Big Data Illusion
Is big data really an objective urgent problem for a business?
Maybe this is just a beautiful marketing course for developers of powerful computers and products for storing and processing digital data.
Maybe this is just an attractive advertisement for market research consultants and customer behaviors.
Or maybe this is just a fashion trend in the field of total observation of market entities and forecasting their reactions.
Perhaps there are no “big” data, but there is a big illusion that it will be possible to somehow gather such a huge array of digital information, process it in some magical way and get answers to all the questions of concern to a businessman.
A still from "The Special Opinion" (Minority Report) by Steven Spielberg based on Philip Dick's novel (2002 - 20th Century Fox, DreamWorks SKG).
')
Big data is essentially an analyst resource. This is a resource for people carrying out research and decision-making preparation. And like any other resource, big data does not work without the ability, knowledge and technology to use them. Someone calls this skill "data mining" (data mining) - by analogy with mining, focusing on deep penetration and complexity. Someone calls this skill “business intelligence” (business intelligent) - showing how important the “mental” component is in this process. Someone like the name "big analyst."
But it is known from theory and practice that even the presence of a resource in large quantities does not mean its successful and effective use . Sometimes, an excess amount of a resource allows you to build a business model, not deeply processing it into a specific set of products, but rather in simple packaging and sale in its raw form. Why look for additional options, when, without putting in excessive efforts, you can simply sell the raw resource.
Big data, as an informational category, has one feature in contrast to material resources. To use them , a truly high level of organization of business objects and company business processes is required . Without such a level of preparation, without a certain qualification from the business, the purchase (or collection) of big data will be characterized by low efficiency. So low that it does not justify the investment in them.
Why should a business spend money on big data if a business management layer (decision making) is not based on analytics? Absolutely right - no need. To this or that extent, come those companies that have begun to use big data without adopting analytical technologies into the control loop, a technician who has prepared analytic decisions and who, by and large, are not ready for change. Such subjects of the market will sooner or later give up big data. Especially the issue will be acute with increased competition for financial resources within the business.
Today, the big data market has focused on information technology. It is understandable and pleasant that tools for working with big data are developing. But the intensive growth of information networks and the improvement of information technology removes barriers to computing power. This will force advanced ambitious businesses to reconsider their current hobby and shift the focus towards new effective methods, tools, and management technologies based on knowledge and training.
Actually, when big data is presented, it is often a question of the possibilities of their storage, transportation and processing. The search technologies of the giants of the Internet are a vivid example of how big data are given to businesses. Algorithms of search is the most powerful processing of the huge growing volumes of information. They are constantly in the process of optimizing, improving the performance of indexing and structuring information. But after all, behind the search technologies in the network are not only big data. Behind them are the teams of analysts who possess high-tech knowledge in the subject areas.
Therefore, a sensible policy of using big data is building a data analysis team , but in no way an exceptional alignment of servers, clouds, data mining systems, machine learning, etc.
It is worth noting that the definition of “data mining” is not very significant. It draws a somewhat simplified picture of reality: there are “invaluable deposits” of heterogeneous and mixed data, and a professional (or tool) takes and “digs out” in these data precisely those that, with the “penetrating” look of a manager, open his eyes to everything that happens and suddenly it dawns on the righteous thought of the hidden reserves of the business model.
Miracles do not happen in big data either. To get valuable information from some storage, it must first be put there, then extracted, processed and visualized. The emphasis is not to shift to extracting information from the repository, leaving out of focus such things as collecting (receiving) data, structuring data, packing data into a repository, checking data quality, organizing the data analysis process, decision problems based on big data analysis and much more .
In addition, even for simple data mining, the correct goal setting will not interfere. Without a competent goal setting, anything can come out, rather than a meaningful result. Let this goal be expressed in the form of a hypothesis or questions, in the form of a problem situation or numerical indicators.
Any data has context and metadata that significantly limits its use in certain situations. If the context condition for the task is not specified, the analyst is not able to make a decision on the data mining and on the correspondence of the task data.
Despite the efforts of the business to reduce the time from the removal of information about its condition to the decision to change this state, there are objective reasons for an irresistible time lag .
The delay between making a decision and changing the state of a business in accordance with the decision made can also be quite significant. Processes and objects are rebuilt, the interaction is changed, the behavior of workers is adjusted, the environment is adjusted. Therefore, any data and even big data are always data about the past. But the management wants to make decisions on their basis for the future . Here the main thing is not to overestimate the possibilities of big data and analytics.
One of the misconceptions about big data is that they are mostly external to the business. It is believed that big data is data about customers (their behavior), data on competitors, data on various factors of business existence (political, social, cultural), data on markets and consumer trends, data on the activity of other businesses. Partially it is.
But big data for business from external sources is linked to data on the internal state , and they are linked strictly and contextually. It is extremely necessary to jointly assess the well-being of the business model and the external environment. Internal data can also be large and powerful for large, efficient analytics. After all, the answer to the question of what management can do to correct the situation can only be given by internal data .
Another illusion that can hinder business is that professional, productive analytics is based only on big data. There is a real opportunity and experience, multiplied by the talent of some experts, to offer solutions within the traditional volumes of internal data, especially when it comes to overt or typical problems in a business model.
It is impossible to deny the enormous importance of collecting and analyzing big data for business development. Especially important are big data for a distributed and information-active business. Perhaps big data is the only effective tool for keeping abreast of all matters for large corporations and associations with an extensive network of business units. Medium and small businesses can also benefit from big data, especially in cooperation with large companies and communities.
But it is impossible to substitute big data for solving urgent problems. It is better to consider them as a direction that supports the central business strategy and allows you to be aware of what is happening, what is happening, and partially predict the development of the situation in the future. But if a business does not have an intelligible strategy and if the business model is seen as primitive and confusing, then no big data can even help passive development. Some managers, understanding for themselves the absence of the need for big data and not being ready for the changes that they promise, do not try to initiate work with them - this is also an example of reasonable rational behavior.
No matter how hard we try, big data cannot solve all the problems. It is impossible to build an effective business model with the help of big analytics. But they can still help to optimize it within the framework of the chosen strategy.
The “magic” of big data, which somewhat remains aloof from the general attention, lies in the obvious and reasonable way of thinking about business and looking for ways to improve it . Indeed, a big data project improves business, not so much because of the value of some kind of information, but because management is starting to look at its business model from a critical point of view , including based on given information indicators and indicators.
If management is closely interested in big analytics, it means that it wants to understand more about its company and this is the beginning of business optimization.
Instead of big data, you can choose another means of business development, for example, marketing research, statistical calculations, economic and mathematical modeling. The result will be different, but work aimed at “understanding” the business model will be initiated and will undoubtedly give some, but more often - a positive effect. Unless, of course, it is carried out objectively, reasonably, professionally and taking into account the influencing factors.
Some companies have accumulated a resource - data , while others have developed powerful software and hardware IT solutions . They will try to bring businessmen to this table under this or that “marketing sauce” and earn “good tips”.
Active marketing and sophisticated marketing, polite consultants and cheerful client events richly flavored with a beautiful brand and impressive terminology will be used. They will talk about building reliable systems for processing absolutely unstructured data, about great algorithms for building multi-level information graphs, about high-speed samples on artificial intelligence, about self-learning neural network mechanisms.
Do not believe the word. Ask for clarifications, explanations, demonstrations, documentation, independent expert opinions, customer reviews, stress tests, a reasonable free trial period.
Judge for yourself, even the “big data” brand looks advantageous.
First, in the title there is the word “ big ”, which means it is something good, positive, beneficial, impressive, convincing, valuable.
Secondly, the word " data ", as it indicates something correct, intelligent, innovative, effective, orderly.
The very essence of the illusion of big data comes from their name.
It seems that having big data, the business solves the questions of the highest (large) order at the professional (large) level. And the more data to accumulate, the more efficient and faster will be resolved more complex issues.
Resist the illusions - big data is not always able to do what the business dreams of.
Maybe this is just a beautiful marketing course for developers of powerful computers and products for storing and processing digital data.
Maybe this is just an attractive advertisement for market research consultants and customer behaviors.
Or maybe this is just a fashion trend in the field of total observation of market entities and forecasting their reactions.
Perhaps there are no “big” data, but there is a big illusion that it will be possible to somehow gather such a huge array of digital information, process it in some magical way and get answers to all the questions of concern to a businessman.
A still from "The Special Opinion" (Minority Report) by Steven Spielberg based on Philip Dick's novel (2002 - 20th Century Fox, DreamWorks SKG).
')
... resource
Big data is essentially an analyst resource. This is a resource for people carrying out research and decision-making preparation. And like any other resource, big data does not work without the ability, knowledge and technology to use them. Someone calls this skill "data mining" (data mining) - by analogy with mining, focusing on deep penetration and complexity. Someone calls this skill “business intelligence” (business intelligent) - showing how important the “mental” component is in this process. Someone like the name "big analyst."
But it is known from theory and practice that even the presence of a resource in large quantities does not mean its successful and effective use . Sometimes, an excess amount of a resource allows you to build a business model, not deeply processing it into a specific set of products, but rather in simple packaging and sale in its raw form. Why look for additional options, when, without putting in excessive efforts, you can simply sell the raw resource.
... requires a high level of management
Big data, as an informational category, has one feature in contrast to material resources. To use them , a truly high level of organization of business objects and company business processes is required . Without such a level of preparation, without a certain qualification from the business, the purchase (or collection) of big data will be characterized by low efficiency. So low that it does not justify the investment in them.
Why should a business spend money on big data if a business management layer (decision making) is not based on analytics? Absolutely right - no need. To this or that extent, come those companies that have begun to use big data without adopting analytical technologies into the control loop, a technician who has prepared analytic decisions and who, by and large, are not ready for change. Such subjects of the market will sooner or later give up big data. Especially the issue will be acute with increased competition for financial resources within the business.
Today, the big data market has focused on information technology. It is understandable and pleasant that tools for working with big data are developing. But the intensive growth of information networks and the improvement of information technology removes barriers to computing power. This will force advanced ambitious businesses to reconsider their current hobby and shift the focus towards new effective methods, tools, and management technologies based on knowledge and training.
Actually, when big data is presented, it is often a question of the possibilities of their storage, transportation and processing. The search technologies of the giants of the Internet are a vivid example of how big data are given to businesses. Algorithms of search is the most powerful processing of the huge growing volumes of information. They are constantly in the process of optimizing, improving the performance of indexing and structuring information. But after all, behind the search technologies in the network are not only big data. Behind them are the teams of analysts who possess high-tech knowledge in the subject areas.
Therefore, a sensible policy of using big data is building a data analysis team , but in no way an exceptional alignment of servers, clouds, data mining systems, machine learning, etc.
... data mining
It is worth noting that the definition of “data mining” is not very significant. It draws a somewhat simplified picture of reality: there are “invaluable deposits” of heterogeneous and mixed data, and a professional (or tool) takes and “digs out” in these data precisely those that, with the “penetrating” look of a manager, open his eyes to everything that happens and suddenly it dawns on the righteous thought of the hidden reserves of the business model.
Miracles do not happen in big data either. To get valuable information from some storage, it must first be put there, then extracted, processed and visualized. The emphasis is not to shift to extracting information from the repository, leaving out of focus such things as collecting (receiving) data, structuring data, packing data into a repository, checking data quality, organizing the data analysis process, decision problems based on big data analysis and much more .
In addition, even for simple data mining, the correct goal setting will not interfere. Without a competent goal setting, anything can come out, rather than a meaningful result. Let this goal be expressed in the form of a hypothesis or questions, in the form of a problem situation or numerical indicators.
Any data has context and metadata that significantly limits its use in certain situations. If the context condition for the task is not specified, the analyst is not able to make a decision on the data mining and on the correspondence of the task data.
... time lag
Despite the efforts of the business to reduce the time from the removal of information about its condition to the decision to change this state, there are objective reasons for an irresistible time lag .
The delay between making a decision and changing the state of a business in accordance with the decision made can also be quite significant. Processes and objects are rebuilt, the interaction is changed, the behavior of workers is adjusted, the environment is adjusted. Therefore, any data and even big data are always data about the past. But the management wants to make decisions on their basis for the future . Here the main thing is not to overestimate the possibilities of big data and analytics.
... external and internal
One of the misconceptions about big data is that they are mostly external to the business. It is believed that big data is data about customers (their behavior), data on competitors, data on various factors of business existence (political, social, cultural), data on markets and consumer trends, data on the activity of other businesses. Partially it is.
But big data for business from external sources is linked to data on the internal state , and they are linked strictly and contextually. It is extremely necessary to jointly assess the well-being of the business model and the external environment. Internal data can also be large and powerful for large, efficient analytics. After all, the answer to the question of what management can do to correct the situation can only be given by internal data .
... big or not
Another illusion that can hinder business is that professional, productive analytics is based only on big data. There is a real opportunity and experience, multiplied by the talent of some experts, to offer solutions within the traditional volumes of internal data, especially when it comes to overt or typical problems in a business model.
It is impossible to deny the enormous importance of collecting and analyzing big data for business development. Especially important are big data for a distributed and information-active business. Perhaps big data is the only effective tool for keeping abreast of all matters for large corporations and associations with an extensive network of business units. Medium and small businesses can also benefit from big data, especially in cooperation with large companies and communities.
But it is impossible to substitute big data for solving urgent problems. It is better to consider them as a direction that supports the central business strategy and allows you to be aware of what is happening, what is happening, and partially predict the development of the situation in the future. But if a business does not have an intelligible strategy and if the business model is seen as primitive and confusing, then no big data can even help passive development. Some managers, understanding for themselves the absence of the need for big data and not being ready for the changes that they promise, do not try to initiate work with them - this is also an example of reasonable rational behavior.
... as a way to think
No matter how hard we try, big data cannot solve all the problems. It is impossible to build an effective business model with the help of big analytics. But they can still help to optimize it within the framework of the chosen strategy.
The “magic” of big data, which somewhat remains aloof from the general attention, lies in the obvious and reasonable way of thinking about business and looking for ways to improve it . Indeed, a big data project improves business, not so much because of the value of some kind of information, but because management is starting to look at its business model from a critical point of view , including based on given information indicators and indicators.
If management is closely interested in big analytics, it means that it wants to understand more about its company and this is the beginning of business optimization.
Instead of big data, you can choose another means of business development, for example, marketing research, statistical calculations, economic and mathematical modeling. The result will be different, but work aimed at “understanding” the business model will be initiated and will undoubtedly give some, but more often - a positive effect. Unless, of course, it is carried out objectively, reasonably, professionally and taking into account the influencing factors.
... under the big data brand
Some companies have accumulated a resource - data , while others have developed powerful software and hardware IT solutions . They will try to bring businessmen to this table under this or that “marketing sauce” and earn “good tips”.
Active marketing and sophisticated marketing, polite consultants and cheerful client events richly flavored with a beautiful brand and impressive terminology will be used. They will talk about building reliable systems for processing absolutely unstructured data, about great algorithms for building multi-level information graphs, about high-speed samples on artificial intelligence, about self-learning neural network mechanisms.
Do not believe the word. Ask for clarifications, explanations, demonstrations, documentation, independent expert opinions, customer reviews, stress tests, a reasonable free trial period.
Judge for yourself, even the “big data” brand looks advantageous.
First, in the title there is the word “ big ”, which means it is something good, positive, beneficial, impressive, convincing, valuable.
Secondly, the word " data ", as it indicates something correct, intelligent, innovative, effective, orderly.
The very essence of the illusion of big data comes from their name.
It seems that having big data, the business solves the questions of the highest (large) order at the professional (large) level. And the more data to accumulate, the more efficient and faster will be resolved more complex issues.
Resist the illusions - big data is not always able to do what the business dreams of.
Source: https://habr.com/ru/post/300342/
All Articles