by After two years, Apple resurrected the HomePod.
This practice is confusing. After all, two years ago, Apple announced the discontinuation of production and the removal of HomePod, sending a clear message to the outside world: this is a failed product.
The new HomePod released this time has not changed much in terms of design, function, and product positioning. It is hard to imagine how it will reverse the fate of its failure.
At the same time, the external market environment has become more difficult. The entire smart speaker market has entered a bottleneck period of development. Google and Amazon, which are market leaders, are abolishing their own speaker product teams.
Against this background, Apple still chooses to resurrect HomePod and insists on occupying a place in this field. It is hard not to wonder about the strategic considerations behind it.
Apple HomePod product map｜Apple
The explosion of advanced natural language processing models such as ChatGPT also provides more room for imagination for the future of smart speaker products.
Will HomePod be Apple’s first product to use ChatGPT technology?
HomePod failures and retries
In the summer of 2017, Apple released the HomePod for the first time.
This is a rare place for Apple to release a brand new product as a “follower”. Before HomePod, Amazon successfully defined the product category of “smart speakers” through Echo, and to a certain extent forced Google and Apple to follow up and develop products with similar functional positioning.
As the leader and disruptor of consumer computing hardware, Apple will not be willing to “follow”.
On the HomePod, Apple is equipped with “computational audio” technology for the first time. It allows HomePod to recognize and adapt to the environment. Users only need to place it on a wall or in a corner, and it can automatically fill the entire room with sound, providing an excellent sound experience.
In order to provide an advanced and excellent sound experience, Apple’s “stacking” of HomePod can be said to be cost-effective. It is equipped with 7 mid-range and high-range sound units, a subwoofer, 6 microphones for receiving voice commands, and an A8 chip for processing audio signals.
That makes HomePod the lowest-margin product in Apple’s history.
At the beginning of the release, there are estimates from market institutions,The material cost of HomePod is as high as more than 200 US dollars, and the gross profit is only less than 40%, which is not only far lower than that of Apple products such as iPhone and iPad,It is also lower than similar competitors from Amazon and Google.
Despite high costs and thin margins, HomePod still sells for well above the market average. The previous generation of HomePod was priced at 2,699 yuan, or $349. In contrast, Amazon Echo’s price baseline is $99, and when it comes to discounts, it can be as low as $50, and sometimes even “free” with Amazon Prime members. This made HomePod quickly fall into a competitive disadvantage in its early days.
The bigger problem is that, although HomePod has excellent sound quality, it is not “smart” enough. Apple has put most of its chips on “sound quality”, and is relatively behind in the crucial voice interaction. The depth and breadth of Siri’s capabilities lags significantly behind Amazon’s Alexa and Google’s Google Assistant.
Apple HomePod product concept map｜Apple
Technically, HomePod means a lot to Apple.“Computing audio” later became an extremely important technical asset of Apple, and was used in almost all series of products such as Mac, iPad, iPhone, AirPods, etc., and was recognized by consumers. But at the product level, HomePod has suffered a major failure.
A year or two after its launch, the HomePod has only a single-digit share of the U.S. smart speaker market. In response to this situation, in the fall of 2020, Apple launched the HomePod mini, priced at only 749 yuan, $99, bringing the front to the same level as its competitors.
A few months after the launch of the HomePod mini,In the spring of 2021, Apple quickly announced that it would stop production and remove the HomePod, directly “abolishing” this product line.While off the shelves, Apple did not forget to emphasize that the sales of HomePod mini are gratifying.
Indeed, according to a report from market research firm Strategy Analytics, after the release of the HomePod mini, Apple’s share of the smart speaker market has almost doubled, rising from 5% to 10%.
Reviewing the failure of HomePod, it is not difficult to find that the psychological price of smart speaker users has an obvious ceiling. Users are willing to buy smart speakers, but they are not willing to pay high prices for them.
In addition, the most important thing for users to purchase speakers is still the “smart voice function”. Amazon and Google, which are relatively leading in this regard, have become industry leaders.
Apple, which tried to save the country through the audio technology curve, failed to reverse the fate of HomePod’s failure.
HomePod works with iPhone products｜Apple
Viewed in this light, the new HomePod is all the more puzzling because so little has changed. Compared with the previous generation, its design and function have almost no improvement, but a small function of detecting indoor temperature and humidity has been added. Apple has reduced the number of sound units and microphones inside the new HomePod. This is most likely to better control costs.
Although by optimizing the audio processing algorithm, the new HomePod maintains the same high level of sound quality as the previous generation. But its price is still not low. The price of 2299 yuan and 299 US dollars means that it is still one of the most expensive smart speakers on the market.
Just when the new HomePod was launched, the external market environment became more and more precarious.
The Curse of “Unintelligence”
Through the story of HomePod, it is not difficult to see that the road to the rise of smart speakers must be “hard” smart. It is a hurdle that cannot be bypassed.
The curse of smart speakers being “not smart enough” is not only lingering on Apple, but also Google and Amazon are facing similar situations:The sales and market penetration rate of smart speakers are still growing, but they have encountered obvious bottlenecks.
In 2022, the global smart speaker market size is about 10.3 billion US dollars. In contrast, smart watches, which also entered a period of rapid development in 2014, have a market size of US$33.4 billion. Another category that was born later and is the same audio equipment as speakers, true wireless earphones, has a market size of 50 billion US dollars, and the growth rate is much faster.
The “market size” calculated by market organizations based on research is not very rigorous. However, from the gap between smart speakers, watches, and earphones, it can still be seen that the overall performance of the category of smart speakers is weak.
The problem still has to be attributed to “not smart”, that is, the voice assistant is not easy to use.
thenWhen the voice assistant was just born and became popular, the industry shouted the slogan of “voice-first” (voice interaction priority). People had expected it to completely subvert the interaction, become a key auxiliary of touch, and even replace touch and become the mainstream.Various manufacturers have begun to bet on voice interaction technology and expand their teams.
Facts have proved that voice interaction lacks depth and precision, often fails to recognize instructions, and the feedback and interaction are not good enough. For example, many users will ask smart speakers some questions, but the speakers do not have the ability to understand the context and cannot form a coherent dialogue; Instructions, the speaker cannot determine what song the user wants to play, after all, it does not have a “search” interface.
The combination of these factors leads to a low usage rate of speaker users, and the commands issued often stay on the surface, mainly for simple interactions such as asking the weather and setting a countdown, which are difficult to bind needs and form a stable usage scenario.A large number of speaker products have not escaped the fate of “eating ashes”.
Amazon Alexa was once a very successful series of smart speaker products｜Amazon
In the end, all manufacturers have to face reality.
In December last year, Amazon announced the layoffs of 10,000 people. Among them, “Devices and Services”, which is responsible for the development of Echo speakers and Alexa voice assistants, became the hardest hit area of layoffs.
Amazon takes charge of ‘devices and services’ after layoffs announcedSenior Vice President David Limp was interviewed by Bloomberg. He said that the focus of the abolition is some experimental projects of Alexa, including telehealth services and games. These projects are full of unconstrained imagination, but the development process is faced with huge resistance, and there is no hope of breakthrough.
Limp said,Alexa’s active users have been growing, but the entire Alexa and Echo business isn’t profitable.Because the hardware profit of speakers is not high, they are often discounted and sold at reduced prices. Limp denied rumors that Amazon lost $5 billion on the Echo and Alexa businesses, but also refused to provide specific figures.
Something similar happened at Google.
Among the large-scale layoffs launched in January this year, the development team of Google’s smart home operating system Fuchsia faced the most tragic situation, with about 16% of employees being laid off. In 2018, when Fuchsia was just launched within Google, there were rumors that Google’s smart speakers would be the first to be equipped with the Fuchsia system. Obviously, it also carried Google’s imagination of “voice interaction”, but failed to achieve a breakthrough.
Now, Google has already cut its investment in smart speaker products,Since March 2021, Google has not updated its speaker products for two years.
Waiting for the next “burst”
Zoom in on the vision a little bit, the dilemma faced by smart speakers today,Rather than saying it is a “recession”, it is better to say it is a “precipitation”.
As a computing device, the future of smart speakers still has a lot of room for imagination, but today’s voice interaction technology is not enough to realize those imaginations.
From this point of view,Although Google and Amazon have cut their respective speaker teams, their exploration of voice interaction and natural language processing has not stopped.Amazon still has nearly 10,000 employees and is engaged in the development of projects related to Alexa; Google has been exploring AI Technology interaction possible.
Including the recently popular ChatGPT, it also brings more hope to the industry.This type of large-scale natural language processing model can just solve the problem of traditional voice assistants “unable to understand the context and not accurate enough”.It can provide better accuracy, scalability, and coherence for voice interaction, allowing machines to understand the ability of human voice commands to achieve an exponential leap.
Of course, the GPT model’s reliance on huge computing power means that it cannot be embedded in voice assistant products in the short term. However, the success of ChatGPT still provides a basis for the feasibility of a “truly intelligent” voice assistant.
interview, amazonSenior Vice President Limp also mentioned ChatGPT. He said that he tried ChatGPT, and the experience of talking with it is very good, and more models will come out in the future, and these models will play a role in more different usage scenarios. There is only one biggest problem right now, that is, the cost of training the model is too high.
The story is far from over. After ChatGPT shocked the world, almost every company began to focus on AI field.
Google has sounded a “red alert” internally.CEO Sundar Pichai recalled the two founders, allowing them to participate in future product roadmap planning, and also mobilized multiple teams, asking them to quickly participate in the AI In the development of the project, speed up the progress. According to reports,Google plans to release more than 20 AI-related products and features this year to provide users with a “conversational” experience.
At the same time, Apple is also holding a AI Summit, although this event is more similar to a “team building”, it does not mean that Apple will launch any new products and features soon.But it still reveals Apple’s attention and emphasis on AI technology, which may also be part of Apple’s “resurrection” HomePod.
It’s easy to think of the story of the Macintosh back then.
In 1984, Jobs identified the “graphical interface” as the future of personal computers and put it on the first-generation Macintosh. But he came too early, and the Macintosh failed to win users no matter from technical or cost considerations.
It wasn’t until 10 years later that graphical interfaces really started to go mainstream.
Now, similar things may be happening in voice interaction, it is waiting for the next “burst”.
