Shulou(Shulou.com)06/02 Report--

Going back to December 2017, a data report showed that there were more than 2,000 smart speaker manufacturers in China at that time, most of which were "zero technical threshold" speakers that bought voice systems to carry their own hardware. At that time, the Chinese market shipped only about 1.5 million smart speakers in the whole year.

By 2018, more than 95% of the smart speaker manufacturers have disappeared, and the industry can even be said to be left with only several giants such as Baidu, Ali and Xiaomi, while the overall shipments of smart speakers may have increased tenfold.

The carnival is over, the market wakes up, what changes have happened to 2018 of smart speakers?

We can start from the origin of all changes and explore how smart speakers and the AI scene of Chinese families can move from dance to maturity.

After the ebb, the industry began to think from the user side, "what on earth is a speaker?"

During the thousand-box war, many people were not optimistic about the bubble of speakers because the product model was too simple, the technology was simple, and the experience was rudimentary. As long as they were infected with AI, they seemed to be able to sell money and raise money.

After the glamorous wave of capital disappeared, the whole AI industry began to reflect on the value of conversational AI. There is no doubt that the origin of the answer lies in the user.

If we rethink the meaning of putting smart speakers in living rooms and bedrooms from the user's point of view, we may find a few logic:

First, users do not want to "be able to talk", but to be able to produce meaning after dialogue, reach services, and complete deep-seated human-computer interaction. For example, it is not enough for voice assistants to communicate with people normally and have multiple conversations. The more ecological partners they connect, such as content creators, smart homes, software and hardware developers, and so on, the more likely they are to pry users' desire to buy.

Second, different users, especially home users need different choices, or even choose combinations, and can not use a simple speaker form to meet all the needs.

The service scene of the speaker in 2018 has been further extended to cars, refrigerators, television, education, pension, and so on. People's expectation of intelligent voice has become to connect everything and serve all members, so the carrier outside the speaker, such as the car screen, home audio and video terminals and other needs also arises at the historic moment.

Third, artificial intelligence speakers can only be toys, can not allow users to continue to open. This means that the technical indicators in voice interaction, such as awakening ability, recognition ability, speech synthesis, dialogue memory and so on, directly affect the user interaction experience.

Xiao du's assistant and Tencent Tinker Bell are all trying to push the natural Chengdu of voice interaction technology to a better level. For example, the small intelligent speaker Pro, which uses Baidu's new generation of intelligent voice interaction technology Endless Conversation, has evolved from question and answer to multiple continuous interactions with active conversations and active recommendations.

In short, when the tide recedes, the only truth left on the beach: if you play with users, users will play with you.

Under such reflection, intelligent speakers began the exit of players and a surge in shipments, and relative "latecomers" such as Baidu's small series, which mastered AI hard core technology, began to complete anti-surpassing. From the macro market trend, it seems that we can dig out three changes that happened to smart speakers in 2018.

Price disenchantment, technical yardstick revealed

The first change that smart speakers experienced in 2018 was that prices began to be reasonable and technology gradually became an absolute yardstick.

In mid-2017, smart speakers worth more than 2000 yuan were not uncommon, and prices have fallen furiously since then. The cheapest smart speakers can be given away almost for nothing.

The reason for this confusion lies in the random premium during the period of blind exploration and the fact that the cost is low. After some chaotic exploration, the speakers began to lock in the range of a few hundred yuan in 2018. This is a price range in which a Chinese family can afford it at will without too much decision-making costs, while ensuring sound quality, communication effects, and sharing the cost of AI technology.

For example, Baidu Xiaodu is at home, although it has the blessing of screen and call ability, it is still reduced to the price of 599 or even 299, which is a price suitable for opening the entrance to the home market.

After the price disenchantment, the change we can see is that the popularity of speakers is mainly driven by the degree of technical completion. As the speaker is an AI technology carrier with clear software technology effect, it is very easy to judge the performance-to-price ratio based on technology. Users who have mastered the AI ruler have started their own AI treasure hunt.

In this process, we can see that the net value of technology at three levels determines the acceptance of the market:

First, AI voice interaction technology. The three dimensions of far-field recognition, multi-round dialogue and semantic understanding have become the most basic technical reference for comparing product capabilities. For example, the unawakened geek model launched by the Xiaodu series has become the technical benchmark in the industry.

Second, the depth of combination with hardware. The solution of loudspeaker with screen, the solution of car speaker and the matrix of sound receiving equipment have become the main reference basis for the integration of software and hardware of loudspeaker.

Third, more and more important "skills": AI is not only capable of dialogue, but also requires various abilities, such as guessing games, learning interaction, and integrated Internet services. Similar skills are based on the dimension of the development platform, and technological hard core and ecological construction constitute another track for speakers.

For example, Xiaodu has more than 800 skill points under the blessing of DuerOS technical capability. This kind of "big gift bag"-like sense of use is obviously easier to bring surprise to consumers. This is indeed the case. Data show that the market share and shipments of smart speakers Q3 equipped with DuerOS are the first, the shipments of small series have increased by more than 711% month-on-month, and cooperative manufacturers such as SUNING and Doss have also developed rapidly.

After the price tends to be stable, AI technology begins to become the main value coordinate system of intelligent speakers, affecting the market's every move. Baidu's Xiaodu series, which has emerged as a new force under such value recognition, has become an effective case for technology to deeply confirm the depth of the market.

Product matrixization and re-segmentation of family scenes

Home scene is a relatively complex smart hardware market. Different families need different products, different rooms need different conversational AI capabilities, and more and more travel scenes make family car scenes become more and more important.

How to meet different needs, different scenarios of hardware connectivity and network access?

Obviously, it is not reliable to use the same speaker to accomplish these tasks, so in 2018 there is a second trend of smart speakers: product matrix and re-separation from home scenes.

From the Xiaodu series, we can see the direct embodiment of this trend:

Xiaodu at home, based on screen mode, provides more visualization skills and remote care capabilities for family scenes, meeting the direct needs of most Chinese families.

On the other hand, the small degree intelligent speaker has cut into the young people's AI family market with a cost-effective mode, and can build an AI network between the bedroom and living room with Xiao du at home, shaping the intelligence of the whole house.

Small intelligent speaker Pro with a higher level of sound quality, cut into the family category that has higher requirements for sound quality and likes the touch of higher-level technology. The birth of the vehicle bracket meets the call for small degrees in the car.

Different families in the family can choose their own living room-bedroom-car scene coherent solution, putting AI into the family has become a lower threshold and more imaginative thing.

The reverse case of this trend is that many single speaker products are gradually running out of surplus value, while the blind indifference product matrix makes many manufacturers fall into huge invalid cost pressure.

Understanding the subdivision needs of Chinese families and having the technical ability and product decision-making power to make matrix actions can be seen as the second change of intelligent speakers in 2018.

The deep water area of content and the beginning of the vertical content market

In intelligent speaker logic, AI is the beginning of interaction and service, while content is the purpose of interaction and the entity of service.

Under this logic, we must admit that in addition to AI technology, the content war is the main path of smart speaker competition. In 2018, the layout of the content field was continuously optimized, and the copyright parties joined forces, squeezing the content war of smart speakers into the deep water area. Content quantity and ecological construction began to become the focus of consumers' attention.

Another change that must be paid attention to is that the smart speaker children's education market, which consists of children's content, early education, Chinese and English learning, opened fiercely in 2018. There is even a trend of self-building ecology in the field of AI voice-children's education.

In order to improve the differential competitiveness in this field, the giants of smart speakers have also started a battle for content.

Taking Xiaodu series as an example, it has accumulated in 2018 content system and established a cooperative ecology with many head copyright parties. Access to NetEyun Music and QQ Music 20 million + authentic music library resources, establish strategic cooperation with CITIC College, integrate iqiyi, Discovery, good video, partner shooting small video, ahaschool, Xueyusi, VIPKID, Douyu TV and other massive video resources, a set of combination punches down, the moat of the content system has also been solid.

The first-mover advantage in the content of children's education also makes the small series in the rapidly opening tuyere of children's education. On the basis of 800000 + boutique children's audio programs and 20 million + certified mother-to-child Q & A content, we took the lead in launching mature children's model services, which opened the door for parents with technical solutions + vertical content ecology.

The integration of content resources will continue in 2019, and the ability to open the second and third vertical content market may be the key to the giant speaker battle.

Generally speaking, in 2018, intelligent speakers began to get rid of the dryness and heat, and developed towards a diversified path of comprehensive comparison in three aspects: the completion of AI technology, real use space and content accumulation. The position battle of the giant and the rapid development of AI voice technology have supported the rapid opening of the market, forming a well-deserved first market in the field of AI hardware.

Product matrix, vertical content and hardcore AI technology should be key words for the continuous evolution of 2019 smart speakers.

Welcome to subscribe "Shulou Technology Information " to get latest news, interesting things and hot topics in the IT industry, and controls the hottest and latest Internet news, technology news and IT industry trends.