As the camera lens swept through the 2006 Google search engine conference, Eric Schmidt, 52, was sitting at the interview table and talking: "it's based on the premise that data services and architecture should be on the server, and we call it cloud computing." "

It was the first appearance of the concept of "cloud computing".

Then, with a trademark smile, the then Google CEO turned around and said impolitely: "I don't think people really understand how big this opportunity is, but Google will go down this road." "

Amazon opened its cloud service platform (AWS) to the public after more than a decade of his voice.

By contrast, Google's program engine Google App Engine (cloud services) was not released until April 2008.

It's really "getting up early in the morning and catching up for an evening episode, and the clown is me."

On the other hand, Amazon AWS, which is "killed halfway", has become Amazon's most profitable business by virtue of its first-mover advantage.

How much money do you make?

According to the data, Amazon's total operating profit in 2019 was $13.681 billion, with AWS contributing $9.2 billion, accounting for 67%. If Amazon earns 2019 yuan, 67 yuan is earned by cloud business.

At present, investment banking analysts have separately valued Amazon's cloud business at $425.6 billion, on a par with its main e-commerce business at $486.5 billion.

Such a "money-absorbing" ability not only gives outsiders insight into the strength of AWS, but also makes the public clearly aware of the broad prospects of cloud computing.

When "Cloud Computing" entered the second half, Thomas Curian, the new CEO of Google Cloud, declared that he would increase the investment in cloud computing sales to compete with Amazon; Microsoft's transformation slogan of "cloud computing first, mobile first" also revealed the "ambition" of cloud computing.

Amazon couldn't keep laughing because of its fierce rivals.

Since 2018, the market share of Amazon Cloud IaaS business has continued to decline, and it is the laggard Microsoft Cloud that has eroded its market advantage.

From the challenger to the challenged, behind the change of the role of AWS is the change of the competitive environment of the cloud computing industry: from the competition of strategic layout speed to the contest of value-added value.

Cloud computing is not only a good business, but also "hard work".

Although the story of cloud computing is interesting, few companies pay attention to it as a stage of conceptual promotion, because it is a "money-burning" business.

According to Ulz Holls, senior vice president of Google's technology infrastructure, the cost of a single Google cloud computing region is $360m.

So, what kind of concept is this?

Just imagine, around 2010, there will be several domestic Internet companies worth more than 1 billion US dollars. At that time, the market capitalization of Tencent was only about 40 billion US dollars, and JD.com 's valuation was only about 10 billion US dollars. If you get ten such projects, almost one JD.com will be gone. Who can afford it?

But for cloud computing, there are no shortcuts other than this "asset accumulation".

Some people vividly compare "cloud computing" to houses. "cloud computing" manufacturers play the role of real estate developers or landlords (building houses), while users are enterprises that rent houses.

Dabbling in cloud computing, providing IaaS services is as basic as building a house. Providers need to build their own data centers and buy infrastructure such as server storage.

This is not over. Cloud vendors want to grow on a large scale and need an expanding cluster of servers and data centers to achieve:

1 set up data centers in different regions to supply the needs of customers in different regions (the data center closest to the location of the user can get faster access).

(2) set up multiple availability zones in the same region to cover the use of users in that region as much as possible.

As a result, cloud computing has become the "hard work" and "hard work" that big companies have the foundation and capital to do, and only Amazon, Google and Microsoft have entered in the early stage.

However, the different areas of business subjects make the attitudes of the three giants towards cloud computing different in the initial stage.

The expansion of Amazon's retail business itself requires a lot of computing, storage and network resources. There is a saying that Amazon's retail industry is seasonal, and when business is depressed, the platform chooses to rent out excess computer capacity.

Just like Alibaba double 11:00 every year, the traffic is concentrated, to buy a large number of servers to ensure the stability of the website. But other than double 11, these servers are idle, and renting them is to make full use of the company's cost.

This explanation may only be part of the truth, but through "eat more than one fish", Amazon has discovered the unity of the nature of e-commerce and cloud computing and become the most active promoter of IaaS.

According to Amazon's 2006 results: with AWS, we currently offer ten different Web services and have built a community of 240000 registered developers.

According to the 2008 financial report: we will invest heavily in Amazon Web Services, and we will firmly believe that it has a considerable scale and can achieve high return requirements.

According to the 2009 financial report, AWS continues to expand its global coverage to provide more services in the European Union and the new region of Northern California, and plans to start business in the Asia-Pacific region in 2010.

This is also cross-validated in the capital expenditure that companies use to buy fixed assets.

Amazon's capital expenditure has continued to rise since 2006, and it did not cut spending after the 2008 financial crisis, but the growth rate slowed down and increased spending sharply again in 2010.

Compared with Amazon, Google is an advertising company, after its listing in 2004, the profit margin of the business is very high, there is not too much pressure to survive, so many of the technical "drag" products and functions, "dazzling technology" is too big to make money.

(around 2006, Google's market capitalization was more than four times that of Amazon. )

Objectively lack of motivation to develop cloud computing, Google has not updated related products for a long time since the release of Google App Engine.

Microsoft, which was born as a software, initially cut into cloud computing with the PaaS layer. But without IaaS infrastructure, the development of Paas layer is limited. It can be seen that Microsoft's early capital expenditure is gradually shrinking.

However, this situation of fighting on its own has changed in 2010.

At the end of 2010, the US government announced that it would restructure the government's IT architecture through the integration of federal data centers and applications and the adoption of a "cloud computing first" policy. And require governments at all levels to transfer some of their work to commercial, personal and government cloud computing systems.

The official attention to cloud computing directly drives the collective warming of the cloud computing market in the United States.

Since 2010, Microsoft's capital expenditure has continued to rise, with a compound growth rate of 23%, while Amazon's capital expenditure has grown at a compound rate of 45%. Google is "catching up" and the scale of capital expenditure is also rising.

The huge capital investment has led to the continuous growth of the availability zone of cloud computing business. (other companies do not have continuous data disclosure, which is unable to reflect the trend for the time being. )

In absolute terms, Amazon AWS has a clear advantage in the number of availability zones with its first launch and continuous investment.

As shown in the figure below, Amazon has 77 availability zones by the end of 2020, ranking first in the industry.

However, the construction of the "house" is only the first step, and whether it can attract users to "rent" is the key to withstand the market test.

AWS destroys the "way of money", but breaks the "road of money" of others.

From the beginning of the e-commerce business, Amazon has a set of "proud" business logic: ultra-low price + huge volume + meagre profit, which has been performed repeatedly in Kindle, AWS, Echo and other expansionary businesses.

Under the riddle of "small profits, high sales and clever management", the core of "selling well at a low price" is hidden.

Cut prices and cut off other people's "money"

In economics, there is a "prisoner's dilemma" of a price war, that is, the cost of a commodity is 10 yuan. if it is priced at 15 yuan at the beginning, the market will soon wage a price war and push the price below the cost price, and no one will have any business.

What causes it?

Because the price is 15 yuan, which represents a high profit of 5 yuan, it will attract a group of people to do this business. There will be a rapid overcapacity in a short period of time, and then it will fall into a vicious price war.

However, if the cost remains the same, it is only sold for 10.50 yuan at the beginning, and the competitors take a long time to do this and earn 50 cents, so why bother?

Well aware of this way of "making a fortune", the price of AWS was deliberately set very low at the beginning.

Low prices (low profits) will only attract customers and will not be noticed by competitors. Wait for the scale of AWS to do, become Amazon's money-making "sharp weapon", competitors smell "profit" and move, but still be led by the nose by Amazon.

Rely on the low-price strategy in the field of IaaS services in the dominant AWS, followed by a "crazy" price reduction. As of May 2020, the price has been reduced for 82 times.

The coquettish operation that seems to destroy the "way of money" is actually breaking the "road of money" of others.

For AWS, the scale effect promotes the cost reduction, as long as the price reduction is less than the cost reduction, the more you sell, the more you earn.

It can be seen that the operating profit of AWS shows an upward trend as a whole after 2015.

But for competitors who are still in the input stage, the costs are high and they are forced to fight a price war, and the results can be imagined. (for more information, please refer to "12 price cuts a year." on how Amazon AWS "cut people's money"? "one article)

Amazon's low price "a fresh move" is not only reflected in the defense against the entry of competitors, but also used as a "spear" to provoke others.

Low price, move the traditional software vendor "cake"

AWS deliberately depressed profits, sure enough, won the "support" of users.

"all of a sudden, Amazon is everywhere," says Eric Schmidt, CEO of Google. The CEO of every startup seems to tell me that they are building their own systems on Amazon servers. "

However, as discussed above, when AWS was launched in 2006, there were only a few basic functions. With the limited "fool" mode blooming everywhere, users are "stupid people have a lot of money"?

In fact, it is Amazon that has a lot of money, but it is very shrewd.

Long-term fishing, AWS another round of "money-burning" mode is opened, the object is the expansion and upgrading of product categories.

Amazon's technology costs accounted for 12.7% of expenditure in 2019, up from 5% in 2010, according to the financial report.

It ranks first among the top 10 companies in global R & D spending for five years in a row (2014-2018), about $15 billion more than Google in second place.

According to Amazon Financial report: the expenditure in this area is mainly used for technology infrastructure expenditure, including AWS, as well as application research and development, category expansion and other construction of digital ecology.

With the money in place, the effect is "immediate". In 2017, the number of functional services on the AWS platform has increased to 1300.

Once again gather the "seven dragon balls", no, it is the cost advantage under the scale effect, AWS also summoned a "low price" offensive.

Take database software as an example:

Although the performance of traditional business software such as Microsoft SQL Server is stable, small and medium-sized enterprises are intimidated by the high purchase price and maintenance cost, while the choice of open-source database may be lack of technical support.

Amazon Aurora, a cloud database launched by AWS in 2015, takes into account the security, availability and reliability of commercial software at a cost of only 1 / 10 of that of a commercial database.

Of course, this is a common advantage of cloud database products.

Hitting the price pain point of users, cloud computing manufacturers have robbed a lot of "meat" from traditional software providers.

According to the Gartner report, cloud database providers such as AWS, Aliyun and Google Cloud have seen a significant increase in market share and are expected to reach 50% by 2021.

However, from a split point of view, the early full "bait" AWS cloud database, caught the most "meat" (market share).

According to the AWS2018 year, more than 64000 databases have been migrated from traditional commercial databases to AWS's cloud databases in the past two years.

As the market grows, those who "bow" to the low price of AWS seem to extend from start-ups, small and medium-sized enterprises, individual developers, to large enterprises or institutions.

In 2020, for example, Samsung moved more than 1.1 billion of its users from Oracle's database to Amazon Aurora. After migration, you can save 44% of your operating costs and other expenses per month.

But it is understood that Samsung's migration took 22 weeks, or about five months. Using such a high conversion cost to "go to the cloud" is not just for cost savings.

So what makes it easier for big companies to "bow"? We think it's the temptation to generate income.

For all enterprises, the motivation to improve revenue is ahead of the motivation to reduce costs when they decide to change an inherent structure.

Just like you are used to nails in your corporate office, if Flying Books come over, let's change our Flying Books, which is cheaper than nails, but this is far from the level of "tearful sale". Will you stir up a crowd for the discount of "three melons and two dates"?

But if Feishu says that using our Flying Book can help your company achieve 50% revenue growth every month, do you want to change it?

The basic IaaS business of cloud computing corresponds to motivation: reducing costs; high-level PaaS and SaaS services correspond to motivation: increasing revenue.

This makes the market confidence hit by AWS at a low price "stirs up waves".

Large-scale advantages and technical barriers are concentrated in IaaS business. But the needs of users have increased in value, how to transition from the good IaaS business to the short board SaaS level, AWS to a "race end point".

IaaS basic disk suffered "reverse sniper"

AWS's weakness in the SaaS business gives competitors an opportunity to start a reverse sniper attack that means "killing you while you are sick".

In recent years, the business revenue growth rate of Microsoft Cloud has been much higher than that of AWS;, while Google Cloud's revenue growth rate has also surpassed that of AWS after 2018Q4.

Under the leading growth rate of revenue, the market share of Microsoft Cloud IaaS business has increased significantly since 2017, while Google Cloud's market share has also increased slightly.

What they occupy is AWS's IaaS market: from 52 per cent in 2017 to 45 per cent in 2019.

Although such a decline in market share is far from impacting the basic market of AWS, how can others be allowed to sleep soundly on the side of the bed? What's more, the current situation is not inexplicable.

Here we introduce the model of Wechat to assist in understanding.

As we all know, Wechat itself is free to use and does not make money, but the giant traffic gathered by Wechat can be grafted onto a variety of realization channels, such as the rise of Pinduoduo, Mini Program value-added and so on.

IaaS services have similar logic.

IaaS itself can not make differentiation, there is no way to get high gross profit, but it is a good tool to get customers at a low price, gather customers on the platform, and then can be realized with PaaS, SaaS value-added products.

However, the truth is this, the reality is that AWS itself does not have any SaaS services.

SaaS services involve complete software development, and the software is independent of each other. If cloud manufacturers want to develop SaaS on a large scale, they need to go deep into the field of software development. This is certainly not a problem for Microsoft, but Amazon is clearly not good at it.

Instead of bothering itself, Amazon has to use its weaknesses to "bump into other people's strengths". Instead, it attracts a large number of SaaS developers to move in, each taking what they need.

For example, Amazon launched AWS Marketplace in 2012 to increase customers' eco-stickiness to AWS and earn some commission by introducing products from software companies such as IBM, Oracle and SAP.

However, it is not clear whether these "counterattack" actions can resist the offensive that Microsoft Cloud has pressed down from the cloud-SaaS layer.

In addition, AWS has another place to make up for-- service.

At present, the group of direct users of cloud services has expanded from technical personnel to non-technical positions. For those who do not understand technology, it will not impress anyone to use the powerful technology as an experience selling point.

The "engineer" culture is strong, and Google Cloud, which is pure "technology brain", is therefore disliked.

When the well-known company Workday chose a cloud computing service provider in 2016, it first considered Google Cloud, but in the end, it chose Amazon AWS because it thought that Google Cloud had a bad experience and did not consider the operation, economy and actual business needs of large enterprises at all.

If the product is to be competitive, it must be easy to learn and use. Only in this way can users who have a headache in learning new tools be persuaded to buy it.

The nature of the retailer, coupled with the simplicity of the IaaS service, presumably Amazon is not troubled by the service experience for the time being, but there is still room to explore how to promote the smooth landing of cloud computing services.

After all, the weakness of the SaaS layer, so that Amazon Cloud no matter how to make up for the shortcomings, are inevitably confronted with Microsoft Cloud head-on, facing users "choose one of the two" situation. And dig deeper in the service value that you are good at, it may increase the chances of the balance tilting.

Summary

The cloud computing industry has been surging for 15 years. AWS, as the forerunner of the cloud computing industry, has eaten a lot of market dividends by virtue of the first-mover advantage, but this advantage has gradually failed under the intensification of competition in the industry.

Especially in recent years, the competition of business value of entrants in the cloud computing industry has become more and more tight. Microsoft Cloud relies on the IaaS penetration of software layer advantage (SaaS layer), constantly squeezing the plate of Amazon cloud IaaS business.

Next, Amazon, which lacks the software product gene, wants to keep its customers sticky to AWS, in addition to trying to make up for its shortcomings, it may also need to do more to meet the needs of users.