来源:新立场NewPosition
Internet companies are always keen to invent "black words", using abstract language to package an ordinary and even barren concept.
This was the first reaction that came to the author's mind in April last year, when Volcano Engine released the so-called "data flywheel" of the new paradigm of enterprise digital and intelligent upgrading. To some extent, this reaction is understandable for media professionals, because words like "data-driven" and "growth flywheel" probably don't appear much less frequently in tech manuscripts than "the Internet" itself, which should be considered common sense.
Therefore, if we analyze from the perspective of "data-driven growth flywheel", the concept of "data flywheel" is actually relatively clear on a macro level, and what I am curious about is whether there is anything substantial new in the discourse of volcano engine.
"Targeting" data center
Volcano Engine divides the data flywheel into two parts, the upper layer is business applications, the lower layer is data assets, and there is a common data consumption as a link in the middle. On the one hand, the business side should truly use data through data consumption, and realize decision-making science and agile action through data, which in turn will bring value improvement. On the other hand, a positive cycle should be formed around the data assets themselves, and the core is still data consumption. The intensity and frequency of data consumption resonate with the quality, breadth and precipitation speed of data assets.
At this point, a natural question is what is the difference between the "data flywheel" and the "data middle platform" mentioned above? The internal management of Volcano Engine has explained this issue on different occasions, and here the "New Position" summarizes it based on its own understanding.
The earliest Internet companies proposed the middle platform because they hoped to integrate technology and data, which are common to all business front-ends, so as to reduce the cost of building these infrastructures and easily support the innovation of various front-end business models. Of course, with the growth of the business, the middle office became thicker and thicker, not only did it not have the flexibility and convenience expected in the initial design, but because the separation of the front, middle and back offices greatly hindered the efficiency of communication between various departments, there was a trend of "going to the middle office".
From the above historical evolution, it can be seen that the emergence of the data middle platform itself is the product of the great guiding principle of "resource integration and intensive utilization". Before the data middle platform was proposed, each line of business also had some needs to use data. The emergence of the data middle platform does not mean that it has played a new function, it is still the same role positioning, but the scattered needs are solved together in batches.
From this point of view, the data center is completely different from the data flywheel in terms of goals, the former focuses on the middle platform, and the center of gravity should be used to achieve the unity of data, while the latter focuses on data, and the flywheel is an effective data-driven paradigm.
For example, Tan Cheng mentioned in an interview last year that the "consumption-centric" data-driven is a bigger upgrade than the data middle platform concept talked about in the past five or six years, and can more effectively solve the problem of enterprise data generating value. He believes that many enterprises have invested a lot of resources to establish a data middle platform, but have not made use of the data, the essence of which is because the goal is biased: to achieve data-driven, we must start with the end in mind, and build with data consumption as the core of data-driven.
Tian Lan, head of data product solutions in the large consumer industry of Volcano Engine, also had a similar expression in a business roundtable:
I think the concept of the data center itself is to do data construction, to do the middle platform to do data unification, which is the biggest core point in that era, or when the data center was put forward at that time. The biggest difference or core difference between us in the data flywheel is that we emphasize the two-way benign drive between data and business, which is the difference in core concepts.
The above perspective given by the official Volcano Engine is reasonable, because when people designed the data center, they did have a strong demand for data access. But it is an obvious fact that in addition to this demand, the traditional data middle office also carries the goal of facilitating the front-end to drive business through data. For example, making the data middle platform easy to use at the front end has the meaning of emphasizing data-driven business. Therefore, only talking about the first half of the functions of the data center, and deliberately ignoring or denying the second half of the functions, is somewhat suspicious of setting up targets to hit.
"Break into" traditional industries
However, the New Position does not mean to deny the positive significance of the "data flywheel". On the contrary, even if there is a certain component of packaging in concept, it is a very beneficial packaging.
Because one of the biggest problems with the "data middle platform" is that it is a set of discourse systems derived from Internet companies. However, the target customers of today's cloud computing vendors are not Internet companies, or even pan-Internet companies, but traditional enterprises in various fields. They don't have a front-end changeable 2C business, and naturally they can't figure out the meaning of the data center, and they don't even have that much data.
Therefore, at this level, the switching from the data middle office to the data flywheel is of key significance to truly promote the digital transformation of the whole industry.
Corresponding to this set of discourse switching, Volcano Engine also takes out the grip of "data consumption", supplemented by a special BP mechanism and a more quantitative evaluation system such as "0987". For many traditional enterprises, these arrangements will make their digital transformation journey more targeted.
In terms of trends, it is clear that digital transformation is already an irreversible trend. Most of the executives of enterprises in various fields have actually seen this, as evidenced by the popularity of collaborative office software such as DingTalk and Feishu. But at the same time, it is an undeniable fact that in addition to replacing basic administrative functions, enterprise digitalization is not smooth in the process of breaking out of the Internet, entering various traditional industries, and participating in the optimization of operation and management processes. This is also the fundamental reason why Alibaba Cloud, Tencent Cloud and other established cloud vendors have experienced sluggish performance growth in recent years.
For example, the digitalization of the dairy industry is obviously not the same thing as the digitalization of home appliance manufacturers. Because the digitalization of various traditional fields is bound to be customized in combination with the actual situation of the industry and the enterprise itself, which can be described as a thousand faces. But the general solution is that gaining the trust of these traditional enterprises and helping them build confidence in digital transformation must be the first key to knocking on their doors. This is the meaning of a concept that is easy to understand, relevant to the business, and can be quantified.
However, at present, only Volcano Engine seems to be pushing the concept of a "data flywheel" vigorously among cloud vendors. In the view of "New Position", this is not only because Byte has the genes of data-driven business far beyond other large manufacturers, but also because as a latecomer in the cloud computing industry, Volcano Engine needs to play differentiated competitive characteristics.
All in all, although these cloud computing companies are now very advanced in digitalization, what is more decisive for their business is how to make customers in traditional industries "have the bottom of their minds" about digitalization, at least to give them a methodology, tools and platform that can be used.
*The title picture and the accompanying pictures in the article are from the Internet.