With the growing IT of Party A, the dilemma and the way out faced by the enterprise service software industry

Author | Fu Xiaoyan

Proofreading | Luo Yanshan

In recent years, with the in-depth development of digital transformation, the enterprise server side of the software industry is "busy", after all, it can be regarded as the superposition of multiple cycles: localization cycle, digital cycle, spontaneous reconstruction cycle, infrastructure update cycle, etc., a variety of factors are promoting the large-scale renovation of software, and the looming "artificial intelligence" cycle has not been added here.

However, "busy" is busy, the income seems to be unsatisfactory, and even to some "dangerous" situation, in recent years, there is indeed no lack of Party B's list of bigger and more loss, so there are also industry experts for this grievances. How should we view this phenomenon? What will be the future direction of software companies that provide enterprise services? These questions are not easy to answer, after all, there are some things that people in the industry are helpless to do even if they know it, and the future is always full of uncertainties. I'll talk about this topic based on my own feelings.

Since these two professional services are now closely related to digital transformation, they will also provide digital transformation services, which can be regarded as an extension of service capabilities. The author has only been on his own for about a year and a half, and has worked in a large state-owned bank for 20 years, 12 years in the business line, and 8 years in the IT line. After leaving the bank, he worked in a leading IT consulting company, a cloud service team of a leading Internet enterprise, and a senior executive in an Internet small and medium-sized enterprise for a period of time. As far as Party A's work is concerned, the author has purchased from others and has been purchased by others, and I have some experience in both aspects, but the procurement experience is not deep.

Regarding the unsatisfactory operation of the enterprise server software industry, the author himself has dealt with some big Party B, so I have also heard a little, in this regard, the author has the following feelings:

The long-term effects of low prices

From the perspective of the phenomenon, in terms of price alone, the long-term "low price" is not a good thing. From the perspective of economics, "price" is also equivalent to the "thermometer" of the economy, and if people want to live comfortably, they must maintain around 36.6 degrees, not too far behind.

The enterprise is equivalent to the "people" in the economy, and the "price" is the temperature of the enterprise, of course, through the look, the "profit" represents the real temperature, but the "price" has a direct impact on the "profit", and the "price" is more "visible", it is the "object" of negotiation in the transaction, and the transaction is also the "price" that has been agreed, so the "price" is usually used to explain everything. If the "price" is too low, the "person" will keep "shivering" to increase the "body temperature", but the "shivering" consumes energy very quickly, and the company may "fight" and disappear.

In fact, the development of the Internet over the years, although the people have won a lot of benefits in consumption, but the long-term "low price" will also make the development of enterprises providing consumer services difficult, and this difficulty will be passed along the industrial chain, from the ToC enterprises to the ToB enterprises little by little, at this time only monopoly resources or industries with a certain monopoly will be better. In fact, economics believes that moderate price increases are most beneficial to economic development, just as people will be very active in summer, especially at the beginning of summer, the temperature is very comfortable, and it is easy to move. If the "price" is too high, it will be troublesome, on a hot day, more than 40 degrees, people may not want to do anything except blow the air conditioner, and at this time the economy is prone to problems and is not active.

The basic principles of economics are the same for enterprises, after all, theories come from thinking about the economic activities of enterprises. If the "price" is too low, the enterprise may "tremble" and "beat" itself to the loss; if the "price" is too high, the money will be hot, and the enterprise will not be able to get it if it wants to, and no one will come out for activities.

Excessive competition, pull down the price, will eventually smash the industry's job, and Party A will not really benefit from too low prices, the author himself has seen, Party B's price is too low, Party A also knows in his heart, at the beginning of the two sides also felt that Party A can be the cornerstone of the customer, to create Party B's products, from other customers to make a profit, the result of this single loss far beyond expectations, Party B "lying flat", Party A has no way, but also knows that Party B has a big loss, can not put forward higher requirements, in the end of course the project can not achieve the initial expected results. This kind of "cake" was there at the beginning, and some of them were eaten hard without "cake", and in the end, it was because of the "face" problem of both sides, so they would not openly "break up". Phenomena alone, this is not good for the industry.

Multivariate cause analysis

From the perspective of genesis, this problem is not easy to solve. There is no point in looking at the phenomenon alone, after all, if it can't be solved, it will be a waste of time to talk about it, and you can only ask for more blessings. There are many reasons for this situation, and it is definitely not caused by a single factor, after all, some Party B in the industry is still in a position with a certain "right to speak", so the "difficulties" of the enterprise server software industry are not generalized, and have a certain relationship with the competition in the subdivided field. I will only mention a few factors that I have observed:

(1) The price is highly transparent

This is the inevitable result and benefit of market competition, especially for Party A, but for Party A, which is good at procurement, this is not only a benefit, it can reach the level of "killer", for Party A, which carries out IT procurement business all year round and has many types of procurement, the price of the whole market is completely clear, and Party B's actual ability level and product performance are almost "clear" For this kind of large Party A (or even smaller), Party B has almost no bargaining power, if you want to maintain the price, you have to risk the risk of losing the bid to "resist", but there are very few that can "carry".

For example, Party A is a very large procurement, and needs to maintain multiple Party B winning bids to avoid the implementation of risks, in this case, sometimes the top Party B in the industry is not shortlisted, Party A itself may not be easy to explain, so Party B has a bit of room to "carry" the price. However, if it is switched to a single winning environment, winning the bid at a high price requires a lot of explanation work, and Party A is very clear about the historical price and the current price, and Party B rarely has any room to "carry" the price, and Party A will not be willing to find this kind of "trouble" for itself.

The formation time of this situation is not short, even if the management requirements for winning the bid at the lowest price are canceled now, it is difficult for Party B to "premium", because the source of "premium" is not the bidding management, it is the "brand" price, "product" price, many domestic Party B in the competition with foreign capital for more than ten or twenty years, can only rely on lowering the price and foreign brands, products competition, even if the competitive products are now withdrawn from the market, it is not a "premium" Return it to Party B in China, because in Party A's memory, Party B is this price, and it is very difficult or even futile to change the impression of others.

（二）双方 IT 力量膨胀

If only a certain character can do something, this character has the capital to "bargain". Ten or twenty years ago, in the IT industry, scientific and technological personnel were still "scarce" resources, even Internet companies, in the early stage of development, there were many stories of programmers who were halfway out of home to get rich on the Internet Express. At that time, it was unimaginable to say that an enterprise had an IT team of 1,000 people, after all, the author's banking industry, even if it was a large state-owned bank with a lot of IT capacity, more than ten years ago, it was also an IT team with a scale of 1,000 people, but about half of it was not engaged in development, and Party A's "abundant" IT personnel were demand analysis and project managers.

Nowadays, the head city commercial banks, which are equivalent to the provincial branches of large state-owned banks (this is in terms of business area, in terms of personnel and scale, they are larger than the provincial branches of large state-owned banks), have IT teams of thousands of people, while large state-owned banks are in the tens of thousands, and the number of people alone may be more than that of Party B enterprises serving them. Some state-owned enterprises have set up digital technology companies in the process of digital transformation, and the size of the staff ranges from a few hundred to nearly a thousand. In recent years, with the development of new energy vehicles, the IT team of automobile companies has also expanded rapidly.

In fact, this is the fault of the software industry itself, and we promote the concepts of "software defines everything", "digital enterprise is software enterprise", "all business digitization, all digital business", "digital transformation is life and death", etc., Party A accepts, but the reaction after accepting it is not more dependent on Party B, but more trying to strengthen its own capabilities. After all, if you feel that one thing is related to future competition, or even life and death, then you must grasp it in your own hands, even if it is a part, otherwise you will not feel at ease.

It is an indisputable and irreversible fact that the number of IT practitioners continues to increase, the strength of Party A continues to expand, and the importance of software continues to increase, and there may be millions of new jobs every year trying to enter the IT field. As I often say in my enterprise architecture class, if Party A gathers so many IT forces and doesn't let him do something on his own, and buys products and suites all day long, does that mean that he wants to disband the IT team again? Do you think that Party A's IT team has no learning ability at all, and the system at hand will never be able to do it?

(3) Passive improvement of Party A's enterprise structure capabilities

Due to the increase in systems, IT personnel, and business complexity, no matter how capable Party B is, Party A has to think about its own internal IT governance. In the past, if the enterprise architecture capability was not available to Party A, it was also dispensable, and there were not so many systems and not so many people, so they were all handed over to the outside to be "idle".

But now it's very different, the system's "family base" is getting thicker and thicker, and the more and more chaotic, Party B is not able to help Party A take care of such a big "big house", Party A can't take care of it by itself, and it's always the kind of planning that is handed over to the outside, and the right and wrong are not clear about the dry method has long been unable to play. Party A's enterprise is now passively improving its enterprise architecture capabilities, or IT governance capabilities based on the overall architecture, and even the State-owned Assets Supervision and Administration Commission (SASAC) has put forward this requirement to central enterprises, which is a good thing for the author, after all, it is not easy for this "cold kang" to be "hot".

Now is not the time to ask whether the enterprise architecture is good or worthwhile, but to draw a "map" for the enterprise to draw a "map" for its own home no matter what method it takes, otherwise IT governance will become a pain point for the enterprise, which will not only solve the business problem badly, but also add technical problems to the enterprise. What's even more "terrible" is that this overall architecture capability itself is a relatively advanced "overall design ability" in the field of software research and development, although it is not said that there can be had, but it must be said that if there is, if Party A's "overall design ability" rises, Party B's "right to speak" can only further decline.

Judging from the current situation, this phenomenon has not only appeared in some industries, but also will spread in more industries with the advancement of large-scale transformation work, which is only a matter of time.

(4) Slow changes in software R&D models

In the field of software design, there has always been a dream of "building blocks", but in fact, implicit in the dream of "building blocks" is the rapid industrial production of software. Regarding the production of the software industry, the author has always had a joking way to describe it, "large-scale small team workshop", I think it is quite vivid, no matter how large the enterprise, no matter how large the IT team, the development is also cut into small pieces, by the small team to do the way of manual code, has not been able to "industrialize", "agile" based on overtime, not based on methods, both parties A and B are like this, up to now, there should be no real "agile" enterprises.

Although the master who was still in the era of shortage of people in the last century said that the software industry is not an industry that can improve efficiency by simply "piling up people" ("The Myth of the Human Month"), there is no better way to improve our overall productivity than adding people. However, theoretical research has been developing, and it has also converged, just like in the field of enterprise architecture, the mainstream methodology has pointed to its own research on the definition of "business capability", and the software design for "business capability" is decentralized and independent to define "business capability", which means that the software should continue to move towards "parts" and "building blocks", the direction is correct, and the method has begun to support (in fact, the design idea of separation of process and capability has been around for twenty or thirty years, and the author writes "Aggregate Architecture" The book has been slightly archaeological), and more practice remains.

As a practitioner in the field of business architecture, I am also coming into contact with more and more people who want to redo capability-oriented and abstraction-oriented architecture design, and this demand also includes some enterprises in the Internet field. After all, only when the design is "parts-based" and "building block-based" can there be the possibility of assembled production and the possibility of approaching industrial efficiency. Many enterprises are trying to "capable" business design, and the "middle platform" methodology, which was once popular for half the sky, is also in this category, and the complete enterprise architecture methodology basically points to this direction.

As the enterprise "passively" improves the enterprise architecture capability (not only recognizes, but can execute), the enterprise's own "business capability" identification, definition, and design level will gradually increase, which will further conflict with Party B's "ideal" output model in terms of software requirements. At present, the change of this model is still relatively slow (after all, the improvement of Party A's capabilities is still in a small number of enterprises in general), but the overall direction is still clear, and the assembly enterprise application based on capabilities is what everyone wants to do, and the assembly should generally be based on "parts" rather than "kits".

Because of the slowness, so Party B may be more "painful", at this stage, the new model is not completely out, the old model is not satisfactory, the customization is getting stronger and stronger (the current IT ability is not so strong Party A also has a tendency to customize), the project cost is becoming more and more difficult to control, maybe it is better to have a "happy".

These causes may not be the root causes, but in the author's opinion, they are basically irreversible factors.

Party B can consider making changes

After all, maybe 5-8 years later, there may be bigger "variables" in the software industry, and if you don't "adapt" first, the "variables" may be more difficult to accept.

The author believes that the biggest adaptation that Party B should do should be the change in the form of software products, change the output of "kit" to a diversified output based on "parts", and if there is an opportunity to do it, then do a thorough "partization" of the existing kit to complete the design of separation of process and ability. The author's experience of "productization" is still shallow, from my own point of view, the most difficult to adapt to the "suite" is the interface, process, roles, permissions, which have little to do with the "real knowledge" (rules) actually contained in the "suite", the reader may think that the "process" is not "real knowledge"? "process" is, but to be precise, it is "unstable true knowledge", "not universal true knowledge", the "process" of different enterprises will be very different due to the difference in organizational size and structure, "process" In contrast, the applicability of "rules" will be broader, and there will also be opportunities to do more comprehensive abstractions and become "productized" capabilities.

Some Party B companies have also asked about the way to do product abstraction, and the author thinks about it based on the experience of business architecture, "object-rule-process-interface", these four layers of abstraction order should be a reference order for doing "productization" abstraction.

"Object" is a logical-level data entity, that is, a business processing object, first abstract the business processing object, this layer of design many enterprises actually consider the abstraction of the library table, not the entity abstraction, which needs to be changed here; then is the "rule" of processing the "object", the rules can usually be "parameterized" and "integrated"; then is the "process", but it is generally combined with the process engine to make some changes, in fact, many times in the integration of the process to consider changing the process, so, the "process" is very unstable, abstract can basically only be connected The process is "lazy" (or it may be that there is no support from business personnel) to think of Party A; the last is the "interface", the most unsuccessful abstraction is often the interface, who can make everyone love? Party A has a unified portal for its own.

From this point of view, if Party B can conform to the current trend of "slow" changes in Party A's product adjustment, and focus on smaller "parts" in professional capabilities, it may have more opportunities to improve "professionalism", after all, the so-called professionalism of Party B is to make greater efforts than Party A's professionalism, after all, Party A is too lazy to summarize knowledge, but also to do business every day, and the business is still changing from time to time It's not easy, after all, it's not the source of business, and business innovation won't happen to Party B.

If the "parts" can be done well, then the output of the large Party A should be based on the parts, here the author especially wants to say that the large Party A should do its own "process", for the large Party A, sooner or later the architecture ability will be in their own hands, their own development resources should at least cover the development of the "process", after all, there are a lot of development resources, Party A does the "process" not only because it has strong enterprise localization characteristics, but also from the perspective of enterprise architecture, the "process" is a horizontal application perspective, that is, the call of "capability", The tandem perspective is a strong business perspective, but also the performance of business innovation, the change of its own process, the rearrangement of the call is the way of business innovation, so this part should be done by Party A's own enterprise, Party B provides a stronger "rule" type "parts", maybe the purchase amount has decreased, but Party B's cost has also decreased, and it may be more professional in the professional place. However, there are many "rules" that Party A may do by itself, and there is no way to do this. For Party A who does not even have enough development resources for the "process", this is the object of selling "suites".

This will not lead to the downsizing of Party B's enterprise, not necessarily, because the "parts" do not necessarily have to be cheap, after all, they are "professional-grade products" (if Party B builds them professionally enough), but this will also "expose" where the software is not so "expensive" in the first place. However, if the development trend is that the scale of Party B is reduced in the end, then if everything that should be done is done, there is nothing to regret, after all, the goal of the development of the software industry has been instilled in our customers, "digital enterprises are software enterprises".

It is difficult to say whether the substantial increase in the production capacity of enterprise server-side software must be accompanied by the large-scale development of specific software enterprises, perhaps this is the difference between the software industry and the manufacturing industry, large-scale software production does not depend on a specific "production equipment" supplier, because as far as the industry is concerned, the basic implementation required for software production is already "distributed" in various industries, only the "distributed" capacity is missing, and this "distributed" is already "in progress".

The future of artificial intelligence

In March of this year, the first AI engineer Devin was born, and then there were several open source versions, enhanced versions, and before Devin could open source himself, other versions were already open source. I remember that last year, many people still thought that AI programming was not very reliable, but in a "doubt", this year AI engineers have improved again. The enhancement of individual capabilities brought about by the development of this round of large language models is very concerning, and it is also the direction of development that the author believes.

If you don't talk about anything else, give it another five years, how many people dare to bet that AI is still "garbage" in terms of programming? Five years, is it not possible to improve the efficiency of individual programming by 50% (the author has not dared to look at it according to several times of improvement), so with the current number of IT personnel of Party A, can Party A handle the main implementation and reconstruction of all business systems by itself? At least there is no productivity problem, and design ability, knowledge extraction ability, The project management ability, as long as the enterprise really attaches importance to it and is willing to work hard, can also be greatly improved, what should Party B do at that time?

We don't have to be "anxious" about the future of AI, and this trend alone is enough for Party B companies to think about. Going back to the present, if you need help with the current difficulties, especially the government, then maybe you can consider some subsidies, business policy adjustments, and so on. However, the core competitiveness in terms of products, software models, development theories, core technologies, etc., still has to rely on the enterprises themselves to solve, outsiders can only help the environment, "adaptation" has to rely on themselves, "protection" can not improve the ability. There is a famous sentence in "Jurassic Park": "Life will find its own way out", and the author will continue to say "as long as the vitality is strong enough".

About the Author:

Fu Xiaoyan is the author of three books: Enterprise Business Architecture Design: Methodology and Practice, Digital Transformation of Banks, and Converged Architecture: Componentized Enterprise Architecture for Digital Ecology. Executive Director and General Manager of Beijing Tianrun Juliang Consulting Services Co., Ltd., Senior Vice President of Digital Twin Technology Co., Ltd., Deputy Director of the Expert Committee of Digital Transformation and Artificial Intelligence Industry Talent Base in Key Areas of Industry and Information Technology, Member of the Software Engineering Special Committee of China Computer Federation, Member of the First Executive Committee of Digital Finance Branch, Enterprise Architecture Promotion Center of CAICT, Technical expert of the Prefabricated Promotion Center, distinguished expert of the Expert Committee of the All-China Digital Talent Cultivation Alliance, professor of industry-education integration industry practice of the SME Development Promotion Center of the Ministry of Industry and Information Technology, researcher of the Financial Big Data Application and Security Research Center of the National Engineering Laboratory, member of the CIC Expert Committee on Financial Technology and Digital Economy Development, and vice chairman of the Expert Committee of the National Internet Data Center Industrial Technology Innovation Strategic Alliance.

Original link: Party A's IT is growing day by day, and the dilemma and way out faced by the enterprise service software industry_Digital Transformation_Fu Xiaoyan_InfoQ Selected Articles