The Success Story behind the Ideal ONE
On December 1st, 2021, Ideal Car announced its delivery results for November: 13,485 vehicles. The milestone of selling over 10,000 units per month was achieved in 24 months, making the Ideal ONE the first new energy vehicle to achieve this feat. Additionally, it became the first Chinese brand vehicle priced over 300,000 RMB to reach the mark.
The Ideal ONE has become a popular choice among families without many realizing how quickly this development has occurred. We interviewed the product development team at Ideal Car to try and uncover the success story behind the Ideal ONE.
The story begins in 2018 – a difficult year for Ideal Car.
At the end of 2017, Ideal Car faced financial difficulties with difficulties in raising funds. In early 2018, their SEV product was stopped due to regulatory issues, causing further damage to the company’s confidence in the capital markets.
Meanwhile, the company started a new round of financing. However, a large institution that was meant to invest was forced to withdraw due to some unforeseen circumstances. Luckily, JWC stepped up to fill the gap and ensure that the round of financing could be completed.
A few months before the launch of the Ideal ONE, vehicle production for the model began, which was an expensive venture. Financial troubles remained a major concern, with spending being akin to a life-or-death situation for the company.
At the Liaison Mansion in Beijing, Fan Haoyu, the head of the Intelligent Cockpit team at Ideal Car, was struggling to solve problems on the four screens in the car. The four screens in the Ideal ONE use two different systems of Linux and Android, and two different computing chips, which presented an unprecedented technical challenge, requiring Fan’s team to put in significantly more resources and manpower than estimated.
At this point, Li Xiang came to Fan with an ambitious plan to include a “whole-car voice recognition” feature.
Fan hesitated, considering that his team was already stretched to their limit. He tried to persuade Li Xiang not to proceed.
Li Xiang refused to compromise: “This is your problem, Haoyu. Remember, we only have one chance to succeed.”
Fan was speechless at hearing this, and Li continued: “Users don’t know about how much resource you put into the four screens. They will just ask why there’s no voice control in the back seat. This is your problem, not theirs.”
“Alright, let’s push ourselves to the limit,” said Fan, regaining his composure and starting to think about the increased workload of this new feature.Li Xiang’s full-vehicle voice recognition system enables four-zone recognition, meaning that sound can be recognized and recorded by microphones located in the front and back seats on both sides of the car. With this system, passengers in the first two rows can easily control the air conditioning, seats, windows, and other functions through voice commands, and can even lock onto a single audio zone to avoid inadvertent activation of the system by children.
Actually, Fan Haoyu had previously considered implementing full-vehicle voice recognition, but it was far more complicated than the original plan.
First of all, changing from a dual-microphone setup to quad-microphone required significant hardware adjustments, and related wiring and engineering projects within the car had to be redone accordingly. The original plan was already close to being frozen, so suppliers would be resistant to making such changes.
Moreover, the industry convention is to use two microphones to form an array for beamforming, which uses the phase difference between sound waves recorded by the two microphones to determine the direction of the audio. This technology is already quite mature.
To implement full-vehicle voice recognition, four microphones would be required, one installed next to each of the four seats. No one in the world has implemented such a scheme before, not even top-tier suppliers like iFlytek or Sibitech, who were reluctant to attempt it.
From two microphones to four microphones, it might seem that the quantity only doubled, but the amount of audio processing power required increased by more than just a factor of two.
After locating the dual microphones, only one set of audio data is actually transmitted, so only one set of audio data needs to be processed. But in a full-vehicle voice recognition setup, all four microphones can transmit audio data at the same time, requiring the audio processing unit to handle all four audio streams in addition to performing echocancellation, which totals to five audio streams.
The system’s increased processing demands were not simply a matter of quantity; furthermore, in order to support the four-zone audio capability, the existing processing power was barely sufficient.
What’s more complex is that the system’s voice recognition model had to be reconstructed, requiring the voice recognition system to be trained from scratch.
But when it comes to decisions that affect user value, ideal solutions are non-negotiable. With no time to spare, the cabin team immediately got involved in this project, developing both hardware and software solutions themselves when no ready-made solutions were available, and leaving the suppliers to implement them later.
In such a situation, the team ultimately succeeded in launching the full-vehicle voice recognition system, making Li Xiang the first car manufacturer in the world to produce and ship cars with four-zone audio capability, and to this day remains the only one. By the end of 2018, Li Xiang’s cash reserves had barely exceeded one billion yuan, and the delay of the full-vehicle voice recognition project had pushed back the car’s overall integration by three months.
And due to the limited four-month time period to perfect the quad-microphone algorithm before delivery, the actual experience still had many shortcomings. Fan Haoyu was not happy with this and hoped that they can improve it as soon as there’s an opportunity to breathe. However, with the continuous optimization of software algorithms in later OTA updates, Li Xiang’s full-vehicle voice recognition system has slowly “turned the tide.”范皓宇 used the term LOD (Levels of Details) to describe the ideal product experience, which is a proprietary term in game engines referring to the rendering of multiple levels of detail. Trees have high resolution and are highly detailed up close, while a lighthouse in the distance may only have a rough outline and be low detail. It gradually becomes clearer as one approaches.
“Things that are completely blurred or perfectly clear are rare, and in the vast majority of cases, things exist in a state of fuzziness and clarity.” Although the lighthouse in the distance and the viewer appear to be in the same piece of land, the river and railing on the way there are not rendered until one gets closer. “You have to gradually walk closer to see the road under your feet clearly.”范皓宇 calls this feeling “fuzzy correctness”.
For the ideal, whether it is a four-screen or full-car voice, it is not just a figment of the imagination or a sudden inspiration, but a path that gradually becomes clear as you approach the lighthouse. The distant hazy lighthouse is called a family car.
The Soul-Searching Question: Who Buys Cars?
Li Xiang’s initial idea was simple, “to build a car for himself and those around him.”
When he founded Ideanomics, Li Xiang had already bought no fewer than ten cars, including sedans, sports cars, SUVs, and MPVs.
After having children, Li Xiang often went to pick them up from school and came across a scene that bothered him: it took too long to get in and out, and children were honked at by the car behind while getting on board. As a result, he began to realize one thing: “Fast entry and exit for three-row seat models is a key requirement for a family car, but there are very few cars on the market designed specifically for this need.”
Li Xiang looked back and realized that the car manufacturers at that time were mainly serving two types of people: drivers in the front row and bosses in the back row. There was not much consideration for the elderly and children. Elderly people had to move their seats manually in the second row, and children sat in the third row like they were in solitary confinement. In Li Xiang’s view, this was all very absurd, and no one had really thought about the feelings of these groups.
Li Xiang’s idea of building a family car grew stronger in his mind.
In May 2016, Zhang Xiao, who had left BMW MINI, joined Ideanomics, where he served as product manager. To be exact, Zhang Xiao was the first whole vehicle product manager at Ideanomics. Prior to that, for over a year, the only person serving as whole vehicle product manager was CEO Li Xiang.
Zhang Xiao saw that Li Xiang’s car-building plan at that time had both extended-range and pure-electric lines, as well as three versions of seats: 5, 6, and 7. Li Xiang told them that the car was to be built for family users. However, many aspects of the car were still unclear at the time.With the approval of the large vehicle project during that period, the product definition of the Ideal ONE project began to converge. Initially, consensus was quickly reached on some basic aspects, such as making it a self-operated business and only offering one configuration because they believed it was the right thing to do.
However, later on, Ideal ONE’s various projects entered the subtraction stage. For non-essential options among the parallel choices, Li Xiang decisively cut them out.
During this period, the pure electric version was halted, and the five-seater version was also cut.
Li Xiang recognized the problems with these decisions: the company’s limited funds and development resources made it difficult to simultaneously accommodate pure electric and extended-range versions, as well as five and six-seat layouts in the engineering aspect.
The pure electric version developed together with the extended-range version could only achieve a range of around 400 kilometers, which was not competitive. By producing a five-seat version with a seven-seat layout, only the third row was cut off. If a truly reasonable spatially allocated five-seat version was to be made, a new engineering plan would have to be developed.
Looking back now, it is likely that doing both would not have resulted in great outcomes. Li Xiang prevented these occurrences as he clearly understood what is suitable for Ideal ONE, which is positioned as a family-use vehicle.
Based on the objective market data at that time, the proportion of five-seater cars was much higher than that of six-seaters. Abandoning the five-seater version at that time was a high-risk move that violated market rules. However, when Li Xiang proposed to cut the five-seater version later on, no one in the company opposed it. This was due to the further clarification of the problem of “who is the user” within the company.
After Li Xiang proposed the concept of family users, he still felt that the definition was not specific enough. After all, families are also divided into those with many people and those with few. For a couple without children, a six-seater car is not a necessity. Li Xiang believes that the best way to define the user is when you describe the user group, 10 people’s mind should imagine the same image.
Thus, the definition of family users continued to narrow down under Li Xiang’s leadership to “families with children.” In the company’s strategy, this group of people is the first target customer as they are the center of the normal distribution of “family users.” Other types of family users belong to the two sides of the curve. They can also still buy Ideal ONE, but they are no longer the primary users.
When the definition and target of family users became clear and aligned within the team, the consensus on subsequent decisions within Ideal became faster, and the unification of the larger direction was completed. Disputes and repetitions were mostly focused on some details.
At that time, Ideal Automobile’s first set of company-level methodology, the “Four-Step Method,” also began to be summarized and integrated. The method included identifying who the user is first, then understanding what the user’s demands are, followed by determining the organizational needs, and finally specifying the goals, and then breaking it down into key outcomes and work plans. When members disagreed with each other, they used this method to discuss problems.It’s now time for the interior review stage. The three-screen concept with the front passenger as the core user appeared in the first interior review stage, but back then the plan was just a rough sketch and the control screen area was still physical air conditioning knobs.
Li Xiang and Fan Haoyu discussed the arrangement of control screen buttons for a period of time. At the time, the two were very busy and didn’t have much time to talk face-to-face, so they communicated through sketching. Fan Haoyu would draw a sketch outlining his ideas for the control screen UI layout, and leave it on Li Xiang’s desk. Li Xiang would then make modifications on the sketch and return it to Fan Haoyu’s desk.
The layout of the first version of the Ideal ONE air conditioning control screen with three buttons on each side was based on Li Xiang’s modified sketch. The choices of four screens, six seats, and extended range were key decisions for the Ideal ONE’s product strategy, and the product team later summarized that these three points laid the foundation for the Ideal ONE’s success.
This kind of decision-making requires product managers to view things from a three-sided perspective. The first side is the user, which requires a clear definition of users and clear user value. The second side is scale, which requires consideration of scale, cost, supply chain relationships, and more. The third side is technology, which requires basic scientific and engineering concepts to be sensitive to the value and application scenarios of technology.
However, product managers, although they need to understand technology, mainly help them establish product definition and strategy. To turn product managers’ ideas from definition into product lies with the research and development team. Therefore, the execution of the research and development team and the cooperation between the two groups largely affects the completion of the product plan.
As a family car, priority was given to the interior space and comfort of the Ideal ONE.
The interior design of the Ideal ONE was once very urgent. The 40.5 kWh battery pack provided the vehicle with a pure electric range of 180 km, but also took up a lot of chassis space. In addition, the constraints of multiple parts such as the three-row seating, electric rails, and extended-range system all affected the interior space. Initially, the feeling inside the car did not meet the expectations of the product team.
The front double-wishbone suspension was also considered at that time. There was no problem with the technical implementation, and the cost was acceptable. However, with the use of double wishbones, the entire space would need to be moved back, making the interior even more cramped. At that time, the depth of the trunk and the space of the second and third row were already the bottom line for everyone, and can’t be made any smaller.
Zhang Xiao was particularly puzzled at that time, wondering why these things were taking up so much space. Clearly, the car was so big, but it still felt crowded. When he asked the engineers, the reply he usually got was, “We need to leave so much redundancy here, we can’t make it any smaller.” However, he was not an engineer, and did not understand the engineering details, he was very perplexed about how to optimize the space and couldn’t get it pushed forward.
Actually, there is no technical problem that cannot be solved. The truth is very simple: the space dimensions of each link are too conservative in design.
Li Xiang’s requirement is simple: Don’t talk about how difficult it is. Talk with data. Data won’t lie.
The overall layout supervisor marks all space dimensions involved inside the car and then prints them into a large chart, which is displayed in the office. All data is clear at a glance. The space that originally relied on feeling can now be viewed on a marked chart, offering a visualized dashboard of interior space.
The dashboard displays not only an ideal car, but also a benchmark comparison car model. Gradually, people can understand what the experience of the actual feedback in three-dimensional space is like.
Li Xiang often looks at the data on the chart when he has nothing to do. Sometimes he spends dozens of minutes thinking about how to create this kind of visualization dashboard in more places.
When it came to execution, there was initially some resistance.
Tang Jing is the ideal research and development operation supervisor. He saw the situation very clearly. He can not only see the rationality of engineering from the drawings, but also understand the engineer’s thoughts.
Tang Jing said everyone’s concerns from the bottom of their hearts: “Maybe this is the commonality among people. Because the first batch of people who made hardware and industrial products came from traditional industries with over 100 years of experience. They have a lot of practical experience. What is their biggest practice? If I’m wrong, I’ll be punished. But if I’m right, I won’t get anything. “
Since the feedback received from practice is that the more one does, the more mistakes they make, and the less one does, the fewer mistakes they make, everyone tends to keep themselves in a very safe area.
“They hope I will never make a mistake. For example, for the interior and the body, it is best to have a 10 mm or even 20 mm distance. We will never touch each other, so there will never be any noise. When everyone thinks and does this way, space is wasted.”
Tang Jing brought everyone together and organized everyone to make transparent all information about the various components. He pulled the engineer out of each individual component’s world, gathered them into one area or even the entire car, creating a common aligned world for all to see.
For every point and every dimension, the engineer had to explain why they do it this way. Is there a successful practice, have they encountered any issues or is it just imagination? Everyone has a 20 mm margin. Is this the product the user wants?
When everyone is on the same chart, the raw data is there, engineers “safe zones” start to slowly shrink and moves a little closer to the limit of actual ability.Compared with the initial design, the second row of floor height has decreased by 25 mm, the second row position has shifted back by 10 mm, the legroom of the third row has been reduced by 20 mm, and the headroom has been reduced by 5 mm. These seem to be small numbers, but each millimeter was carefully considered during the design process. Even the glue and carpet were scrutinized, and the sliding rails of the second row seats were re-molded after hundreds of thousands of dollars were spent trying to perfect them.
During this process, Tang Jing mentioned an important corporate culture at Ideal Auto: “As long as you approach a task with the goal of maximizing customer benefits at Ideal Auto, even if the task is not completed perfectly, we will not punish anyone for it.” This creates a crucial sense of psychological safety for the engineers, who are now more proactive and invested in pursuing the ultimate product experience and engineering innovation. The use of visual data dashboards has also continuously improved and evolved, and eventually, everything related to the seats was linked with detailed data to establish a methodology.
As the engineers became more familiar with the comprehensive data, Li Xiang was able to better communicate his needs and the engineers were better equipped to meet those needs.
In Li Xiang’s view, the engineering team was fully capable of solving these challenges. Many times, it’s not that they lack experience or skills, but their traditional methodology in the automotive industry may be flawed.
For Chinese automakers, management systems and standards often come from mature foreign car companies, which are often adopted without much modification. Few companies have systematically rebuilt their communication and management methods. The process of building an organization’s methods and systems is also a journey from a personal-experience and authoritative-based approach to a scientific and efficient approach, and every iteration is an evolution for the organization and the company.
The biggest challenge for the Ideal ONE’s interior design was overcome with Tang Jing’s initiative to align the team’s data, which allowed for the effective release of interior space. Zhang Xiao was very impressed with Tang Jing’s engineering ability and problem-solving methods during this incident.
In her eyes, Tang Jing is proactive and mission-driven. Li Xiang also highly praised Tang Jing’s work: “We have a complete and rigorous process, and Tang Jing’s responsibility is to uphold these processes. But the difference is that he is not just upholding them for the sake of it. He believes that everything that is valuable to the customer should be improved. This does not conflict with the process, and he processes these problems.”
Tang Jing understands that his current mentality and values were not formed immediately upon joining the company. He was also a “traditional” automotive engineer before joining Ideal.Before joining Li Auto, Tang Jing was already an industry veteran in the automotive industry. At that time, he didn’t think much about the position of product manager. Years of experience in traditional automotive manufacturing told him that the product manager was a retirement position, usually where the boss’s secretary went after retiring.
And those so-called product managers didn’t really understand the product or the car. In Tang Jing’s eyes, those “product managers” were just configuration specialists.
“We call them configuration commissioners. Some product managers don’t even know the difference between automatic rain wipers and speed-synchronized rain wipers,” Tang Jing said. This used to be very common a few years ago.
At the end of April 2018, Tang Jing formally joined Li Auto, and he felt that many of the things here were different from before.
Tang Jing and his colleagues found that the product manager team headed by Li Xiang was always proposing various modifications. Many of these were unnecessary or self-inflicted problems from the point of view of the engineers.
One project change stands out in Tang Jing’s memory. During the engineering sample car phase of the Li ONE, Li Xiang proposed changing the second-row seats to electrically adjustable. The reason was that he believed that in the family car scenario, the elderly, children, and women often sat in the second row, and manual seat adjustment was inconvenient. Especially for seats with child safety seats installed, manual adjustment was extremely difficult.
“In the traditional automotive industry, project leaders usually only need to plan the schedule. Their top mission is to deliver the product on time,” Tang Jing said. This is an objective and widespread phenomenon in the industry.
Therefore, once something that increases project time or uncertainty, such as changes or innovations, is involved, project leaders mostly hold opposing views. Because delivering on time is their mission, whether the car is competitive or how well it sells after being produced has nothing to do with them.
Returning to the demand change for the second-row power seats, Li Xiang believed that electric seats were a must-have for second-row users such as the elderly and children, and it was also very convenient to install child safety seats. But at that time, it was calculated that changing the second-row seats from manual to electric would cost more than 600 yuan per car in addition to material costs. The larger cost was the engineering change: not only a new set of electric seats had to be made, but also a specialized seat rail for the second-row electric seats had to be developed, with a very high cost.
Even so, Li Xiang was still very insistent on electric seats, and the cost increase was acceptable. The only problem was that changing the seats after the engineering sample car phase would be very urgent, and the supplier delivery would be delayed. In the end, Li Auto’s chief engineer personally contacted the chief engineer of Faurecia Seats and through unremitting communication over several weeks, the supplier agreed to provide the seats on time.
This incident had a significant impact on Tang Jing. Before changing the second-row seats to electric, many colleagues would usually try the manual seats, pull the lever with their hands, and push the seats forward or backward with their legs. At that time, they just felt it was a little inconvenient, but no one would think it was a problem.He still remembers the question Li Xiang asked everyone: Why don’t we make our seats electric?
“Because everyone thinks that other cars are the same, and all cars at the same price point have manual seats. Engineers subconsciously assume the rationality of this matter, so they didn’t think that using manual seats is inappropriate.”
It wasn’t until Tang Jing started using the second-row electric seats that he realized how great this feature was. Looking back at the manual seats from years ago, he found them completely unacceptable. But without this real experience, such awareness is difficult to establish out of thin air.
“Most of our company’s employees used to work for companies that made cars priced between 100,000 and 200,000 yuan. Making a 200,000 yuan car was considered quite remarkable in the domestic market at the time, and many people haven’t driven cars priced over 300,000 yuan, or at least don’t drive them regularly. But our product managers spent a lot of time researching luxury brands and top-of-the-line luxury cars, their understanding of what makes a good car far exceeds that of our engineers.”
“Starting from the idea of changing to electric seats, Tang Jing began to realize the huge difference between ideal product managers and his past understanding of ‘product managers’. They are at the forefront of pursuing user experience. Even if they encounter resistance in engineering or are criticized by their boss, they still stick to their original intention and insist on what they believe is right.”
Later, the second-row electric seats received a large amount of positive feedback from users. Similar things made the engineering team begin to feel positive feedback from users. This also enabled the R&D team to gradually build trust with the product managers, and even try to actively understand their perspective and see what can be done to help from the engineering side.
Unconsciously, the engineering team’s values began to align with those of the product team, and the goals of the two groups began to unite.
Then came the airbag project, which Tang Jing called “the biggest challenge of his life.”
The engineering difficulty of this project was that the three-screen structure in the front row severely limited the layout of the passenger-side airbag. The addition of the passenger-side screen directly reduced the possibility of the airbag popping out from the front to 0, and it could only pop out from directly above the dashboard.
In order to ensure a comfortable viewing angle, the position of the passenger-side screen was already high, and there was also a sunshade above it. The challenge for the team was to make the passenger-side airbag pop out from the narrow triangle area between the dashboard and the windshield, and protect the passenger just like a normal airbag, without interfering with other components.
“The airbag has to pop out and come out. Going up is the windshield, going down is the cover and the screen. You only have about 150 millimeters of space to pop out the airbag without interference.” Tang Jing still remembers that his first thought after seeing these conditions was “how is this possible?”
At that time, domestic auto companies had already proposed solutions for passenger-side screens, which were later rejected due to the airbag issue, as no supplier could provide an airbag deployment scheme with this layout. And if the passenger-side screen is made low enough to ensure safe airbag deployment, the viewing experience for the passenger will be extremely poor.”This is the biggest engineering challenge I have ever encountered in my life. Sometimes I still find it unbelievable when I think back to it,” recalls Tang Jing, still somewhat excited about the project.
The three-screen co-pilot airbag project was an unprecedented engineering challenge, facing the same dilemma as other unprecedented projects: there was no successful experience to refer to within the industry and no supplier willing to take on the risk to do the job for you.
After completing the outer circle, Tang Jing was prepared for this action without external support.
With this awareness, Tang Jing and his team devoted themselves to the research of the project, fighting alongside his colleague Qiu Peng.
The space this time was too extreme. The size of the explosive force, the angle of the airbag popping out, and the angle of compression all needed to be strictly controlled. The team focused their efforts on this and conducted specialized research and testing every day. All the data they studied was unprecedented.
It can be said that from the moment the decision was made to take on the challenge of the three-screen airbag, Ideal Car’s engineers have been at the forefront of the industry, the only light in the dark being the torch in their own hands.
As time and testing progressed, engineers found that airbags had random problems during testing. In one round of airbag explosion tests, sometimes a whole group was fine, but in others, there would be a problem with the eighth or ninth one. Sometimes the seventh was normal, the eighth had a problem, but the ninth was okay.
Tang Jing and Qiu Peng went to find the reason and discovered that the airbags that had problems had ruptured. Then they analyzed why these airbags would break and began to study what technical details they had missed.
To find out the cause, they also consulted many industry experts, including some experts who were not yet Ideal suppliers at the time, but no one had an answer. At this point, the supplier had to give up and said, “We’ve already told you that this layout is not feasible.”
Finally, after a test, Tang Jing discovered the secret of the airbag ruptures. The co-pilot airbag cover had a plastic cover plate with some connecting ribs in the middle. It was similar in structure to the surrounding structure and was almost integrated with it. When the airbag popped out, the cover plate was lifted up and those connecting ribs were also broken.
The problem lies in the fact that these connecting ribs are made of fiberglass, which is unordered, and has no directionality when it is broken. Therefore, the breaking position of the connecting ribs is also random. Under this uncertain premise of the breaking position, in some cases, it may cause the airbag to come into contact with it. When the airbag bursts in just a few tens of milliseconds, sliding on the fiberglass fracture point that quickly may cause damage to the airbag.After identifying the root cause of the problem, Tang Jing and his team discussed solutions. Tang Jing’s initial proposal was to remove all the connecting ribs of the cover plate so that the airbag would not be cut when it popped out. After making the prototype and testing it, they found that the direction was correct and the airbag did not break anymore. However, this solution had the drawback that the upper part of the instrument panel on the copilot side would sink a little and there would be a shallow indentation visible upon closer inspection.
Although the effect of this solution was not perfect, the problem was finally resolved. If it were another company, the project would probably have stopped here and everyone would have followed this path that they had finally explored.
However, in Ideal Auto, there was always a group of people who asked themselves, “Have we done our best at all times?”
Qiu Peng, who had been working with Tang Jing on the project from beginning to end, had not stopped exploring. He thought of another solution to prevent the airbag from being cut: If there were stones on the ground that could hurt someone’s feet, wear shoes or lay a carpet. If fiberglass could cut the airbag, then he could add a layer of cloth to pad the airbag.
This idea was also successfully verified later. Compared to the previous solution, the instrument panel without cutting the connecting ribs inside looked exactly the same as it always had. Tang Jing was very happy about this and thought that Qiu Peng had found the perfect solution. Ultimately, this solution was successfully implemented.
Tang Jing imagined what would have happened if this had occurred in the past: “Engineers and project managers would have found ten thousand reasons why this problem could not be solved. It’s useless for you to try. Their way of thinking was, has anyone used this technology before? No, then I won’t use it. I won’t be the first one. This is usually how it goes. After all, proving that you can do it is very difficult, but proving that you can’t do it is too easy.”
“What we want to do is to create something that others cannot truly create. If we are the first in the world to use something, it will be very fulfilling.” When Tang Jing said this, his eyes were shining and the pride in his tone was palpable and admirable.
Today, Tang Jing often tells his engineers that Li Xiang will challenge them to their limits and they should think carefully whether their solutions have reached the extreme. Over time, the team’s “pursuit of the ultimate” thinking was forced out, and the engineers began to possess this kind of thinking. They would even take the initiative to find Li Xiang and say, “Boss, I have come up with a better solution. Do you want it?”
Later on, “mastering your own destiny, challenging the limits of growth” was written into Ideal Auto’s talent mission. Although it looks like a life mentor-style motto, it is very important for everyone in the Ideal team. Perhaps making cars is really too difficult, so everyone has to go all in.
When features and performance are in crisis, who will you save?# 2019 Shanghai Auto Show in April, the first user test of the Ideal ONE was conducted. At that time, the feedback on the car’s infotainment system was good. However, when the testing team hit the road, they found that there were still many problems with its performance and stability. The product manager, Shan Yi Fan, then initiated the action to cut off many of the car’s features.
The 701 building of the Beijing Liaison Building was the war room where the product, research and development, and quality control teams gathered every day to draw and cross out a flood of functional modules on the blackboard covering the room’s entire wall. Shan Yi Fan was heartbroken with every added function that was to be struck out, as he felt that the car was already approaching the unbearable bottom line of user experience.
However, there was a much more important bottom line within Ideal, which was the system’s fluidity and stability.
Fan Haoyu’s rival at that time was Apple. He was very excited when Apple decided to make cars, and he saw it as a great honor to be able to compete with Apple in his lifetime. He told the product team that they had to become the No. 1 product team in the industry, even surpassing people with dark hair and yellow skin like Apple. Apple first, Ideal second.
Apple had benchmark level fluidity and stability in the industry, and Li Xiang initially set an objective for the car’s infotainment system to achieve the same touch response experience as the iPads and flagship smartphones of that year.
The three functions that Shan Yi Fan reluctantly cut off were the Bluetooth earphones, Android Bluetooth keys, and Baidu Maps. When he was anxious, he would squeeze his forehead in his hands and start crying uncontrollably. Fan Haoyu vividly remembered taking Shan Yi Fan to the B3 parking lot of the Liaison Building where they smoked. It was a space communicating with the stairwell, and it was particularly high with a metal ring hanging from the top. Shan Yi Fan gazed at the ring and said, “You see that? If we don’t deliver it when we start producing, I’ll hang it here with a rope.”
Fan Haoyu made a bet with someone. At that time, the head of MIUI at Xiaomi, Jin Fan, had reserved an Ideal ONE and talked to Fan Haoyu about the new version of MIUI that they were developing. As the release dates of both were similar, they bet on who would release it first. From that point on, Fan Haoyu refused to cut his hair, waiting for the release, and many people saw him with long shoulder-length hair at that time.
On August 9, 2019, the Ideal ONE was preparing for a media test drive, and the marketing team and Fan Haoyu provided feedback that the infotainment system was very slow, indicating a severe issue.
When Fan Haoyu heard this feedback, he felt that something was off. The team had already conducted a rigorous test on the testing stand, and they had confirmed that the system’s fluidity was not a problem.
The next day, August 10, was Fan Haoyu’s birthday, and his wife had prepared a birthday cake ready for a small celebration at home. However, a large-scale media test drive was about to happen, and Fan Haoyu faintly felt that something was going to go wrong, so he found Shan Yi Fan on the same day to plan a test drive to Changzhou.On September 9th, Beijing was hit by heavy rain. Two engineers who were responsible for voice and app development were supposed to fly to Changzhou, but the flight was canceled due to weather problems. Therefore, they were driven to Changzhou with Fan Haoyu and Shuai Yifan.
As soon as they hit the road, problems arose. Initially, everything was fine, but once they left the urban area, the entire car’s software began to lag to the point of being unusable.
The engineers immediately connected a laptop to the car’s system to monitor performance consumption on-site. They found that the CPU was already running at 200%, and it was strange that the system was not crashing.
As they drove, the engineers tested each function and sent screenshots of the performance consumption of each application to their R&D colleagues. At first, they thought it was an isolated problem with that particular car.
After driving for about an hour and a half, they discovered that the system had restarted more than ten times. Whenever it restarted, it only took about five minutes before it started lagging again. The GPS location was also drifting, and the program couldn’t run at all. They realized that there was a big problem.
Shuai Yifan was furious. On one hand, there was a large-scale media test drive right in front of them, and if this problem were shown, it would be a disaster. On the other hand, they had already finished two rounds of reliability testing before this, but a big problem still occurred.
He immediately took out his phone and started scolding his colleagues over the phone. The person on the other end of the phone was also confused at first. The car had previously worked perfectly fine on the test bench, so why was there a problem now?
The situation was urgent, and more than ten specialized groups were immediately established to analyze each process’s system performance consumption. After analyzing the log data returned, they found that almost all applications were lagging, instead of just one or two.
The team speculated that the problem was not in one or two processes, but in the system’s performance. They further analyzed why the system worked fine during the test bench and ordinary city tests but developed a problem once they hit the highway.
Upon comparing the scenarios, they found that the issue was caused by the car’s speed. As the speed increased, the refresh rate of the GPS also increased, causing increased performance consumption. Furthermore, another module that consumed more performance as the speed increased was the signal parsing service. However, during the test bench, these two modules did not run at high loads.
Without hesitation, Fan Haoyu and Shuai Yifan in the car immediately called their colleagues on the phone and asked them to remove unnecessary special effects, animations, and high frame rates. The heavy rain that day caused poor visibility on the road, making driving very difficult. Furthermore, the atmosphere in the car was very tense after the system crashed.
Exhausted and nervous, they reached the vicinity of Linyi when it was already 3 am. They could not hold on and decided to stay overnight before continuing their journey the next day.
When we arrived in Changzhou the next day, the problems had already been mostly accounted for. At this critical moment, we received news that due to the influence of typhoon weather, travel had been obstructed in some areas, and the large-scale media test drive had been delayed for more than ten days.
The delayed date provided the software team with an additional two weeks of remedial time. During this period, in order to rescue the stability of the system, many people didn’t return home for several weeks. They would just lie down and sleep in the factory or office when they were tired, and then pick up their equipment and go out for road tests after they woke up.
During this time, Shuai Yifan ran an average of four to five hundred kilometers of road tests each day, with only four hours of sleep. He nearly covered all the roads near Changzhou. In this life-threatening state, the software team was able to produce a version that everyone could accept before the test drive began.
After the optimized version was put on the vehicle, the marketing department was amazed at how the fluency had improved so much in just two weeks during their experience with the car’s system.
After experiencing this event, the software team’s emphasis on road tests was raised several levels. They established a professional national road test team to run software testing in various regions. Combining the feedback from the road tests before the thunderstorm, they also found another problem: “their colleagues in road tests mainly focused on issues such as the chassis, durability, and sound, and were not sensitive enough to software. The situation when the test car stalled was similar to that of a traditional car, and it was considered relatively normal.”
At the end of 2019, with the formal production and delivery of the Ideal ONE, the long-awaited brand was finally registered. The maturity of the car’s system at the time of Ideal’s delivery was beyond many people’s expectations. From the delivery, the Ideal car’s system start-up speed ranked first among the new forces, rarely experiencing crashes. The system’s 100-millisecond response speed and fluency have also received a lot of user praise.
Fan Haoyu’s long hair was finally cut, and what surprised them was that many senior executives of domestic Internet companies were impressed by the software experience after driving the Ideal ONE and actively sought cooperation. Everyone saw the product’s capabilities and potential and wanted to stand together with a good product.
At that time, everyone exclaimed, “the product is the best spokesperson.”
The software team, which was almost at its limit before delivery, also had a breathing moment. The previously cut functions, such as in-car karaoke, QQ music, NetEase cloud music, Bluetooth headsets, and Android Bluetooth keys, were gradually added back within a year after delivery. The voice system also received a major update in September of this year, bringing the ability of full-car voice to a new level through OTA.
Thanks to the integration of the user account system and R&D task system, many problems raised by users were tracked and updated promptly. “The issues raised by users in the ideal app generate work orders directly within the system, and owners in the ideal test circle can also directly see the progress of feedback processing in the app.”
As of October this year, the Ideal ONE has had 18 OTA updates in total. This iteration speed has placed the Ideal in a leading position in the industry.# Different Facets of a Product Manager
For a automotive product manager, once the core hardware decision is made, it is difficult to change it for at least three years. However, for software product managers, unexpected things often happen when it comes to the practice of combining software and hardware in the all-new arena of intelligent electric vehicles, even though there are opportunities for quick fixes.
It is the details of product implementation that differentiate a really high-quality product from a mediocre one.
According to Fan Haoyu, a product manager needs both perseverance and endurance. Without perseverance, a product manager will be overwhelmed by numerous troubles, problems, and risks. Without endurance, it’s hard to persist in one thing for a long time. Fan Haoyu believes that the most enduring person in the intelligent cockpit product team is Shuai Yifan.
Li Xiang was completely unaware of the conditions that occurred within the team. He didn’t know the story in Changzhou or the specific pains of other team members. At the time, he was facing tremendous pressure on the company’s funding chain. This was all due to his insistence on adding range-extender.
A Contrarian Movement
As one of the first Tesla owners in China, Li Xiang drove over 100 kilometers to Gushuibei town in the winter of 2014. When he arrived, he found that the charging pile was broken, and he could only rely on the remaining electricity to return. He was very worried all the way but fortunately, he returned safely down the hill.
He realized that it was lucky he travelled alone. If he had brought his family with him, he would have been more anxious and uncomfortable. One of Li Xiang’s friends had encountered such a situation. His family was enjoying themselves, but they started arguing over charging, with his wife angrily saying that they shouldn’t have bought an electric car. Li Xiang realized that pure electricity could not solve all the problems if they want to make cars for family travel.
In 2016, when the idea of making a car for families became clear, Li Xiang led a few people responsible for engineering and development to hold a meeting in a second-floor office at Wuyuan Bridge. Li Xiang mentioned that he wanted to make a car that could be powered by electricity in the city and by petrol on long-distance travel.
Li Xiang explained the thinking behind his idea. In his opinion, charging infrastructure would not achieve a true large-scale popularization in the next few years, and long-distance travel would still need to be addressed by refueling. His solution was to make a large-battery range-extender. He continued to put forward specific requirements at the meeting. The pure electric range of this range-extender car should be able to cover 100 to 150 kilometers, meeting the commuting needs of first-tier cities, and it should be convenient to fuel up, that is, “powered by electricity in the city and by petrol on long-distance travel”.
“Many plug-in hybrid models that can run on either petrol or electricity already exist both domestically and abroad, don’t they?” Someone in the meeting immediately raised the question.# Li Xiang believes that many companies take subsidies just to cope with regulations. They often set the pure electric range of plug-in hybrids to 51 kilometers, but the actual pure electric range that continues to be discounted is not enough. Users completely treat them as gasoline cars. He emphasized that this is not what users really want, and the ideal is not to make such products.
The team carefully studied all the range extender and plug-in hybrid products on the market, and they found that:
The range extender is fully electrically driven and without a gearbox. Regardless of the level of electric power, the driving quality consistency of pure electric drive is high, and the system is more reliable.
The range extender can work in a more efficient range, and the fuel consumption is lower in the city, which accounts for the majority of the user’s driving time.
The large battery can provide better electrical foundation for the vehicle and is beneficial to the development of intelligent systems.
Soon, the range extender was finalized, but the engineering team was nervous. The reality before them was simple and cruel: there were no successful range extended car cases in the market before.
Ning Yunpeng, the person in charge of the Ideal vehicle control system, recalled the situation at that time. At that time, no one in the company knew what to do. When everyone talked privately, they said, you look at the market, who can sell range extenders except for BMW i3?
In the winter of 2016, Ning Yunpeng drove a benchmark range extender to the third ring bridge and was about to enter the highway. He had already stepped on the accelerator to the end, but the number under the instrument pointer only went from 30 km/h to 70 km/h after a while. During this period, cars constantly drove past him, honking their horns. Ning Yunpeng tightly held the steering wheel, praying not to be rear-ended by fast cars. At the same time, he also had an extremely profound idea in his heart: The company’s range extender car cannot be made like this in the future, if it is made like this, it will be finished.
After developing for a period of time, Ning Yunpeng had a deeper understanding of the range extender system.
The range extender system structure consists of three parts: the power generating range extender, the energy storage battery, and the motor that drives the vehicle. The strategy of the range extender system is how the three parts work together. The most difficult thing is to grasp the balance point.
Ning Yunpeng needs to find the balance point between power, economy, and NVH on the range extender system. This seemingly “balancing” task consumes a lot of his and the team’s energy, because the relationship between these three indicators is contradictory to each other.
Whenever an economical range for the range extender to work efficiently is found, the noise is often higher, so at this time, some efficiency must be sacrificed to reduce the noise. Li Xiang has very high requirements for noise reduction, so the engineers have tried various methods to improve every link that may reduce noise.
This work continued until the formal release of Ideal ONE. When Ideal ONE was launched, it was still the first plug-in hybrid with range extender product on the market. With Ideal’s success, more and more brands have launched range extended models.Even though the sales of the Ideanomics One keep growing and the customer satisfaction remains high, many people still think that the range extender technology is easily achievable without much technical expertise. However, Ning Yunpeng and his team are fully aware of the amount of effort they put in to build the system from scratch over several years. “Everyone can travel this road, but doing it well is another dimension, and no one can achieve it overnight.”
During the development of range extender technology, an investor who had completed the due diligence on the company called Li Xiang and expressed his doubts about the range extender technology, suggesting that as long as Ideanomics was willing to turn their product into a pure electric vehicle, he would be willing to invest.
Without hesitation, Li Xiang refused, stating that “this car was not built for investors.” Ideanomics has always had the least amount of financing among the top three new energy vehicle companies, and fundraising has been a constant issue for Li Xiang. Financing has been the company’s lifeline, and there was a time when the company was at risk of bankruptcy. Li Xiang’s Polymorphic Light Eruption, which caused a minor symptom, occurred when he couldn’t raise funds in 2019. Although Li Xiang did not hesitate to reject the investor, the pressure he felt internally was far greater than the product decisions he made.
Earlier this year, the range extender team reduced the vehicle’s fuel consumption by as much as 1 liter per 100 kilometers through a refined control strategy upgrade, with no hardware changes made, completely implemented through software OTA, something that has never been done before.
Regarding the new model X01 to be launched next year, Ning Yunpeng stated that many of the problems the team encountered and the experience it gained from the Ideanomics One’s range extender control can be reused, and from a software perspective, the team will introduce model-based predictive control, which is a completely new technology attempt.
Who would have thought that a few years ago, a team like this had never even touched on mass production range extender systems’ development, and the topic they talked about the most in private was “Why should we do range extenders when everyone who worked on them has already died?”
After the delivery of the Ideanomics One, more and more people began to learn about the range extender technology, and the urban electric and long-distance fuel consumption mode was highly recognized by users. Many car owners drove their Ideanomics Ones to Xinjiang, Tibet, and other places and shared their experiences on social media. They were often asked the question: “Can new energy vehicles go to Tibet?” In a third-party survey of Ideanomics car owners, 81% of respondents said they would not buy an Ideanomics One if it were a pure electric vehicle.
Positive market feedback after the launch of the Ideanomics One made the industry aware of the potential of range extender technology, and many companies began to follow suit.
In November 2021, the Ideanomics One delivered more than 13,000 units, becoming the first single-model monthly sales of more than 10,000 vehicles among the top three new energy vehicle companies. It is also the first Chinese brand to achieve monthly sales of more than 10,000 vehicles priced over 300,000 RMB, and the monthly revenue of a single model exceeding 4 billion RMB is also extremely rare in the history of the entire Chinese brand.Behind such sales, there is only one Ideal ONE model with an ideal configuration, and the average car buying decision cycle for Ideal ONE users is only 10 days, with an average inventory cycle of only 7 days. By comparison, Tesla, known for its pursuit of extreme efficiency, had an inventory cycle of 6 days in the third quarter of this year.
With precise market positioning for families with children, a strategy of offering only one configuration, and a high level of comprehensive product power, Ideal ONE has become a popular model in the Chinese market today.
The success of the explosion also owes much to the definition of the product, as well as the ability to implement it and the people who shoulder the mission and challenge their limits in the process of landing the product.
For the Ideal ONE team, of course, there are regrets, such as the child lock that users have been complaining about since mass production began, which could not be added until the redesign was released because of the long development cycle.
These regrettable decisions and successful persistence continually stimulate Ideal from two directions, helping this fast-moving company take better steps forward.
In the process of creating Ideal ONE, key decisions such as extended range, six seats, four screens, and full-vehicle voice control were all led by Li Xiang. Many people did not expect these seemingly non-disruptive things to later lead to industry follow-up. In the early stage of the company’s product team formation, Li Xiang was the most important product manager and played a core role in the most critical definition stage of Ideal ONE.
Next, Ideal’s product line will become even more diverse, and the organization structure of the product team will need to be further adjusted and upgraded accordingly.
In mid-2021, Ideal’s three product teams for the whole vehicle, cabin, and assisted driving merged and upgraded into a major product department, led by Fan Haoyu, to make the product team a development engine for the company.
In Li Xiang’s view, today his role as a car product manager is only 20%-40%, while his role as an organizational product manager is 60%-80%.
After the launch of Ideal ONE, many Ideal employees bought this car without internal discounts. Fan Haoyu, who once forced Shuai Yifan to cut functions with tears in his eyes, now occasionally sings karaoke with his children in the car.
As the company expands, the office becomes more and more cramped. Tang Jing’s Ideal ONE has become his mobile office. He completes almost all video interviews in the car, and takes a short break in the car during the interval between interviews.
In the process of using this car, more and more people have gradually recognized what the product team has done and slowly understood the company they are in.
Initially, when writing Ideal ONE sales forecasts, many employees, including Tang Jing, wrote numbers around 3,000 on the whiteboard. When this number was broken and surpassed many times, more and more people began to feel that the years of effort were valuable.After the announcement of the company’s plan to achieve annual sales of 1.6 million in 2025, everyone has begun to take this goal seriously. Compared to the sales expectation meeting of the Ideal ONE a few years ago, people now have more yearning for this goal.
In Li Xiang’s plan, to achieve this goal, ten cars need to be developed, and each of them must become a hot seller.
This article is a translation by ChatGPT of a Chinese report from 42HOW. If you have any questions about it, please email email@example.com.