Guided reading

       The data generated by the Internet of things is gradually becoming the main force of big data. This paper aims to explore an issue that most innovative enterprise executives are very concerned about: how to convert the data into sustainable business value step by step along the Internet of things data chain?

"Although the data has not been included in the enterprise's balance sheet, it is only a matter of time.”

       ——Big data age Victor & middot; Mel Schoenberg

       Some people say that the role of data on society is like oil. Others say that the impact of data on society will surpass oil. I think it's all right.Data and oil are both energy sources to promote social change, and their generation and application processes are the same;However,Data has many characteristics unmatched by oil:It can be reused, copied and transmitted at a high speed, so the influence of data on society will far exceed that of oil!

       In any case, comparing data to oil helps to better understand the context of the Internet of things data chain.I happened to have worked in an oil company and have some knowledge of the business process of oil. Let's compare the data to oil.

       Before oil was exploited, it was only a mixture of hydrocarbons with complex components deposited in the earth's crust. However, once it was sprayed out from the underground, it could be used for a variety of purposes and become various necessities of modern society, such as fuel (gasoline, diesel, etc.), lubricating oil, chemical raw materials, plastic products, etc, Its commercial value has also been gradually displayed incisively and vividly.

       The process of mining the commercial value of data is like refining virtual digital oil.

Figure 1: oil production and marketing process of Internet of things data link

       As shown in Figure 1,For the Internet of things, mining data value should also start from looking for data mines, and then, by continuously improving the ability to collect, communicate, store and analyze data, further convert those filtered and analyzed data into commercial value that can reflect the advantages of enterprises.This process forms a data link similar to the oil production and marketing process:Data acquisition (oil exploitation), data communication (oil transportation), data storage and analysis (oil storage and refining), application and conversion to commercial value are shown in Figure 2.

Figure 2:Schematic diagram of Internet of things data link

       In reality, new data are still generated in the process of data application and conversion. These new data are fed back in real time and return to the initial stage. Together with other data, they go through the whole chain process again, forming a "acquisition application feedback acquisition" cycle process different from the traditional "production consumption waste" linear process of oil, This is quite similar to the sustainable process of "production consumption recycling" experienced by oil under circular economy, which is why sustainable business strategy is very important for the construction of Internet of things ecosystem.

       The following is a step-by-step discussion on how to convert data into sustainable business value along the Internet of things data chain according to the steps shown in Figure 2.

Data collection: find big data target mining area

Figure 3:Schematic diagram of Internet of things data acquisition

       The composition and appearance of crude oil vary from place of origin, and the original data are all inclusive in form.Words, pictures, sounds, symbols, signals and other information that can be digitized by the computer can become original data!In addition, oil has specific threshold requirements for the geology, landform and mining technology of its origin, and the mining technology is always keeping pace with the times.Similarly, the discussion on the collection of Internet of things data must also correspond to specific scenarios and use evolving technical solutions for collection.

       Before the rise of the Internet of things, our traditional data generally refers to relatively static, structurally fixed and accurate data that came from within the enterprise, interacted with customers manually or read regularlyNow the Internet of things extends the scope of data collection widely. It includes real-time massive dynamic data from production equipment, users and products over time. It is more comprehensive and timely than ever, and fills many data gaps in traditional business processes.

       It is the characteristics of Internet of things data that tend to be "complete", "comprehensive" and "massive", which makes us have to look at these data in a different way of thinking. We no longer have to focus on the rigorous causal relationship between limited data, but have more insight into the possible relationships between various things represented by data, so as to find out the potential value and grasp the future trend.let me put it another way,The core function of IOT data is to create a more colorful future through "prediction".This new way of thinking is closer to the ecological thinking of understanding life evolution, rather than the precise thinking of understanding mechanical movement.

       So what are the IOT data that can be converted into commercial value?From the perspective of its effect on business processes, I think it can be roughly divided into the following four categories:

 
       Data for improving equipment energy efficiency and performance：Internet of things technology can intelligently manage and optimize the energy consumption of equipment, reduce manual participation, and save the energy loss of the whole business process. This is why people generally classify Internet of things technology as "energy-saving technology".According to the actual application scenario, the energy consumption data we need to collect include real-time power, heat, water volume, pressure or calculated coal equivalent, as well as the associated voltage, current, volume, etc.

       Data for equipment diagnosis and early warning：For example, specify the temperature of the equipment location, equipment vibration frequency, fluid flow, environmental noise, and even the pressure of the equipment contact gap.The specific parameters to be collected vary according to the application scenario, so as to prevent or reduce possible faults or disasters in the production process.

       Data used to improve customer service：Desensitized customer transaction data, such as transaction quantity, transaction volume, transaction frequency, transaction location, product type, customer score, as well as customer use data, such as various spatial experience data for smart home.These data can not only provide customers with personalized reference opinions and real-time services to help customers make transaction decisions, but also help enterprises provide more personalized services for new customers and future cross-border partners.

       Data used to implement "as a service (xaas)" authorization："Communication equipment" in the era of Internet of things includes various objects. Their normal operation time, displacement distance, failure time and so on determine what services users have the right to use under what conditions and when.These data play a particularly prominent role in the transformation of various "as a service" business models.On the basis of authorization, users can also be directly priced, settled and deducted online.

along withThe application scenarios of the Internet of things tend to be diversified, and the types of data available for collection and analysis will be continuously expanded. The data types mentioned above may only be a part of the future Internet of things data, but the new data are basically inseparable from the core function of "finding out relevance, predicting results and realizing value".Around this core, it can be predicted that Internet of things data will lead business into a new era with more intelligent technology and more ecological thinking!

Data communication, storage and analysis: treasure hunt in data

Figure 4:Internet of things data communication, storage and analysis

       As IOT data is virtualized, changing over time and reusable, the communication, storage and analysis process of IOT data has more complexity and flexible development space than that of oil transmission, storage and refining.This process involves a large number of emerging cutting-edge technologies or technology collections, as shown in Figure 4, including various new communication technologies and protocols in the field of data communication, local and wide area Internet of things networking technologies, as well as public or private cloud storage, cloud computing, edge computing, machine learning, artificial intelligence, data visualization, blockchain, etc. in the field of storage and analysis.Because these forward-looking technologies continue to penetrate into business, this link is also extremely easy to give birth to subversive new technologies and new business models. At the same time, it is difficult to avoid the smoke of short-term and white hot trade disputes and commercial wars.

       Among these new technology solutions and business models, only those new technologies, new platforms or new business models that can integrate more diverse technologies and business resources are the final winners in the future technology market, such as smart chips, 5g communications and Internet of things operating systems that occupy the core area of Internet of things technology, Or connect the upstream and downstream ecological partners of the Internet of things, provide an Internet of things data platform with rich APIs, and so on.The commercial formats in the Internet of things era, from personnel organization to technical architecture, are tending to be flat and open-source, various traditional technical barriers are being broken, and commercial authority is being delegated. Therefore, the dying struggle of some leaders in traditional industries has also been exposed one after another.

       How to grasp the technology and business trend of this key link, take the following typical Internet of things architecture extending from cloud to fog as an example.

Figure 5:Architecture diagram of Internet of things edge analysis platform

       As shown in the figure, with the increasing diversification of intelligent terminal equipment and Internet of things business scenarios, it is required that the data transmission and feedback speed be faster, especially for smart grid, oil exploration and most industrial manufacturing fields with remote operation, The deployment idea closer to the mobile Internet in the initial stage of the Internet of things -- omitting the gray background in the middle of the figure and directly centralizing the data collected by terminal devices to the cloud -- is becoming more and more ineffective.Of course, for many actual industrial Internet of things application scenarios, this architecture needs to be further personalized, which may involve the intellectualization of various old equipment, man-machine interface management (HMI), etc.

       In any case, the technical solution deployment position of the Internet of things, especially the industrial Internet of things, is moving from the cloud to the fog layer closer to the terminal equipment.This is like the cloud general manager delegating the data processing authority to the fog area managers in various places for independent processing. In this way, the fog layer analysis platform can analyze the data in more real time and make more timely and appropriate business decisions.The power distribution between cloud and fog is a balance art between centralization and partition. The cloud should delegate power, but not too much. The key is to keep the real-time information flow channels between the two smooth, timely and effective.

       From this point of view, the massive dynamic data generated by the Internet of things architecture is like flowing water. It evaporates from "terminals" around the world into steam, fog, cloud, rain and returns to the earth.If we say that the water cycle in nature breeds evolving natural things with the help of the energy of sunlight under the action of the earth's gravity, the data flow of the Internet of things develops and evolves in the direction conducive to the birth of each vertical industry and the construction of their own business ecosystem with the help of the wisdom of people and machines.

       In general, the expansion of the Internet of things technology framework from cloud to fog is in line with the ecological trend of business development. It makes the technical division of labor in the commercial society more detailed and the management more flat.Among all subdivision technologies, it is undeniable that data analysis technology is very key. It can be said that it plays a connecting role in the whole Internet of things data chain and has an immediate effect.The most eye-catching field in data analysis is probably artificial intelligence (AI). Many people think that AI is a single technology. In fact, it is not. It is a collection of various technologies related to intelligent data analysis, from "visual sensor" and "Ai chip" at the hardware level to "face recognition", "unmanned driving" and "speech recognition" at the application level "Building intelligent housekeeper", "energy saving analysis" and the software technology, patents, standards, protocols, etc. contained therein are all part of the artificial intelligence technology set.

       In the collection of artificial intelligence technology, the most prominent subset is undoubtedly "machine learning", which uses computer computing power to mine valuable information or laws from data.Relatively speaking, traditional programming calculates the results according to the algorithm rules given by the programmer, and the machine in machine learning can act as the programmer and re-use each calculation result as data for self correction and continuous verification.Among them, with the help of multi-level artificial neural networks (ANN), the application of calculating and analyzing massive data to the optimal solution brings "machine learning" into its most important subset "deep learning".The difference between them is shown in Figure 6:

Figure 6:Comparison between traditional programming and machine learning

       There is no doubt that "deep learning" will be the core competition in the field of Internet of things data analysis for a long time in the future. The process of deep learning is equivalent to using multi-level neural network to find a treasure hunt map in the big data maze, and it is a treasure hunt map with optimal path. With this treasure hunt map (algorithm), We are more likely to eventually find new babies (commercial value) in many complex data.

       Speaking of this, I can't help feeling:Even machines are trying to learn and make intelligent independent judgments. Why don't we redouble our efforts to learn and form our own independent ideas?This resonated with the first French representative organization in China that hosted the doctor of Business Administration (im-dba) project in intelligent manufacturing and proposed inclusive learning. I roughly expressed this view at the internal Seminar:In the future, business decision makers should continue to learn and think like the "deep learning" artificial intelligence machine, refine personal practice (data) and performance (results) into guiding ideological values (business laws) for their industry, so as to maintain their competitive advantage.Since machine learning is a business direction worthy of increasing investment in the future, personal lifelong learning is also a life direction worthy of continuous investment?

Converting data to business value

Figure 7:Schematic diagram of data conversion to business value

       For example, refined oil enters all walks of life to become fuel or production raw materials. The results of Internet of things data analysis must be integrated into commercial service processes and become products or services in order to reflect its commercial value.Generally speaking, the more data services that can improve the integration ability of commercial ecological resources, the higher the value of commercial sustainability.

       In addition to the value as a raw material like oil, the actual value of data is equal to the sum of all possible uses. It can be repeatedly used to change business processes, improve operational efficiency and generate value-added.

       So, how much business value can data be converted into?In other words, how to value data, a new type of intangible asset in the future?At present, there is no industry evaluation standard in this regard.However, just like the companies that evaluated various assets in the past, some refreshing data value evaluation companies will certainly appear in the market in the future.

       In any case, in order to convert data into commercial value, we must not only have the corresponding available technical capabilities, but also have the strategic awareness that data can be converted into commercial value. Both are indispensable.The following three cases can well illustrate this truth.

       Case 1:After collecting vehicle operation data and cooperating with an external data analysis company, a European automobile manufacturer found that the fuel tank detection sensor provided by its German supplier had the defect of frequent false alarm. After careful consideration, the automobile manufacturer did not choose to directly tell the supplier the information and order it to make rectification. Instead, it applied for a patent for the improvement by improving the software and the corresponding parts on the fuel tank, and then sold the patent to the supplier to establish a new partnership with the supplier, so as to find a sustainable commercial return for its early investment, We are happy to enhance the business ecological cooperation ability of enterprises and improve the technical development level of the whole automobile manufacturing industry, including competitors.

       If case 1 was accidentally encountered by a high-tech enterprise in the process of data analysis, and they seized the business opportunity in time only because they were aware of its value, then case 2 is just the opposite.It is a new marketing scheme actively planned by a South American traditional manufacturing enterprise I came into contact with in the recent smarties x 2019 review activity. Its unique idea is an eye opener.To be exact, the data of this case comes from the mobile Internet rather than the Internet of things, but their ideas of associating irrelevant data into business value are consistent, which also has reference significance for business decision-makers.

       Case 2:A century old shop in Brazil that produces and sells saucepans is called tramontina. Seeing fewer and fewer young people cooking by themselves, the product sales prospect is worrying. In order to change this situation, the enterprise has established an international project team composed of chefs, neural network experts, band conductors and other professionals to carry out a win-win cooperation with spotify, a world-famous music network loved by young people. They spent half a year studying how to use the "synesthesia" algorithm to associate music hearing with food taste. Firstly, the music characteristic parameters that can correspond to cooking parameters are screened on spotify, such as negative tone corresponding to bitter taste, Concerto played by multiple instruments corresponding to heavy taste, music duration corresponding to the number of ingredients, note jump corresponding to cooking temperature, etc., and a corresponding large database is established. Then, use artificial intelligence to convert the data into a recipe of tens of millions of orders of magnitude and publish it on the new website of "flavor of songs". Finally, when potential users browse, they timely give a guide on how to choose a pan to turn recipes into food, so as to attract more and more young users of music network to experience. In this way, the old-fashioned pan combined with personalized cooking has become as popular as music, changing young people's eating habits and guiding them to the kitchen.

Figure 8: screen capture of tramontina|flavor of songs

       From these two cases, it can be seen that the data collected intentionally or unintentionally have potential commercial value. In the future, enterprises need to spend more energy on mining, sorting, screening, analyzing or reorganizing the collected data, and choose a conversion path suitable for their own enterprise's actual situation, just like the automobile manufacturer in case 1, Convert data into a new business with intellectual property attributes;You can also use the data to establish a new marketing channel like the kitchenware manufacturer in case 2.

       Of course, some enterprises, especially those that can raise the value of mining data to the height of enterprise strategy, can go further. They often take the data as an opportunity to change the whole business process and reshape the business model.In my previous articles on "Internet of things ecosystem" and "PAAS (product as service) business model", I introduced many cases and gave relevant analysis. I just saw another case mentioned in Dr. LV Jianzhong's article "sustainable development path of manufacturing industry transformation to digital intelligence", which is very clear and easy to understand, It is also very representative.Excerpts are as follows:

       Case 3:"In the book" intelligent transformation "(digital transformation, David Rogers) There is such a true and interesting story: TWC, a company mainly engaged in weather forecasting services, takes collecting, processing and sending weather forecasts, and then publishing advertisements on the publishing platform as its main business model. However, TWC soon realized that the potential value of data far exceeded the revenue from advertising. So they set up a group of data scientists led by Sofia. The task of this group is to turn the data tool into a strategic asset to create higher added value. The research team found that the change of weather had an impact on the fluctuation of one-third of economic activities in the United States. So they formed a joint working group with Wal Mart to establish an economic analysis model linking meteorological data and sales categories and sales volume, and use this model to speculate on which products consumers will have higher purchase demand and willingness under what weather conditions, so as to formulate advertising strategies for popular product categories affected by the weather, So as to obtain the maximum sales revenue and improve the return rate of advertising input-output. The research team also worked with insurance companies to develop a small app called hailzone to remind car owners to move their cars into the garage before hail or snowstorm. This small app has greatly reduced the car damage rate, alleviated the inconvenience and annoyance of car owners, and also reduced the claim claims and costs of insurance companies, which has been praised by consumers, insurance companies, communities and other stakeholders. All of a sudden, many people registered with hailzone. Some meteorological enthusiasts also brought their own meteorological observation tools to send and share the meteorological change data monitored in their own place to the hailzone platform at any time. In this way, the hailzone platform can integrate a large amount of data from all directions every 1.5 seconds, give more real-time and accurate prediction through special algorithms, and enter a positive cycle of quality improvement and value creation. "

       In case 3, the enterprise consciously linked the data with the external environment, actively developed external ecological partners, established a real-time interactive network with retailers, insurance companies, consumers, communities and other institutions and individuals, and transformed the original relatively static weather forecast database into a real-time data update and interactive platform with the participation of people and equipment, Provide more timely high value-added services for stakeholders, and obtain sustainable business ecological effects that can grow by themselves.

       A brief summary of the ways in which relevant data are converted into sustainable business value includes the following three categories or a combination of these three categories:

       Lease the use rights of their own products or services to form a data platform that can interact, pay and settle with users (such as cases 1 and 3), such as some high threshold industrial product leasing, IP, consulting, information services, etc;

       Establish convenient data service portals for product partners and users, and then sell these portal permissions or establish new marketing channels (such as cases 2 and 3), such as providing intelligent logistics services, product use and payment service portal permissions, etc.

       Help customers add external data to the specified data set and provide value-added services (such as case 3), such as providing remote heart rate monitoring services and interpretation and diagnosis of detection reports, and the value-added relationship between environmental data and economic activities.

       In a word, the key to transforming data into sustainable business value is to build your own ecosystem around the data chain, and its success depends largely on how to formulateAnd implement sustainable business strategies.

       This article is transferred from the official account of the Internet of things think tank. If there is infringement, please delete the background message.