the past there actually haven't been many voice rings truly scaled and delivered.

So we first have to ask ourselves: what problem are we actually solving?

Before voice rings, how did people record inspiration and fragment info?

According to our research, 40%+ people use Apple Notes or phone built-in memo, another 20-30% message themselves, or use pen and paper.

These are the old experiences we first need to replace.

So the first thing is very simple: it must record.

This requires microphone, ASR, the entire basic chain from voice capture to recording must be done well.

Second is AI.

Everyone now has very high expectations for AI, but we looked at many Shenzhen AI hardware, the true AI experience actually hasn't been delivered that well.

So in the AI part, we instead won't overhype.

First run through the most basic process of user-AI interaction, do it well.

Including in pre-test and warm-up stages, if users in comments have expectations for the product, we reply one-on-one.

Some unrealistic expectations, actively pull back.

Which ones we can truly do, explain the experience clearly.

I think this is building a long-term brand trust very important thing.

We won't, in order to make a particularly beautiful crowdfunding number, make the product sound particularly exaggerated, then not care about delivery.

We still treat it as a long-term thing.

So won't blindly chase numbers; users and delivery are definitely first.

Kang Zhengzhong: First have long-termism, then defend the first version product's truly deliverable boundaries.

Mr. Li, you also said before you did two crowdfundings.

Open-source robots are actually a very special product.

In Maker, maker communities it might be relatively easier, because everyone can accept products constantly updating, solving problems themselves.

But truly turning it into a standard consumer product, I think is actually very hard.

So after your crowdfunding ends, what's the first experience you must defend?

Li Rongzhong: If to say the simplest, it's a bottom line: must guarantee unboxing success rate.

If we keep doing the maker market, it's actually quite comfortable.

Because makers solve problems themselves, check docs themselves.

I also used to sigh, when the product video went out and suddenly huge traffic came, we served the smartest batch of people in the world.

They know what the current technical level is, we suddenly being able to do it like this, they find it stunning, and start forwarding for us.

But once truly starting delivery, the whole technical team and product team's thinking must completely change.

You originally served the smartest batch in the world.

Now you must assume you're serving the person in the world least able to use this thing.

Of course you can't call users stupid.

Users hand you their hard-earned money for your product, how can you call them stupid?

So if users encounter problems during use, ultimately it's our own failure to design well.

This mindset must be low enough.

And even if we sell a hardware robot, I increasingly felt later that what we ultimately deliver to users is actually a service, an experience.

From payment, to logistics timeliness, to unboxing after receiving the product, to reading the manual, to how to solve problems after encountering them.

The whole chain must connect.

In the end users feel happy, willing to give good reviews; or minimum requirement, at least not return it.

Only then does the value our team creates for users truly form a closed loop.

Including myself, to this day I still watch all Support Emails.

After-sales emails I all read, and about a third I personally reply to.

Sometimes users need remote support, I simply don't care who they are.

Maybe just a kid, as long as they need English remote communication, I directly jump in, help them solve remotely.

In this process you receive lots of very direct, very sincere user feedback, plus spiritual support.

You suddenly discover, this isn't some capital game, nor some partner's order.

It truly solved the user's problem.

And you'll also encounter some意外 surprises.

Our product is after all a high-tech product, so quite a few tech circle people buy.

When I did Support before, I even saw Mark Zuckerberg use his personal email to buy our product for his daughter.

This kind of thing you really only encounter when you personally serve users on the front lines.

Kang Zhengzhong: So finally return to the same point: the founder is the biggest product manager, must truly see every user's feedback.

Danni, you did crowdfunding around 2023, also truly delivered products.

And your category isn't the hottest tech crowdfunding category, but in your vertical market results are still good.

Now Europe and North America are also advancing.

So from crowdfunding "hit orders," to truly starting to form later repeat purchases and long-term sales, what hurdles did you cross?

Also, from day one you wanted to build a global brand. In early brand days, did you first prove the product's immediate experience, or long-term efficacy?

Wu Danni: Let me answer the later question first.

Just now I kept saying, why when first making the product did we do a lot of subtraction.

Because before entrepreneurship, I actually sold hair growth devices for a while.

The first huge challenge I encountered in this category was: user trust.

Because people coming to buy this kind of product, including myself, essentially hope to get a good result; everyone comes with expectation.

But as a brand new brand, to convince them, you need lots of clinical, lots of Clinical Trial to validate efficacy.

This is the first difficulty.

The second difficulty is compliance.

Before we spent a lot of energy building WeChat groups, various groups, telling users this product is good, come use, come check in.

But in the end there's still a lot of users who drop off.

So when I remade this product, I started redefining this thing.

First, at this current stage, we don't first make efficacy-type promises.

We first make it a tool.

Let user experience be good enough, compliance high enough.

I think this thing has actually already gotten some validation.

We are a very atypical startup.

I most really started alone.

From entrepreneurship, to first batch goods delivery, to second version product iteration, the whole time was very long.

For me, the first truly big hurdle entrepreneurship hit was actually supply chain and production.

In 2024 I squatted in a factory for about half a year.

Daily watched the factory, watched the production line.

I used to be a pretty gentle person, then was forcefully turned into a shrew, arguing with the factory every day.

This was the very first very big tuition I paid since starting a business: over-optimistic about production, over-optimistic about supply chain.

So if now let me redo the next product, let me do crowdfunding again, I might wait until the product has already opened samples, the overall form basically settled, before starting crowdfunding.

This way after crowdfunding ends, the later traffic you can truly take.

At the time my team had no one.

Every day after arguing with the factory boss, come back and have to calmly, email by email reply to overseas users, often until 3 or 4 AM.

At that time pressure was actually huge.

I'd think: "I've worked this hard, why do you keep urging me for goods?"

But now looking back, this experience gave me two abilities.

One is judgment on supply chain.

The other is, I really had a lot of direct contact with users, collected a lot of very real feedback.

This year after our product officially launched, we started seriously doing independent sites, then slowly moving toward channels.

For globalization, we actually never defined ourselves as a "Chinese brand" or "overseas brand."

From Day One we thought of the global market.

We're now at Shenzhen Science and Innovation Academy; Li Zexiang has a line I particularly agree with: don't obsess over whether you're a Chinese brand or overseas brand, don't obsess over whether to do China market or overseas market.

What you truly want to do is a global market, just where your starting point is.

Our starting point is now more overseas, lots of users in the US and Europe.

So we particularly value building connections with overseas users.

Including we now value customer service a lot.

Why?

Because customer service is the brand's first layer of connection with users.

But doing overseas customer service domestically is actually quite hard.

First is customer service's own depth of understanding of business and users.

Second is language barrier.

Now of course there's AI.

But to truly use AI to the point the other party feels "you're a local," there's actually still a threshold.

We always emphasize one word: Go Global, Go Local.

We'll constantly grind scripts with customer service teammates.

Another is, the founder must truly go out.

I'm also going to IFA soon, then spend a month in Europe, meeting various partners.

Because we're now going through professional channels, so truly building relationships with professional channel people is especially important.

I increasingly feel, domestic and overseas are actually the same, essentially relationships between people.

Before I looked for various doctors in Ireland on LinkedIn, they basically ignored me.

Even if they reply, after one phone call say: "Oh, very cool."

Then nothing.

But when I went to Paris in July, truly meeting these people, it was completely different.

They see the product, truly know you as a person, then build trust with the brand.

Come back and call again, closing is much faster.

So if you truly want to do a global brand, I'd very much suggest founders and core teams, from Day One, interact as much as possible with people from target countries.

Even spending a month, two months there building relationships, I think is very worth it.

So returning to "world shelves."

We now mainly do a few things:

First, who the user is must be clear.

Second, channels must be very clear.

Like certain mass channels aren't suitable for us at this stage, we won't touch them.

Third, go to local more, truly meet real people.

Supply Chain: First the小人, Then the Gentleman

Kang Zhengzhong: Just now actually mentioned again a problem all hardware founders can't avoid: supply chain.

Everyone's backgrounds may be completely different.

Some from marketing, some from product, some from tech.

But once doing hardware, you eventually hit supply chain.

Luckily today we're in Shenzhen, Greater Bay Area supply chain is very mature.

But "supply chain is mature" and "you can find supply chain truly suited to you" are actually completely two things.

Wu Danni: I think supply chain's first class to me was very interesting.

Our founding team people might have all had decent work experience before, in companies average cognitive gap isn't particularly big.

But supply chain is completely different.

Supply chain bosses are all very "jianghu."

I actually quite like them now, because I've been learning from them.

The first mistake I made was, meeting a supply chain boss, he immediately loved my philosophy, said he really liked the product.

Instantly my trust in him maxed out.

Then many pitfalls I later stepped on were because of this.

So now when I meet supply chain bosses, I first evaluate from a slightly "evil" angle, then slowly relax as trust truly builds later.

Kang Zhengzhong: First the小人, then the gentleman.

Gu Mingjun: From the platform angle, you've seen too many crowdfunding projects.

After crowdfunding ends entering the delivery stage, what do you think brands most need to watch?

Delivery is truly the most headache part of crowdfunding, and also the part everyone most easily ignores.

Many people only focus on the amount on the page.

Today hit $1M, think crowdfunding succeeded, start celebration banquet.

Actually not.

The second half of the pitfalls has only just begun.

What crowdfunding hardest really is is delivery.

We've seen too many companies finally stuck on delivery.

Crowdfunding users are very inclusive, and also very viral.

Water can carry the boat, also capsize it.

If after getting this batch of users you give them good delivery, they're satisfied, they become very good spreaders.

Conversely, if you keep delaying, and don't communicate transparently, this batch of users will definitely turn on you.

Many teams indeed ignore this point.

But actually delivery can be pre-warned.

Now crowdfunding is essentially very much like a digital marketing game.

In the entire warm-up period, including marketing period, through data you can already roughly predict the project's final performance.

Now many service providers are very professional, through various models, in the crowdfunding warm-up stage predict roughly how many units can sell, how much spend needed.

So I particularly suggest everyone: before officially launching crowdfunding, first use this estimated sales to negotiate with supply chain.

Pre-run through supply chain and factories.

What's the approximate delivery cycle?

What's the capacity?

What's the cost?

These all must be calculated in advance.

Cost is especially important.

We've seen many brands, crowdfunding amount is very beautiful, marketing ratio is also good, result truly entering delivery, supply chain is completely different from what was negotiated.

Supply chain boss doesn't honor the previous quote, wants to raise price.

Crowdfunding profit itself might not be high, add price and the whole project directly has no profit.

In the end crowdfunding looks "successful," but goods can't be delivered.

So including what Danni just said about prototype state, I particularly agree.

Early years doing crowdfunding, cost was very low.

Even some people's view, one idea, one hand-built prototype, didn't even formally open mold, dared to do crowdfunding.

But now it's already unrealistic.

Why?

Users have been scammed too much.

Now users looking at crowdfunding pages will frantically search whether your product is actually feasible.

Will find third-party review videos.

Without review videos, many people simply won't buy.

Before there were some projects making pages especially exaggerated, like claiming "world's smallest projector, also highest lumens," finally reached $2M+.

But truly delivered was completely different from what was advertised.

After enough of this, everyone feels crowdfunding isn't credible.

So doing crowdfunding today, I'd suggest the product at least already entered mold opening stage, or very close to formal mold opening.

Do it at this time, later pitfalls will be much fewer.

Kang Zhengzhong: Companies spend a lot of time on brand and channels; if finally goods can't keep up, those earlier works basically equal zero.

Gu Mingjun: Really equals zero.

And will also hurt the brand you've accumulated before.

Some companies aren't first-time crowdfunding, already have some brand foundation overseas.

But this time even if crowdfunding amount is particularly high, if delivery fails in the end, previously accumulated users will still overturn your brand.

Because crowdfunding users' spread is especially strong.

You serve them well, they're truly die-hard fans, even actively help you refute people dissing your product.

But if you serve them badly, they'll bring others to curse you too.

Kang Zhengzhong: So the first batch of real paying users is very important.

Because of time, let's finally discuss a globalization question.

Just now in the sharing ahead, a judgment was repeatedly mentioned: globalization is essentially localization.

But no product can naturally fit all markets in the world.

Like Mr. Tang's ring, different countries' people have different wearing habits, aesthetics, usage habits.

So I want each guest to answer:

When doing globalization, which product definitions and capabilities should stay globally uniform?

Which parts should actively localize?

Mr. Tang first.

Tang Chang: This question we actually did user research across different regions.

When crowdfunding pre-launch advertising, everyone doesn't necessarily only spend on final target markets.

Besides Europe and America, maybe Japan, Korea, East Asia, Southeast Asia some small markets will also be tested.

We finally found, everyone's most basic need for a hardware is actually very universal.

Longer battery life, shorter latency, better-looking appearance, better experience, and privacy protection.

These are basically all common.

Also, for language localization, everyone also has requirements.

Like our product in Europe, Germany, France, Italy, Netherlands, Poland these places, customer acquisition cost might be relatively low, but their localization requirements will be very high.

This kind of time needs to do lots of user research, find the most suitable localization strategy for the product.

But returning to the most core place, especially for tech geek users, hardware experience and final delivery are always first.

Supply chain and the product itself must first truly deliver, satisfying consumers' most basic needs for a hardware — these are actually universal worldwide.

Kang Zhengzhong: Mr. Li, robots' definition differences across markets might be bigger.

Some treat it as teaching tool, some as family companion.

When you do globalization, what must be unified, what is localized?

Li Rongzhong: This can be seen from many angles.

Robots themselves are still a relatively niche market.

Ideally of course wherever there's market, we hope to catch it.

But startups also have a big premise: resources are truly very limited.

So our principle is simple: economic base determines superstructure.

Whichever region pays most, when we develop features, we prioritize adapting that region.

Like now our robots, US users buy most.

Then docs, interface, first make English good, then match Chinese.

Basically can already cover a large part of the market.

Like Japanese, French — luckily we're an open-source product, some users even help us do it themselves.

This part can be fully delegated.

Also some regions need localized service.

Like equipment maintenance, local teaching.

This kind we'll as much as possible find local service providers, have them solve.

Basically gradually adapting based on actual market demand.

Kang Zhengzhong: Danni, your category has demand globally, but different countries' understanding of scalp care and beauty tech actually differs a lot.

Especially Europe and North America.

When you enter these markets, do you split the product into completely different models?

Wu Danni: Our product is a bit special.

First, hair loss is a very Universal problem.

Basically as long as you're human, you might have this problem.

Domestically about 1 in 5 people, Middle East might be more severe, about 1 in 4.

So if looking at overall human demand, it's actually relatively unified.

But if further splitting demographics, some differences may appear.

Like our current product, isn't particularly suited to Black users.

But Black hair loss is actually very severe.

In the future we might specifically develop some products to serve this group.

From current hardware features, because our demand is relatively simple, function is relatively single, so for now haven't done too many differences for different countries.

The first difference we might do in the future is on the market side.

Like now the product is relatively neutral gray, silver; for certain markets, we might do different CMF.

Second, if future add more intelligent capabilities, local compliance becomes very important.

One is data security.

Another is, local users' actual demand for certain features is high or not.

Like US users' acceptance of new products, new tech is overall relatively high.

But Europe overall, even acceptance of many beauty tech is lower than the US.

So "intelligence" this thing, for some European users might not be that attractive.

Of course, also related to channels.

At least in our next two or three products, I don't think we'll do too much customization for each country.

Because cost is too high.

Finally it's still ROI.

We might more first based on a good product, go find the market most suited to it.

Kang Zhengzhong: This is also a problem hardware companies will definitely face.

After too much customization, hard to form standardization, long-term costs get higher and higher.

Wu Danni: Right. CMF is relatively OK. But if truly opening a new set of molds for one market, immediately 100-200K RMB, later there's maintenance, material management, very complex.

Kang Zhengzhong: Mingjun last question.

From end of 2018 entering Indiegogo, to now you've seen very many crowdfunding projects.

Just now you also repeatedly said, crowdfunding is only 0 to 1.

So enterprises truly doing 1 to N globalization, especially soon IFA, global media, partners, channel buyers all gather.

For brands that have completed crowdfunding and are preparing to enter global markets, what signal do you think they should bring to IFA this kind of global trade show for validation?

Gu Mingjun: I still want to break it down from the crowdfunding angle.

Because we've been in touch with especially many companies.

In the crowdfunding track, many are startup teams.

The biggest problem for startup teams is: too much uncertainty.

So what they want to validate through crowdfunding is actually several layers:

Can my product be globalized?

Can my this feature be globalized?

Can my users be globally replicated?

Crowdfunding is especially suited to doing this.

Because crowdfunding essentially lets you take a product that's not yet fully mature but can already show core capability, relatively early do global validation.

Currently the whole crowdfunding market is still North America-led, Europe-secondary.

At the same time Japan, Korea, Singapore these mainstream developed markets are also covered.

So you can through these users' most real feedback, judge whether your feature can truly be globalized.

This is very important.

Can't just because the founder himself thinks "my product can be globalized," it really can.

Before we chatted about many domestic small appliances.

Like someone making a small washing machine specifically for underwear.

He'll naturally feel this must be globalizable.

Because Chinese users have the habit of washing underwear and socks separately.

But North America might not be like that at all.

We had a colleague who just went to the US, saw her husband washing socks, shoes, all together, she felt completely broken.

But after 10 years in the US, she also started doing it herself.

So behind many demands is actually culture and lifestyle.

At this point don't immediately mass open molds and produce.

First take a prototype to test, see if different markets actually have demand.

If really OK, then go toward globalization.

Also, globalization also has to "look at the person."

What kind of person is actually buying your product?

Which countries' people are interested in your product?

Crowdfunding has a big charm, which is you can take a prototype to the global market, then get a batch of real user data and profiles.

During crowdfunding, user information all goes to the brand.

Email, address, you can all have deep exchanges with them.

You can see exactly what kind of person is willing to buy.

Which country is he in?

Which region?

Why is he willing to buy?

With this batch of profiles, you can then replicate your globalization.

So I think one of crowdfunding's greatest values is letting companies wanting global markets quickly validate:

Is globalization actually feasible?

If not fully globalized, where exactly is your small piece of the global market?

Wu Danni: Let me add to Mingjun's point just now.

When we first did crowdfunding, actually chatted with Mingjun a lot.

I knew from the start, our project definitely wouldn't be a Million-Dollar Project.

But we still very much needed crowdfunding.

First reason is, whether crowdfunding succeeds or not, we'll make this product anyway.

Second reason is very practical: when just starting a business, there was really no money on hand.

So we could take crowdfunding orders to go to suppliers to make goods, then take the money received from crowdfunding to open molds.

This was actually very helpful in early entrepreneurship.

Another point, today wanting to completely rely on organic traffic to make crowdfunding work is already very, very hard.

Most projects still need digital marketing and advertising to validate.

But when we launched, the timing was particularly awkward — November, exactly when traffic is most expensive all year.

At first everyone, like all Chinese companies, poured into North America and Europe's big countries.

US, Canada, UK, Germany, France, Spain, these most key markets.

Result because traffic was especially expensive, no matter how we spent, ROI wouldn't work out, miserable.

Later we instead found, some other countries' data was particularly good.

This thing actually, in the crowdfunding stage, completely opened our understanding of the global market.

Europe actually isn't "one market."

Europe is a dozen countries, a dozen markets.

So when doing globalization, your vision can be a bit broader.

Don't naturally think I only do the US because US spending power is high; or just stare at Canada, UK these traditional markets.

Including Middle East, I now really feel the potential is very big.

Crowdfunding to some extent, is helping you find these markets you originally wouldn't have thought of.

Kang Zhengzhong: So doing globalization, there's no uniform template.

Sometimes the market truly worth going to isn't the one you first think of sitting in a Shenzhen office.

Crowdfunding isn't just selling the first batch of goods; more importantly it lets you see in advance: who is willing to pay for the product, why they buy, and where to go next.

---

Original publication: https://uniqueresearch.substack.com/p/src-20260902-01html
On-site reading page: https://ffcap.cn/en/research/src-20260902-01html