Shulou(Shulou.com)06/03 Report--

Data is not only a technical problem, but also a key factor related to the life and death of enterprise digital transformation. the primary problem of digital transformation is the problem of data strategy, so how to solve the problem of data strategy?

1 the essence of digital transformation is that information technology drives business change.

What is digital transformation?

How should digital transformation be defined?

A thousand people may have a thousand answers to the above questions. As far as I can see, these answers cover different areas of enterprise management, such as strategy, management, technology, business, user center and so on, from industrial 4.0, consumer Internet to industrial Internet, each has its own description and explanation.

So, after so many years of transformation practice, we look back, what is the most important driving factor?

Let's go back to the beginning of the industrial age.

Despite the help of steam engines, human physical and mental strength is still an indispensable factor in industrial production. After so many years of development, machines have increasingly replaced human physical strength, but brainpower is still the most important driving force in the last century, and even the original computers are just strengthening human brainpower rather than replacing it.

It wasn't until the last decade that things changed.

In the past 10 years, the development of cloud computing, big data and artificial intelligence technology has continuously driven machine intelligence to replace physical and mental power, and those high-frequency repetitive physical and mental tasks are increasingly replaced by machine intelligence; and the empirical formulas on which brainpower depends (methodology, thinking tools, etc.) are also increasingly replaced by algorithms. In this sense, computing is the real driving force of this era, and its replacement for physical and mental strength is one of the greatest challenges facing mankind in the past decade, with past operations and business models constantly being subverted.

Therefore, from today's point of view, we may be able to clearly define digital transformation, which in essence is the information technology represented by computing power, algorithms and data, which drives the change and change of society as a whole in a digital way. The future of digital transformation clearly points to the development of artificial intelligence, which allows people to do what they should do, machines to do what machines should do, and the coordinated development of human and machine brainpower.

The challenges facing mankind do not stop there.

A book called abundance once pointed out that great changes have taken place in the organizational model and business model of human society when the supply of materials exceeds demand to the state of excess of supply. Lack of food and clothing has been the theme of human society for thousands of years, but today, when these are no longer problems, people are confused, what on earth do we need to pursue?

If we look at the data from the same perspective, what will we find?

The following picture comes from the report of the Queen of the Internet, from which we can see that the amount of data in human society has increased greatly since 2010. What the original picture does not tell us is that 2015 is a key point in the growth of data in human society, because in 2015, the amount of data generated in one year is the sum of the amount of data generated in human history. In other words, the amount of human data has grown exponentially since then, growing by 40% Mur50% a year after 2015, also known as the "new Moore's law": the total amount of data in human history doubles every 18 months.

If you remember the story of the king and the accountant in the Arab fairy tale about filling the chessboard with rice, you will know how huge the data increment of this exponential growth is. Human society may be about to change from the supply of data in short supply to the oversupply of data. From here, we may understand why Mr. Ma said that all companies in the future will be "data companies".

Alibaba was once a company that experienced this kind of data growth. for a long time, the contradiction between the increasing cost of data storage and the still scarce data applications has been the main contradiction of Alibaba. Why did Alibaba "go to IOE"? It is said that Wang Jian calculated an account for teacher Ma that if the IOE architecture is still used to store data, then, ten years later, Alibaba's data storage costs will be 10 times his income, when Alibaba will go bankrupt.

At present, most companies are still in the data addition stage, but in the car companies, we have seen signs of data growth by leaps and bounds. This sign is called "car networking".

With the motorization, intelligence and networking of car companies, car companies have also collected a huge amount of data, 4G a day, 250 million lines a year, and at least more than 200T three years later. Descriptions like this can be heard in most car companies. Perhaps, the data growth of car companies can not be called exponential growth, but it is at least a multiplier growth than before.

So, what do we do?

As a company that has surpassed the exponential growth of data, Alibaba's experience may give us inspiration.

2 how does Alibaba surmount the singularity of data exponential growth

In 2007, Alibaba decided at a strategy meeting that Ali would become a data company in the future. But really make some achievements in the data, but to postpone this time to 2009, Ali Yun was also born in this year. In general, Ali's whole leapfrogging process can be divided into three stages:

The first stage: 2009-2012, the theme is "see".

Since its establishment in 2003, Taobao has collected a large amount of data, 90% of which are unstructured log data. when we have these data, everyone wants to see the truth behind the data: where my users come from, what they bought, why they bought it, and what the conversion rate is. These questions generally boil down to two basic questions: what happened? How did it happen?

What happens at the same time as "going to IOE" is that Ali increases the demand for BI (business intelligence). BI, which "sees" the answer with data, is the main force of Ali storage and computing resource consumption at this stage. Alibaba was also the first company to set up CDO (Chief Data Officer, Chief data Officer). The first CDO was later Alibaba CEO Lu Zhaoxi. Interestingly, the data platform department established later was habitually called CDO. By the way, Lao Lu is also an angel investor in Singularity Cloud.

The second stage: 2012-2015, the theme is "use".

A landmark event was the establishment of the data platform Department in 2012, a department called CDO, which originated from the data platform team formed by Qigong, which gave birth to a series of data analysis and mining tools, including in the cloud, data Rubik's cube, Taobao time machine, Taobao Index, TCIF and so on. In particular, I would like to mention TCIF (Taobao Consumer data Factory), founded by the founder of Singularity Cloud, which has pulled through all Alibaba's consumer data and completed the construction of a 3000 + tag system. Every student who uses Alimama for precision marketing should have seen these tags, that is, the checked options in the Dama plate.

The landmark event in 2012 was that the storage and computing consumption of TCIF exceeded that of BI, and the population represented by TCIF was directed to become a major consumer of computing resources. Another landmark indicator is that 50% of Alibaba's servers no longer handle any transactions, but are only used to process data.

At this stage, Alibaba began to realize the question of using data to predict the future, and better help the business to answer: why did it happen? What will happen in the future?

The third stage: from 2015 to the present, the theme is "enabling".

Similarly, there were also two landmark events in 2015: first, the establishment of the Aliyun digital plus platform (in the process of creation), which means that Alibaba began to externalize the internally formed big data capabilities to empower the society to build big data capabilities; the second is to promote the thousand-person thousand-face algorithm, the recommendation algorithm has become the number one consumer of storage and computing resources.

The recommendation algorithm is not only as simple as the Taobao interface we saw. To some extent, the recommendation algorithm allows Alibaba to cross the "singularity from human command machine to machine command human". Today, more than 75% of Alibaba's GMV is run by machines, and traffic is accurately distributed by machines. In contrast, Tmall Taobao and others all together have only a few thousand operators, and the human efficiency is frighteningly high.

After these three stages, we can think that Alibaba has completed the construction of the data industrial production chain around the data, and established a rich data ecology around the data chain. In contrast, too many companies are still in the stage of manual work of data, which has a very bad impact on the digital transformation of enterprises.

Including the initial stage of connecting buyers and sellers, connect-see-use-empower four stages, allowing Alibaba to successfully cross the singularity of exponential data growth.

PS: for specific technology development stages, please see "four stages of Alibaba's Business Development in Taiwan" and "four stages of Alibaba data Development in Taiwan" in practice.

3 active and enterprising digital transformation of automobile enterprises

Apart from Internet companies, the automotive industry is one of the most active in embracing the Internet. Especially with the rapid development of vehicle networking, electric vehicles and new energy vehicles in recent years, automobile companies have also invested a lot of money and energy on IT infrastructure and advanced technology research and development.

In general, in addition to the manufacturing side (this is another big topic, more industrial 4.0), these attempts can be divided into

01 Technical reconfiguration centered on cars and services

For example, the future automobile strategy guided by electrification, networking and intelligence, as well as building omni-directional service capabilities around digital travel, and so on. According to PricewaterhouseCoopers Sliot's estimates, the profit share of traditional industries such as supplier business, vehicle sales and aftermarket will fall from 71 per cent to 41 per cent by 2030, and car companies will need to shift to travel service providers.

02 shift from car-centered to people-centered

Car companies have traditionally built their organizational structure around the value chain of cars, that is, the semicircle on the right side of the picture, and assign people's value chain to sales or marketing. But today, more and more car companies have found that the human value chain should be in the same position as the car value chain. New car-building forces like Weilai, and even its organizational structure are built entirely around users. From the core scene, car companies need to get through the human and vehicle data, build the life cycle management capability of people and vehicles, and break through the performance bottleneck through data capitalization and business intelligence.

03 consumers' massive personalized demand forces car companies to promote the four modernizations

Since 2010, with the development of e-commerce, social network and mobile Internet, personalized and diversified consumer demand has emerged. For example, some consumers want to see the whole logistics process of buying a car like Taobao shopping, and some consumers want to customize personalized colors. However, these requirements can not be met only by relying on the traditional information management system and architecture of the mainframe factory. When will the bus arrive at the store? When can I pick up the goods? None of these simple information can be queried online, let alone provide complex personalized services. Car companies have long been aware of this, and started the process of cloud and service to achieve large-scale and fine matching of people and cars, but the results are difficult to be satisfied, and the reasons will be explained later.

More and more car companies realize that those repetitive mechanical labor will certainly be replaced by artificial intelligence in the future, in which there are not only transactional and labor jobs, but even a lot of knowledge jobs, which are also likely to be replaced by machines. Then, next, car companies will face severe challenges, the need to reshape artificial intelligence-empowered employees and teams, and, the core competitiveness and attractiveness of car companies in the future will also come from their ability to support employees, predict new needs and promote job functions through artificial intelligence.

(4) the neglected data may become the Achilles heel of the digital transformation of automobile companies.

As far as possible, many attempts have been made in technology and business, and even traditional consulting companies have been found to advise on digital transformation, but car companies have encountered more and more digital problems.

The widespread use of new technologies not only partially solves the old problems, but also brings more new questions: why is half of the data I collect null? I have so much data, how can I cash it? How to generate value to the business? Even with the new system, why make decisions in a black box?

When doing consulting, I will immediately observe, or directly ask, is there a core project for the company's digital transformation? (that is, the so-called "Project one") is the top leader of the company or department directly in charge? And, does the technical planning effectively solve the existing problems?

Usually, the answer is relatively clear on the importance of digital transformation, but the overall project planning and the contribution of the project to the business will become blurred. And, there is a common expression, "Digital transformation is the number one project of the company." however, further down, the actual definition of "digital transformation" varies greatly, some build databases, some build platforms, and some carry out innovative projects. The only commonness is probably "testing the water first".

The advantage of testing the water is that it exposes the problem; the downside is that the experience learned from the test often does not help to solve the problem.

To sum up, at present, there are three main problems exposed by car companies in testing the water:

✨ 1. Digital transformation lacks top-level planning and design, and each department carries out the transformation work based on their own understanding of digital transformation.

✨ 2 business problems are often reflected in data problems, such as poor data quality, inconsistent data statistical caliber, data inconsistency, and so on.

The problem of the coordination between ✨ 3.IT and the business, and the needs expressed by the business are usually not the real business demands and pain points, but are more reflected in "I want to have the functions that others have".

Let's start with the first question: top-level design.

First of all, it needs to be clear that all technologies and projects serve to achieve the company's strategic goals, but the company's strategic resources are limited, and it is necessary to maximize the use of existing resources and technology under the premise of reasonable planning. then there is the problem of project management, which is the original intention of the top-level design.

Therefore, all top-level design problems need to go back to the source "what kind of customer provides what value", from the perspective of business value to the priority of the project, and then consider the issue of project management.

Secondly, it is generally believed that the quality of the project is determined by money, time and scope, which also constitutes an impossible triangle of project management: if you want to spend less money and cover a large number of business requirements in a short time, then the project quality will not be good. Either spend more money, spend more time, or identify a small number of business requirements, only in this way can digital transformation projects be transformed into business results with high quality.

Let's talk about the second question: the problem of data.

This is a problem that is often talked about by managers and employees, but not promoted to a strategic level by enterprises.

According to the SKOTT evaluation method, the digital transformation of enterprises needs to pay attention to the five dimensions of strategy, KPI, organization, digital technology and digital talents in order to reduce and avoid the risk of transformation. We very much recognize and quantify this (quantitative method is described later), and according to the characteristics of information technology, digital technology is divided into three dimensions: computing power, data and algorithm, to evaluate the current situation of digital transformation of enterprises.

The result of the evaluation can be described in a word: the lack of a clear data strategy has dragged down the digital transformation of car companies. In other words, at the present stage, the digital transformation of automobile enterprises must first solve the problem of data strategy.

Usually car companies are relatively ahead in strategic deployment and computing deployment, but, unfortunately, if all seven dimensions are pulled into progress bars, then the progress of the data is the slowest. Not only that, usually car companies in KPI, organizations and talent are also relatively poor performance, and these are more or less related to the progress of the data.

For example, due to the lack of data strategy, companies are usually slow to set up data-related organizations, and lack a corresponding compensation system and incentive mechanism for the recruitment and training of relevant talents needed by data and algorithms. it is not clear what talents and roles are lacking at present. By the way, according to our consulting experience, the most missing roles in enterprises today are business architect, product manager, and data analyst.

Because of the lack of data access and data quality problems, it is also difficult for enterprises to define the digital transformation in terms of KPI indicators. For example, digital business accounts for a proportion of profits, so which businesses are digital business profits and which are not? What is the reasonable proportion? This also leads to the lack of effective KPI to guide the transformation of the company, which can only be defined by each department.

Taking Ali as the benchmark, we also evaluate the current situation of digital transformation of automobile enterprises from the perspective of "saving and general purpose" of the data. as can be seen from the following figure, the data has not yet been centralized, which is a common phenomenon of automobile companies now. Data impassability and data quality problems further lead to car companies have algorithms but no results, data collection but no quality, arithmetic deployment but no data realization.

From this assessment, car companies are like looking at the Greek hero Achilles, who has been armed to the teeth by IT, but because they do not pay enough attention to the heel of data, they are often unable to move once they are shot by the flowing arrows of business.

So we often say that if you don't think about the data, the data will think about you.

Finally, let's talk about the third question: the problem of business cooperation.

This is actually a very big proposition.

First of all, this is an organizational problem. Traditionally, as a foreground department, there is a set of traditional process for business as a foreground department to divide and cooperate with IT, but today, this process is increasingly unable to meet the needs of enterprise digital transformation, and needs to break the system, including project budget, scheme design, KPI and so on.

Second, it is also a business problem. Traditional foreground business usually regards IT as the provider of systems and tools. IT does not need to understand foreground business. IT is only a guarantee role in the division of business. When enterprises talk about business value, they usually talk about front desk work, and rarely mention the value of IT. However, today, when more and more technical ideas lead the company's business, when the business needs technology delivery, IT not only needs to solve technical problems, but also needs to understand the business and be able to solve business problems with the receptionist. This requires not only cross-border talents, but also system design.

Finally, this is, of course, a technical problem. Let IT do business, let business do technology, these are unrealistic solutions. Then, IT needs to be able to face the future and provide technical support for the business. The trouble is that today's technical guarantee is not only as simple as system construction and IT deployment, but also includes data and algorithm solution planning, offline data collection capacity building, data application capacity building, and a series of construction parts beyond the traditional IT definition. There are also some projects such as tagging that require business to participate in and precipitate together, as well as cross-business projects such as connecting the front and rear supply chain data. These projects even require IT to plan from a technical point of view, and IT will drive the whole project cooperation.

5 data, not only technical problems

In 2015, MIT Sloan Management Review, together with Deloitte, conducted a global survey of digital executives, and the findings became the title of a later report, "Strategy, not technology, is driving the digital transformation." Its research shows: "the ability to digitally structure enterprises depends in large part on a clear digital strategy, one is a culture-supported strategy nurtured by leaders that can change and innovate." "

But as far as the research itself is concerned, it only stays in the application of digital technology, and there is no mention of data. Perhaps, its research objects regard data as part of digital technology, or regard data as materials for customer-oriented analysis.

The neglect of data is also the reason why big data is often caught in a dilemma today.

For most enterprises, there is no urgent demand for massive data processing, business is good in the incremental era, and there is no need to use data mining to assist fine operation, data decision-making and so on. However, when the market enters the stock fighting era, each company needs to seize the stock market from its competitors while keeping its own share, and the demand for data quality and data mining suddenly erupts. Everyone wants to know what the business problem is. How to promote the continued growth of performance?

In addition, the development of artificial intelligence in recent years has also attracted the attention of people of insight. The questions we often hear include: if all high-frequency repetitive mechanical tasks will be replaced by machines, and if all transactional and labor-oriented jobs will be replaced by artificial intelligence, what form will the enterprise be? What is the most important strategic asset of an enterprise in a society dominated by artificial intelligence?

Further split the acquisition of computing power, data and algorithms, we can find that computing power and algorithms can be obtained through public supply, such as cloud, or open source algorithms, but the data is difficult to obtain from the market. Moreover, as the country and the public attach importance to personal privacy and personal information, it will be more and more difficult to obtain data through purchase before there is no mature data exchange market.

To put it simply, data is the only factor that needs to be accumulated by the enterprise itself. however, similar to gasoline, once there is no refinery to process crude oil to turn the data into usable gasoline, and there is no gas station to refuel the car, then, the data is just crude oil lying in the data warehouse, which is useless.

Car companies should have a deep understanding of this, especially with the promotion of car networking and electric vehicles, car companies have deposited a large number of related data, at least a few G growth every day, but does not produce any business value. "what can I do with this data? "this is the most frequently asked question in the course of research.

And, as described above, the processing and consumption of data has never been just a technical issue, but an integrated strategic issue in the areas of business value, technology planning and organizational security. In order to solve the technical problems, we also need to start from these three dimensions.

6 how to promote enterprise data strategy from the perspective of data asset management

According to my research results, there are three main ways to promote enterprise data strategy:

01 overall assessment, organization comes first

This is an extremely rare form of promotion, and it is only reflected in a few companies such as Alibaba and Huawei. In these companies, with the strategic adjustment of the company, organizational adjustment will be carried out first. For example, after teacher Ma put forward the "new retail", Alibaba will quickly add the "HR- new retail line" at the human resources level to unify the planning and arrangement of human resources. How terrible is this? To borrow a teacher's point of view. He said that the traditional theory is that "the ship is hard to turn around." However, the current situation is that giant enterprises like Alibaba can quickly turn around in the face of market changes because of their excellent infrastructure, while small enterprises cannot make a U-turn because of lack of data and effective planning. This subverts the traditional management theory.

02 business first, take small steps and run quickly

According to empirical data, 90% of companies, 95% of companies will choose this way to promote the digital transformation of enterprises. To put it simply, the IT department will start with the most urgent needs of the business, or the most productive needs, and use digital technology to promote the business to "test the water". If there are results, the project can be turned into a star project, persuading executives and other business units to continue to adopt digital technology; if it fails, then continue to look for the next star demand. Mainly through small steps to test the water to accumulate data, accumulate technical capabilities, and to provide judgment for the next step of technical planning.

03 advance in technology and stride forward

This kind of enterprises usually have relatively clear requirements for the construction of data platform (platform) and data asset management system, although it may not be clear about the important position of data in the enterprise. and how to plan and build the data platform and data asset management system, but after preliminary research and discussion, clear the construction needs. From my understanding, these enterprises all hope to use the data platform (although the meaning of the word is ambiguous) to realize the great-leap-forward development of enterprise digital transformation. For this kind of enterprises, the demand for digital transformation consultation is also the most urgent.

No matter which way it is, we believe that its inherent development logic will be similar to Alibaba's experience, first "connecting" enterprises and customers online through low-cost business, and then "seeing" performance and customers through data online. Then build the data "use" ability to predict the future, and finally use data intelligence to "empower" the business into a service company or platform company.

Relatively speaking, we prefer the "technology first" approach, because no matter which way we choose, we will need a data center and a data asset management system to manage the enterprise's data assets like human assets.

Of course, you can also seek big data consultation with data asset management as the core.

Big data Consulting is the definition of service provided by Singularity Cloud in response to market demand.

As can be seen from the above picture, traditional consulting is usually strategic consulting and planning for one or two areas, such as management consulting, financial consulting, brand consulting and human resources consulting, and its methodology is based on experience summarization. through cross-level and cross-departmental research, to provide third-party neutral data and planning suggestions for enterprises to facilitate high-level decision-making. However, these experiences do not include data asset management experience, so companies with corresponding experience are needed to provide consulting services. This is also the original intention of big data consultation.

On the other hand, the special feature of big data Consulting is that in addition to commercial and organizational factors, IT and data need to be taken into account, and suggestions for solutions should be provided from the perspective of capacity building, that is, not only to solve problems on the demand side, but also to provide capabilities on the solving side, which also means that big data Consulting needs end-to-end (demand) to end (solution) capabilities. This is also the difference between big data consultation and other consultation methods.

Ps: more big data consultation content and detailed data intelligence solutions are the "big data Consulting White Paper" exclusively released on the "Digital Intelligence Business Forum" of Yunqi Conference on September 25. Please look forward to it.

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.