Exclusive Interview by 36Kr｜ Quan Zheng from ZBeetle Intelligence: We choose to take on really tough problems amid involution in the cleaning appliance industry

Written by/Ya Gao   Edited by/Xiaoqiu Peng

ECOVACS’s soaring stock price from RMB 17 to RMB 252 within one year shocked all investors in the primary market, and they wanted to find the next ten- bagger in the sweeping robot industry.

On this background, almost all sweeping robot enterprises finished financing in August or September last year have seen their valuation increase by 3-5 times in December. In the first 8 months of this year, there were 79 financing events in the service robot field, with a total amount disclosed reaching RMB 23.8 billion.

Star projects had a staggering rise. In 2020, NARWAL completed the A+ round investment led by ByteDance, B-round investment of nearly RMB one hundred million led by Source Code Capital, and C-round investment led by Sequoia Capital China in succession. Its evaluation has seen a 15-fold increase from RMB 400 million to USD 1 billion within half a year.

With the advent of industry inflection point, this growth has its grounds.Young consumers became the main force of the market, and boosted the purchasing power for sweeping robots.

According to public data, the retail scale of sweeping robots in the first half of 2021 was RMB 5.3 billion, with a year-on-year growth of 39%; the expected price among consumers also increased. The average price of products hit RMB 1,914 in Q1 of this year, up by RMB 384 compared with the same period last year.

Fierce competition comes after a capital boom. When low-hanging fruits have been picked up, you need to do more to reach fruits in higher places.

In the sweeping robot industry, more fruits are growing high and are also difficult to pick up.

To better complete cleaning tasks, today’s sweeping robots have “eyes” and “brain”, carrying a series of technologies such as LDS laser radar, 3D structured light, dToF, AI algorithm which were once applied in aerospace, autopilot and smart phones.

Although the technology seems cool and market space is attractive, the market penetration rate of sweeping robots in urban households in China was less than 8%, according to Special Report on Machinery Industry - Sweeping Robot Industry by CITIC Securities. For this reason, entrepreneurs continue to enter this sector.

In January of this year, Quan Zheng who once worked in Microsoft and Aqara, also marched into this industry and created the household service robot company called ZBeetle Intelligence.

“Involution” is a word used to describe the state of the industry by Quan Zheng.“After NARWAL rolled out dual-roller mopping robot, some manufacturers started to make dual-roller sweeper. This is a kind of involution but not innovation.

Have sweeping robots reached the bottleneck period of technical development? What kind of sweeping robot is needed by consumers? Is sweeping robot one of the starting points for smart house?

From this interview between 36Kr and Quan Zheng, we learned that he had 16 years of experience in internet & smart hardware products, and once created a Top 1 popular intelligent door lock from scratch and this single SKU’s

shipment has hit 1 million sets. He had unique insights and opinions on smart house portrait, vision and technical trend of sweeping robots.

What is the starting point for smart house?

36Kr: Why did you choose to manufacture household service robots at this moment?

Quan Zheng: From the original intention, we wanted to make household service robots.

We started this idea from 2016. At that time, I was working in Microsoft. It has deployed robot arms and vision algorithm at a large scale in factory. The entire factory was in order under management of robot arms and big data algorithm. We thought that these devices were cool and conceived whether we can introduce these automatic

devices in our home to create intelligent household scenes.

Later, I went to Aqara and made the first generation of Xiaomi intelligent door lock. At that time, I came into contact with smart home scene and created a popular intelligent door lock. Until last year, I felt that it was time to open up a new battlefield when intelligent door lock businesses were relatively stable.

The idea was simple at first. For example, could we use robot arms on a sweeper body and make it pick up garbage, and so on.

Later, we found that there was a long way to go for producing a real product of this kind, and cannot bring true values to users. So, we decided to start from sweeping robots and shrink the big dream of household service robots to cleaning robots.

36Kr: In recent years, people often thought that smart speaker was the starting terminal for smart house. However, can it be sweeping robot?

Quan Zheng: In reality, this is the difference between dynamic and static. Currently, we don’t think that there is a starting point for smart house. We can only say that what kind of scene might be a good point to start. So far,

sweeping robot is one to realize man-machine coexistence environments.

36Kr: Compared with other smart household devices, is sweeping robot the form that is mostly likely to be realized?

Quan Zheng: If we started to make household service robots from the beginning, they can almost do nothing at home, because current family devices including house devices and home appliances are designed from the perspective of human interaction. If robots can provide true services at home, the entire family needs digital transformation.

An ergonomic environment can be built only after digital transformation. Then, robots can be used in thousands of homes. This is a necessary process.

36Kr: Talking about smart house, although it brings convenience to consumers,some consumers reflected that their usage is far from an interconnected status in terms of feeling. What do you think of smart house of this kind?

Quan Zheng: Existing smart house is in an extremely primary stage, which only achieves backstage data connection without automatic physical interaction between devices.

What is direct physical interaction? For example, can a cup, desk or stool make some physical movement?

Supposing that we install some chips on a desk for automatic lifting, who will control it? This is a question worth considering. This is the endgame of smart family.

In today’s smart family, we realized automatic turning on of light, automatic curtain opening, and so on. However, this is far from the ultimate form.

I remembered that Samsung launched a robot with a mechanical arm last year, but the root problem remains unsolved: Currently, the living space of most ordinary families has not become digitalized. Then, robots can hardly adapt to user families and can do limited things at home.

36Kr: In the future, is it possible that we can operate a smart house without using a phone app?

Quan Zheng: If it is in the ultimate form, the technology is not perceivable.We hope that technology is a necessity like air, but you will be not interrupted by it. For example, the technology can be convenient and integrated into the entire home environment without causing any effect from visual perspective.

Simply speaking, it needs little effort to maintain. An ideal form scene is like this: one day when you enter your door, the home layout is what you desire.

For instance, when you want to workout, everything, such as the music playing,the light, the items placed on a desk, is the way you want.

As to sweeping robots, if we do not need to take efforts to maintain it, it

means that its sense of existence is reduced.

For instance, sweeping robots are prone to collide with wires when moving. We can prevent them from being tangled by wires or stuck somewhere through obstacle avoidance. For another instance, mop can be automatically cleaned instead of being washed manually This can also reduce sense of existence.

The first purpose is to achieve true cleanness; and the second purpose is to achieve true convenience.

36Kr: As we observed, the market penetration rate of sweeping robots is relatively low. To improve it, what should we do?

Quan Zheng: The penetration rate is indeed low, and there remains a huge market space. We can compare this from two measures of data. First, licensed houses across China have hit 400 million. Second, the annual market scale of washing machines is RMB 70 billion to RMB 100 billion. In contrast, the annual sales amount of sweeping robots in China has not reached RMB 15 billion. Therefore, there is a huge space.

We need to improve penetration rate from two dimensions.

First, for cleaning products, cleanness is one of the most important factors for consumers to consider purchasing. From cleaning dimension, most sweepers on the market cannot do a better job than manual sweeping, except for handheld floor scrubber, which cannot truly free up hands. This concerns the second dimension: convenience.

High annual sales and penetration rate of washing machine contribute to the fact that it is cleaner and more convenient to wash clothes than human labor.  In comparison,

sweepers are far from replacing human labor in terms of cleanness, which results in a relatively low penetration rate.

For these two reasons, a gap exists between products and ultimate form desired by users, which explains why the penetration rate cannot increase significantly in the short run.

36Kr: We noticed that the price of current handheld floor scrubbers can reach RMB 3,000 to RMB 4,000. Why are consumers willing to purchase one at such as high price?

Quan Zheng: Because this is a rigid demand. When looking back the development history of washing machines, we found that washing machines also faced with the same development bottleneck as sweepers at the time of invention in the beginning of 19th century. With the development of era and technology,household per capita incomes increase and people have higher requirements for living quality.

Washing clothes and sweeping floor are common work scenes in house keeping.High-quality home appliances help save time, improve living quality and create more values. Handheld floor scrubbers can do a better job than manual mopping in terms of cleanness. This is an accurate user pain point that most sweeping robots cannot realize.

XWOW robot combines and improves superior cleaning performance of handheld floor scrubber on the basis of sweeper. It outperforms manual cleaning and does not require maintenance within 3-4 days.

36Kr: If the cost is relatively not high, can we expand customers and get a bigger market by reducing profits and lowering prices?

Quan Zheng: Currently, handheld floor scrubber market has already competed on price with different brands of products on sold. In this case, market price is not determined by a single company, but from the choice of market regulation. The question is why do different brands choose to make this product? Because this solves consumer

pain point. This is also our original intention when we started our business, namely, doing tough but worthwhile things.

Do new brands have an opportunity?

36Kr: We know that algorithm is reformed by accumulation of scenes and data. Some enterprises that have long engaged in sweeping robot have relatively mature algorithm. Do you think new brands have a cutting edge at this moment?

Quan Zheng: In terms of engineering practice, some incumbents indeed have an early edge. For example, they have a large customer base and can optimize products with massive user data.

These mature experiences bring industry learning and exchange opportunity. Practically speaking, technology is evolving, and many new sensor technologies have narrowed the gap in algorithm.

From technical level without obvious difference, if current mature technology can be transformed to products with new forms, it will be a good opportunity for new brands, for example, when breakthroughs can be made in usage process,usage effect, and product design.

We know the fact this market has high demands and low penetration rate. To make products useful is the key. Considering these factors, a brand that can offer good products will never lose consumers.

36Kr: From current market occupation rate, top five brands have already occupied a huge portion, and consumers generally tend to purchase familiar brands. What should a new brand do to shape its brand?

Quan Zheng: Let’s use a four quadrant to measure a cleaning appliance. The vertical axis represents convenience, and the horizontal axis represents cleanliness. We found that in this quadrant, most products launched in recent two years have not exhibited good performance in convenience and cleanness.

If we further segment these two pain points and seek for a higher goal in the first quadrant, we have the opportunity to impress consumers.

Crack a hard nut, and do something tough but worthwhile

36Kr: Regarding sweeping robot innovation, there is a view that “integrated sweeping and mopping” is a mode innovation, and moat is algorithm innovation. Do you endorse this view? What is the moat for XWOW?

Quan Zheng: Let’s talk from two aspects. We have innovation in algorithm and structure. For example, to refill clean water and drain sewage, we need to achieve five-location alignment when the robot returns to the station for

charge. In ordinary sweeping robots, only two charging sheets are used for alignment.

This poses huge challenges on algorithm, sensor, and structure design. How to make a larger dust bin, clean water box, and sewage box, while keeping the entire sweeping robot size unchanged? We have made a lot of structural and material innovations and multiple technical iterations when manufacturing products.

This era does not have the so-called moat. In order not to be abandoned by the times and users, we need to always focus on offering practical values to users and providing innovative products with better experience.

For example, we keep considering how to provide truly clean user experience, and finally integrate such functions as mopping, mop cleaning, and real-time collection of sewage and dirt on the floor in XWOW.

36Kr: Why cannot the combination of cleanness and convenience be achieved by previous manufacturers?

Quan Zheng: Some manufacturers in the industry have been trying to combine cleanness and convenience. For instance, many new machines launched recently require no manual cleaning of mop. In addition, dual-roller mopping mode exhibits much better performance than previous attachment of one piece of mop on the station bottom shell. These are attempts and efforts made in cleaning dimension.

However, how to do better on this basis is a goal that all ZBeetle Intelligence staff strive to achieve. I believe that this is also a direction of continuous efforts in the entire home cleaning industry.

I hope that with ZBeetle Intelligence’s entry in this industry, we can deliver high-quality product for consumers, bring a new communication window for the entire industry, and together create more possibilities for the industry.

36Kr: How can a robot clean every corner at home with better performance than manual cleaning at home?

Quan Zheng: This requires combination of structural design, sensor and algorithm. We use LDS global navigation with three lasers. Ordinary products only use single-point laser or infrared sensor to make a robot move along the

wall path. Therefore, our products can move along walls or approach obstacles closer.

36Kr: From the consumer's point of view, what kind of products does sweeping robot belong to considering its technical properties according to your opinions?

Quan Zheng: Compared with high-tech products, I think sweeping robots are more like home appliances. We hope to reduce their sense of existence and improve their family property.

36Kr: Will you develop another sweeping robot after XWOW?

Quan Zheng: Yes, we will design our next product into an ultimate product in a real sense. In the future, we can meet different user demands with just one product, unlike now when you need different cleaning appliances.

36Kr: Then, we mentioned something about future smart home, including some ideas on Internet of Things. What will a household service robot look like in the future, in your opinions?

Quan Zheng: Everything we see in science fiction films may come true in the future. When we go home and open the door, smart home has arranged something we want: all rooms are tidy and one cup of hot drink has been prepared when you sit on couch. Nothing is impossible as long as you can image it.