Author: Liu Kai, Founder of Collaborative Data Technology

With more than ten years of experience in data governance and BI analysis, he has worked in Huawei, four major accounting firms and non-profit organizations. The institutions he has served include: Asian Development Bank, People's Bank of China, China Judicial Big Data Research Institute, Shandong City Commercial Bank Alliance, Nanjing, State Grid, PetroChina, China Building Materials , China Railway Construction, China Resources Group, Financial Street Group, Yuexiu Financial Control, Haier Financial Control, Hong Kong Monetary Authority, etc.

Text:

In this world, how can the data business make money?

There are three paths:

Marginal compensation increment * Data turnover acceleration * Data leverage amplification

Why do you say that?

In this article, we look at how to make money from the perspective of financial thinking. In the digital era, data has become one of the most valuable assets of enterprises, and its value creation and profitability are increasingly becoming the key to measuring the success of enterprises. These three paths are the core elements from the perspective of data management and finance.

1. Marginal compensation increment: make money with the core competitive advantage established by data elements.

Marginal increase in remuneration means that with the increase in the amount of data elements and the deepening of data application, the additional benefits that can be brought by each additional unit of data elements gradually increase. This is the opposite of the law of diminishing marginal remuneration in traditional economics. The marginal value of data increases with the deepening of use and analysis.

Why is it so important?

The increment of marginal remuneration means that we cannot evaluate all data elements equally, but should pay attention to those data that can bring positive feedback loops. We should pay attention not only to the amount of data, but also to the quality of the data and the depth of application. Through technologies such as data mining, machine learning and artificial intelligence, enterprises can extract valuable information from massive data to realize the in-depth utilization and value maximization of data.

Imagine that you are the owner of a garment factory. Recently, you have found that the production cost is getting higher and higher, while the profit is getting thinner and thinner. You try to increase the number of workers in the hope of increasing the output, but it seems that the effect is not obvious. At this time, competitors on the other side of the ocean use some mysterious new methods to produce more clothing at a lower cost. What's going on?

You may think that now college students are of high quality, and graduates only need 3,000 yuan, which is less than half of migrant workers. If you increase the number of workers, the output will naturally increase. This is common sense. But why doesn't this method say now? In traditional economics, this phenomenon is called marginal reduction in remuneration.

However, if we introduce data elements and advanced technology, the situation will be different. An incremental increase in marginal remuneration is possible. Let's take American super factories as an example. Through big data, AI and automation technology, the time of clothing production is compressed to 22 seconds, and the cost is less than 2 yuan. This is the power of "incrementative marginal remuneration".

So, how can we achieve an incremental marginal remuneration in our own clothing production? Here are some specific practices:

(1) New quality productivity input to realize the input substitution of data elements: use robots and automation technology to replace part of the labor force in repetitiveness and high intensity, reducing dependence on manpower and improving production speed and quality. The initial investment is large, but in the long run, it can reduce costs and improve competitiveness. For example, by switching different process schemes and adjusting the process parameter data, the assembly line can automatically produce different products. For relatively simple T-shirts and denim clothing, at present, American super factories can use a piece of clothing in 22 seconds, costing 2 yuan!

(2) Data sharing, optimize the production link and supply chain through data elements: use big data analysis to optimize fabric procurement, cutting and sewing and other links. Accurate supply chain management can reduce the cost of raw materials and improve the response speed. This requires close cooperation with suppliers to establish an efficient information sharing mechanism. For example, in the early days, only contract transaction data was shared with suppliers, which was further expanded to real-time inventory data, production progress data, quality inspection data, logistics data, etc., so as to optimize the production link.

(3) AI data analysis to achieve value doubling through data elements. Improve the traditional single-factor production efficiency, optimize the design process through AI algorithms, quickly respond to market changes, and launch clothing styles that better meet the needs of consumers. For example, SheIn Company can achieve more, fast, new and economical clothing production at the same time. According to the research report, SHEIN has about 7,000+ new daily, which is 200 times that of ZARA, and the proportion of unsold inventory has been reduced to single digits and the inventory turnover is only about 40 days, far lower than the industry average (about 30%) and 90 days.

The decrease in marginal remuneration refers to the phenomenon that in the production process, with the increase of a factor of production (such as workers), other factors (space, machine) remain unchanged, and the additional output (the contribution of each new worker to the total output) generated by the newly added factors of production will gradually decrease. Through the three modes of resource optimization, input substitution and value doubling, data elements are reshaping the production mode and value creation process of the clothing industry, improving production efficiency, reducing costs, and achieving incremental marginal remuneration.

The incremental marginal effect is usually the embodiment of core competitiveness, and core competitiveness is precisely based on data assets to form a flywheel effect. Take Lianjia's real estate dictionary as an example, it can be seen that many services have been derived to realize data realization and continuously strengthened. Below is a simple diagram based on the real estate dictionary:

Two. Real-time data, make money with the speed of data flow turnover!

Data turnover acceleration refers to improving the liquidity and efficiency of data, accelerating the whole cycle of data from perception, judgment, decision-making to action, so as to quickly respond to market changes and business needs.

In command theory, there is a famous model: the OODA ring. It abstracts the combat process into four steps: observe, orient, decide and act. These four steps form a ring. Whoever can shorten this loop means that whoever can complete a strike process faster than the opponent, the greater the overall chance of winning. The faster the operation speed of big data, the shorter the OODA ring is compressed, that is, "integrate and fight".

This real-time data closed loop brings two benefits:

(1) It is that every closed loop can create value, such as unmanned intelligent mines. Each transportation process is a closed loop to create value, and the amount of value is directly proportional to the number of turnovers.

(2) It is the continuous improvement of accuracy (conversion rate) and the rapid formation of a core competitive advantage. For example, with the increase in the number of unmanned driving and the continuous improvement of accuracy, the intelligent driving system can become an industry standard or infrastructure.

The OODA cycle of big data is similar to the inventory turnover in the clothing industry. A business data can not only be used in multiple application scenarios such as industry, finance, ticket, tax, etc., but also create value in every decision and action. Therefore, the faster the cycle turnover, the more value, and each closed loop brings action to create value. The monthly business analysis meeting is usually an offline data report or PDCA cycle, which is not in the same order of magnitude as the value created by real-time online decision-making. Therefore, in today's rapidly changing market environment, the timeliness of data and high-speed flow turnover are crucial. We need to divide the data into two categories: offline data and real-time data.

Offline data, as we know, is the data that is no longer updated or updated very often. This kind of data usually exists in the form of data sets, packets and data reports. They are very useful in historical analysis and batch processing. As a commodity, offline data can be traded and recorded as assets. Its special feature is that it is difficult for buyers to judge its true value before buying. Once the buyer grasps the content of the data, they can easily copy the data, thus losing the need to buy again. This phenomenon is fully exposed in information economics and is called the "Aro Paradox" by Nobel Prize winners. Another feature of offline data is that it has low timeliness and cannot be intervened immediately. It may have paid for the mistakes in decision-making.

Real-time data. Continuously updated, very sensitive to time. The transaction of real-time data is more like a service that provides data flow in real time through API. The value of this data changes with time and data updates, so it is easier to quantify. Due to its online characteristics, real-time data is easier to monitor and protect through technical means, thus improving the security of data. However, in principle, the real-time data of the subscription can only be recorded as operating costs, not as assets, because after a certain period of time, the value of the data will be depreciated to zero. For real-time data, online transactions, confirmation, pricing, circulation, security and other problems do not exist, and have been proven to be successful in the financial and securities market. In the data trading market, the trading mode of offline data is usually one-time purchase, while real-time data is more inclined to the subscription service mode. This means that enterprises can obtain the right to use data as needed, instead of purchasing ownership of the data at once.

In order to make full use of the value of real-time data, enterprises need to integrate real-time data into their operating systems, make real-time decisions and turn it into action. The acceleration of data turnover means that enterprises can respond to market changes faster, use real-time data to obtain hundreds of times more returns than offline data, improve decision-making efficiency, and thus gain an advantage in the competition. At the same time, the technical threshold for building a real-time data trading platform has been greatly reduced, usually with a budget of one million RMB. Any enterprise with real-time data can quickly build a trading platform to provide services.

The OODA cycle of data elements also enhances the learning curve effect and improves the conversion rate or prediction accuracy. The learning curve effect refers to the phenomenon that with the accumulation of experience, the efficiency of individuals or organizations increases and the cost decreases when performing specific tasks. This concept was first applied to the manufacturing industry, describing the decrease in the production cost of a unit product as production increases. In a broader field, the learning curve effect means improving performance and efficiency through practice and experience accumulation. By collecting and analyzing historical data, enterprises can identify patterns and trends in business processes, accelerate the learning process, and further use it to predict future trends and optimize decision-making, so as to reduce the cost of trial and error. Through machine learning algorithms, enterprises can improve the accuracy of prediction and decision-making by continuously learning and adapting data, and The dynamic system can automatically adjust the operation and improve efficiency according to historical data and real-time feedback, which continuously improves the accuracy of the OODA cycle.

3. Data asset leverage: use leverage to leverage other people's data to make money!

Data asset leverage refers to the continuous amplification of the radius of data that can be used for free through a small number of proprietary data assets and a series of data replacements and other means, so as to amplify the value of data and the profitability of the enterprise. The formula of data leverage can be regarded as the ratio of available data/own data, and other available data are free usage rights obtained by the connection.

A typical method is to directly establish ecological precipitation data sharing by establishing a digital platform with network effects. Take SheIn as an example, SheIn not only shares its own data, but also outputs its own system. Through the system, it can obtain a large amount of real-time data from various small clothing and workshop production and processing, Amazon, independent stations, social media, etc., from demand forecasting, product development, ordering, dispatching to production process. The management realizes the automatic closed loop in the system. From supplier introduction, order, evaluation, and elimination of automatic management of the entire life cycle, all depend on this digital operating system. This is the reason why SheIn can achieve the industry ceiling, and these digital systems are also his protector.

Another way is to establish a digital system for the upstream and downstream of the industry/industry. For example, in the field of electricity. An industry supervision system can be established, so that real-time equipment manufacturing progress data and workshop quality inspection data can be obtained from upstream suppliers (such as large and small cable fields), control quality risks and delivery risks at the source, save the cost of traditional on-site supervisors, and improve the number of upstream suppliers. According to the degree of standardization, more system functions can also be expanded based on the supervision system, which can not only monetize existing users, but also sell the system for profit.

A smarter way is to borrow flowers to offer Buddha and make full use of cross-industry multi-source heterogeneous data. For example, a large provincial energy data center can be established for the provincial government. While sharing its own electricity data and providing public welfare power data product reports for the government and society, it can actually use water, coal, wind, gas, oil and other multi-source heterogeneous data for free, while providing more comprehensive upgraded energy data for the government and society. The product report can provide electricity-using enterprises with a comprehensive, energy-saving, emission-reducing and consumption-reducing dual-carbon solutions. This method not only saves the cost of external data procurement, reduces the difficulty of data access, and avoids the cost of operation and maintenance of data centers, but also obtains better cash flow through the construction of large energy data centers. At the same time, it also helps the government establish a dual-carbon industrial park to attract investment and improve political achievements, which can be said to be a win-win situation. This method is essentially similar to financial leasing. It can be used by the government, industry associations, etc. as lessors (leasing companies) of data to bear the construction and operation costs of the public service platform for data assets, and each enterprise that shares and opens up data is both the lessee (enterprises that need to use various data assets) and the supplier Yingshang (the provider of its own data assets), since data is usually used for free, whoever has the core competitive advantage of data development and utilization (which has formed a flywheel effect) and who has the ability to operate real-time data will benefit the most.

In summary, the key to amplifying data leverage is that by contributing a small amount of own data, it can leverage a large number of ecological data upstream and downstream of the industrial chain and multi-source heterogeneous, and carry out light assets, light operation and low-risk business. The core is not how much data you have or how many data assets you gather, but how to more conveniently obtain real-time data, connect, and quickly iterate to use real-time data. The starting point of the ability to use data may not be high, but through continuous iteration and accumulation of learning experience curves, it can grow rapidly and establish its strong competitive advantage barrier.

The Last Word

From the financial perspective of data operation, the marginal remuneration (increment), data turnover speed (acceleration) and data leverage (amplification) of data assets. The multiplication of these three factors constitutes the profitability of data management.

The incremental marginal remuneration tests the core competitive advantage you build with data assets. The more types of value-added services derived from data assets, the stronger the flywheel effect, and the greater the value.

Data turnover accelerates to test your real-time data operation ability. The faster the data flow and the more closed the loop, the faster the data turnover speed and the greater the value.

Data asset leverage tests how you can use your own data to leverage the resource integration ability of third-party and fourth-party data. The more data you use for free, the stronger the data development and utilization ability, and the greater the value.

In addition, you must consider free cash flow, which is better than profit. For example, you should turn the capital expenditure originally used to purchase data into the income from the construction of profitable digital systems and the income from the construction of large data centers (project prepayment). Because profit is only one of the components of free cash flow, it cannot truly reflect whether a company is profitable and whether it has development potential.

So, which of the above three paths do you choose to use?

Are you making money by relying on the core competitive advantage and flywheel effect formed by data innovation, the real-time iteration operation speed of data, or the ability to integrate data resources?

Many enterprises have made great efforts in one of the paths.

If you can do a very good job on these three paths, then congratulations.

If you have just entered the field of data assets today, you must think clearly about which path to build your core competence on.

Use the financial perspective of data management to find your success factors.

I hope to give some inspiration.

(Article Source: Liu Kai Official Account: Yecai Data Management)