This article by”
AI the front“Original, original
Ali lost money to sell millions, how can Amazon dad easily sell millions of smart speakers?
Author | Xu Fei
Editor | Qianqian, Betty, Emily
In the United States, Amazon is known primarily as an e-commerce and cloud computing company. This impression has been greatly changed in the last two years with the launch of Echo, a smart speaker, and it turns out that Amazon is also good at making smart hardware.
So what’s the history of Amazon’s smart speaker? How does it strategize? Let’s start with a brief review of Amazon’s hardware development path.
This article is excerpted from Dr. Fei Xu’s geek Hour column “Technical and Business Case Studies”. For the full version, click here to subscribe.
From Kindle to Fire Phone: Amazon’s Hardware path
Amazon has been making hardware for a long time before it made smart speakers. Two of its best known products are the Kindle e-reader and the Fire Phone. But while the Kindle was a huge success, the Fire Phone was a complete flop.
Amazon’s experience selling e-books gave it experience running digital products, which it expanded into music, video, games and more. Amazon, like Apple, is one of the few Internet companies to gain experience with digital products early on.
Compared to the Kindle, the Fire Phone, Amazon’s highest-profile, upmarket launch, was a complete failure and was ignored, forcing Amazon to discount the device aggressively. Such a fiasco is undoubtedly the first and only time in amazon’s entire history.
Why did Amazon fail miserably? Amazon’s leadership guidelines give us the answer.
Amazon’s leadership principle emphasizes “customer first”, and the development of this phone, in terms of price and function, does not reflect this leadership principle: expensive, dynamic 3D flashy. Firefly is designed to make it easier for people to make purchases on Amazon, while the Cloud browser is a naked collection of user data. When Amazon itself doesn’t follow its own leadership guidelines, a product fiasco is “inevitable.”
This fiasco, let Amazon’s hardware research and development team, the mysterious Lab126 suffered a heavy blow. Amazon’s hardware development and sales were radically changed by the Fire Phone fiasco, with the entire Phone line and a virtual reality-related product that had been in the works at the time shut down. There has also been soul-searching within Amazon about the fiasco.
And that’s when a development project that began in 2011 and has been constantly downgraded to make way for mobile phones turned into a historic opportunity. The project became known as the Echo.
Start the smart speaker battle: Echo conquers territory
About six months after the Fire Phone failed and the Echo was finalized, Amazon decided to release it. The launch was very low-key, just gradually pushing the message to its Prime members that they could buy the speaker and give it a try. At the time, the stock of speakers was also small, but the Echo was an instant hit for Amazon.
Amazon sold more than 2.5 million speakers in 2015 and doubled that to 5 million in 2016. In just two years, the Echo smart speaker has knocked the entry-level professional speakers out of the market and wiped out brands.
The rise of the Echo has a lot to learn and think about:
For one thing, amazon itself wasn’t quite sure whether the speaker would be successful in the first place. So testing the waters in a small area is a safe bet. If it works, keep adding more, and if it doesn’t, stop your losses in time. In the event of failure, Amazon will not produce too much and run out of control.
Second, amazon’s speed and ability to respond to the market was admirable after it found that the speaker was selling well. Realizing that an open speaker ecosystem was the foundation for their growth, Amazon opened up its third-party skills platform to allow speakers to access third-party music providers. This approach is in line with Amazon’s leadership mantra, which is that leaders should take a long-term view, not just focus on short-term profits. Of course, there is, is to do things need to do “customer to fashion”.
Generally speaking, there are some accidental factors in the birth of Echo, but after the birth, Amazon fully implemented the principles of leadership and adopted a series of strategies around the development of Echo, making it truly successful.
The battle for smart speakers: Voice assistant Alexa
Voice interaction, or “voice assistant,” as it’s more commonly known, is the most important module for Echo, a core feature of smart speakers. The hottest voice assistant of the last two years is Alexa, which is owned by Amazon.
Go all out to build voice assistants
How did Alexa come about? Back in August 2015, about 10 months after the Echo was sold, Amazon suddenly realized that its voice assistant could exist as a standalone cloud service and didn’t have to be bundled with the Echo.
So Amazon quietly changed the name of its voice assistant, Echo, to Alexa. Alexa was originally a web analytics tool owned by Amazon, but the voice assistant has taken over. Before long, no one remembered the site analytics tool, and Alexa became the exclusive term for Amazon’s voice assistant.
But in order to ensure backward compatibility, the Echo voice assistant’s original Hello Echo reminder remains, even though the default is Hello Alexa. Only in the reminder words do we get a hint of the past: Alexa was once called Echo.
When Amazon first developed the speaker, it didn’t know the importance of voice processing, so the technology of the assistant was mediocre. Amazon was quick to recognize the importance of voice processing when Echo speakers sold well, but the talent pool for voice processing was small. Nuance, a Seattle-based company that is not particularly well known in the consumer market but is well known in the field of professional speech processing.
Nuance has a research and development center in Seattle that specializes in voice-processing technology, and Amazon is paying a lot of money to scour the company. Eventually, much of Nuance’s core Seattle r&d center moved to Amazon.
In order to further strengthen its technology and talent pool in the field of voice processing, Amazon moved quickly to acquire two startups, Yap and Evi. At this point, Amazon finally some peace of mind, to Alexa reserve enough voice processing talent.
Using machine learning to tackle speech recognition
Echo is first and foremost a speaker, and a speaker’s main function is to play music. But interacting with speakers at events such as dinners and dances, where there are many other sources of sound, presents a big problem: How do speakers recognize real interactive commands amid the noise?
Alexa, the voice assistant on Echo speakers, has very different technical requirements from Alexa on mobile phones due to different application scenarios. Simply put, if the difficulty of voice interaction on a mobile phone is one star, then the technology of voice interaction in a deep noisy environment should be at least four stars. This is a conundrum that the Echo smart speaker will have to solve.
In a public exchange, Rohit Prasad, the lead scientist for Alexa and a noted machine learning scholar, briefly mentioned in an interview that the Lab126 team had stalled the Echo smart speaker project because of this problem, Finally, he had to ask for help from the whole company.
The next solution is machine learning. Amazon has released an audio clip on the web that compares raw sound in a noisy environment with machine-learning-processed sound. As you can see from the audio, the audio processed by machine learning has achieved almost perfect noise filtering, which is the main reason why Echo can still show very good speech recognition function under extremely bad and noisy environment.
Echo’s ability to recognize voice interaction in noisy environments is one of the most important reasons for its rapid emergence.
At a company that has long had a reputation for making business decisions over technology, Amazon’s core technology isn’t that impressive. In this case, however, Amazon’s ability to use technology to solve difficult problems made a big difference to my impression of amazon.
Voice platform promotes growth
Alexa has been integrated across Amazon devices since it became independent of The Echo, such as amazon’s Fire TV. Alexa is a big deal inside Amazon, and any project that can be integrated with Alexa takes a lot more priority inside Amazon.
In addition to domestic sales, Amazon has exported the voice assistant to smart refrigerators, cars and Even Huawei phones. The ubiquity of Alexa puts Amazon in a lot of good positions as a latecomer to the voice assistant market.
From the technology development, Alexa also has its own independent R & D team, and has been very large, including a director, a lot of people. At Amazon, if a director-level person is involved in a project, it means that the project is really important enough for c-level executives to see and focus on directly. With the director-level role, Alexa’s status at Amazon has become more prominent, and it’s no longer just a side project under the Echo.
Inside Amazon, many people believe that voice interaction is a very important new traffic channel, and will become even more important in the future. So the team has been swelling ever since Alexa went independent. Interestingly, I’m getting a lot of offers from Amazon recruiters on LinkedIn, and the percentage of hires for Alexa is consistently high.
If Echo as a smart speaker was important to Amazon’s conquest of the living room and kitchen in the first place, Alexa, amazon’s voice assistant and the skills platform it opens up to third parties, is now the company’s top priority. If Amazon had focused more on speakers before, Alexa’s development as a separate piece of software has shifted its focus from hardware to software and platform: hardware, including smart speakers, serves the software platform. Indeed, the hardware amazon is developing today has become primarily for Alexa, and the hardware itself is far less useful than the voice platform.
With this platform, Amazon can promote voice not only on its own hardware, but also open it up to third parties. Get third-party hardware into the platform. There are huawei mobile phones, LG electronics, the future we can see, such as car systems and so on.
Voice platforms have a much bigger impact than hardware, and ultimately platforms create an ecosystem, and any ecosystem is very sticky. Amazon could also use the platform to expand its influence beyond its own business, allowing Alexa to become an ecosystem that many companies across the Internet and IT industry would like to join in a way that smart speakers have never been able to, and amazon has never been able to, but is now able to succeed.
The shift from Echo speakers to Alexa voice assistant is a big change in priorities. This transition took amazon 10 months, and I was amazed at the speed with which Amazon responded to the market. At other Internet or software companies, it’s not uncommon for the transition to take three to five years. So there’s a reason why Amazon is able to dominate the voice market.
The battle for smart Speakers: Amazon’s strategic layout
In addition to being an early entrant or inventor of smart speakers, Amazon is also a company that is very responsive to the market, so it naturally has its own strategic layout. Whether or not another competitor comes in, Amazon needs to develop smart speakers.
Amazon’s initial approach to Echo was to cast a net everywhere, which was not very clever. First up is the Tap, a portable smart speaker that you can take with you when you travel. But we know that carrying a speaker on a trip is a rare event. What’s more, Alexa needs to be connected to the Internet to work, and Tap doesn’t have a built-in phone chip, so it has to connect to WiFi to connect to the Internet, so it’s not selling well.
The second is the Echo Dot, which you can think of as a cheap Echo. The Echo Dot was upgraded quickly, reaching its second generation in less than six months. It has a microphone, speakers, Alexa, and the ability to connect to other high-end speakers and turn them into smart speakers.
In this way, those who like higher quality speakers can have the functions of smart speakers and the sound quality of high-end speakers. The cheap Echo Dot is also a good option for people who need Alexa voices everywhere in the house, but not necessarily listening to music.
You can put an Echo Dot in different rooms, and only in the hall, and “combine high and low” to form a complete smart home voice control system. Amazon is encouraging people to buy more Echo Dot, so it’s offering discounts on the price of multiple Echo Dot purchases.
Now that Google is offering video interaction on TVS through Chromecast and Google TV, Amazon is offering a counterstrategy: The Echo Show.
The Echo Show is simply an Echo with a screen, a decent speaker, a small screen, and Alexa to provide video and voice answers. When this product came out, a lot of people liked it.
Google is aggressively pushing smart speakers, and Amazon did an unprecedented Prime Day sale. The promotion is focused on Amazon’s smart speakers. The Echo series were all discounted, and the Echo speaker was even discounted by as much as 50%, so many members bought it that day. In my moments, many people posted pictures of themselves buying speakers or cutting off their hands that day.
At the time, there were two theories: one was that Amazon was using massive subsidies to quickly capture the market and further consolidate its dominance; Another theory is that Amazon is clearing out its inventory because it’s likely that the next generation of Echo products will come out and dominate the market.
It turned out to be Amazon’s Echo 2. The second generation Echo comes in Echo and Echo Plus versions. The Echo is half as tall and magically costs $99, less than Google’s smart speaker. But the Echo is said to have better sound quality than its predecessor, despite being half the size, thanks to a new woofer.
At $149, the Echo Plus looks a lot like the older Echo, at least from the height. Of course, at $149, it’s also cheaper than its predecessor.
The Echo Plus’s biggest selling point is its control of the smart home. Unlike its predecessor, Echo Plus will automatically detect and set up all smart homes connected to the same LAN. Once out of the box, you can control your smart home directly through Alexa, the voice assistant, without any manual setup.
This does seem like a very useful feature. To demonstrate the value of this feature, the Echo Plus also comes with a Philips smart light bulb. As for the effect, the market may be tested.
In its announcement, Amazon also said that after studying user behavior for three years, it found that the top three things Alexa users like to do are turning on lights, starting the coffee machine, and reading new messages. So in the future, users can program routines to do all of these things, say, “Alexa Good morning.”
Amazon also unveiled a new product, the Echo Spot, a 2.5-inch device with a rounded display that is smaller than the Echo Show. The display displays the time, the weather and even video conferencing with the user. For a product with such a small screen, the Echo Spot doesn’t strike me as cheap, at $129.
Last came the $20 Echo Button. It has very limited functionality and is currently used for voice interaction, such as answering questions.
When Amazon released the new Echo series, I was a bit upset. Yes, I’m the one who bought the first Generation Echo speaker at half price on Prime Day. It doesn’t feel good to be liquidated. After I bought the first generation, whether to buy the second generation, the question becomes awkward.
In the face of Google’s aggressive push into the smart speaker market, Amazon has introduced a new generation of products that mainly address the following issues.
The first is Google’s low price. For a company whose leadership code is frugality, whose cost control is life, and whose core value is low price, it is hardly pleasant to be outsmarted by a dark tech company in the smart speaker business it invented. So with the Echo 2, Amazon dramatically lowered the price of the speaker and suddenly surpassed Google in terms of cost performance. That gives Amazon a price advantage.
The second is a backlash against Google’s Chromcast and the TV. Google offers a unique experience through the combination of speakers and Chromecast, as well as Chromecast and TV. Amazon’s counter-attack is roundabout in two ways: The Echo Show, which can be placed in a living room or something; One is the new Echo Spot, which is sleek enough to sit on a desk and video chat with other people. It’s hard to tell whether a big-screen TV is a better experience or the new Echo hardware is a reboot. Guess it’s different to each other.
Third, and perhaps defining, is amazon’s desire to be the “controller” for the smart home. Why is it “controller”? Because in the last generation of products, the control of smart home is to be completed through the setting of skills. Although set up after good also very good use, but unavoidably some people do not understand or not familiar with the use of computers. Occupy intelligence to live in the market better so, “let the control of intelligent equipment is simplified, silly change” also carried on the agenda. This Echo Plus device is designed to do just that.
Lower prices, more ways to interact, and enhancements to the smart home controller are basically the main goals of amazon’s new product launch, and amazon has a clear business vision.
So in the long run, the battle between Amazon and Google over smart speakers will continue. I think it’s more likely to end up in a showdown.
The battle for smart speakers: The big boys are marching in
Google’s entry comes a year after two other IT giants, Apple and Microsoft, announced their own foray into the smart speaker market. Apple made its entrance at WWDC 2017 with the announcement of a smart speaker called HomePod, which had been scheduled for delivery by the end of the year, but has since been delayed.
Not long before that, on May 9, 2017, Karman Kardon, a maker of traditional speakers, announced a partnership with Microsoft to launch a smart speaker powered by Microsoft’s voice assistant Cortana, called Invoke.
Enter intelligent speaker market, not only have foreign enterprises, also have domestic enterprises. The most famous and earliest one in China is JINGdong, which is affiliated with IFLYtek. Iflytek as a pioneer of Chinese voice technology, after more than ten years of accumulation, for Chinese voice processing and recognition technology master, is basically the world’s most cattle. The “Ding Dong” speaker, the product of a collaboration between IFLYtek and JD.com, was launched not too late, its functionality was not satisfactory, and the ecosystem did not do well. In general, Ding Dong should be no match for Amazon and Google.
In addition, the domestic intelligent speaker still has alibaba’s tmall genie X1, Himalayan Small elegant AI speaker, lenovo launched lenovo intelligent speaker, miui launched Miui AI speaker, and so on. These speakers are difficult to reach the level of Amazon or Google speakers in a noisy environment, whether the domestic speakers can ride the dust, it still needs a period of time.
About the author:
Xu Fei, doctor of Computer science, senior architect, big data infrastructure expert, author of Geek Time App column “Technology and Business Case Interpretation”. Currently living in the United States, he has a profound and unique view of the global computer industry and the development of cutting-edge technology.
For more content, you can follow AI Front, ID: AI-front, reply “AI”, “TF”, “big Data” to get AI Front series PDF mini-book and skill Map.