How far is enterprise from bi?

Previous Article <do enterprises need bi?> After the presentation, some netizens responded: Bi is still not popular; Bi is only useful for some large enterprises, and small and medium enterprises can only look at its back and be out of reach; Bi is just a little cute, if you really want to practice it, it will take a while and so on. I am confident that, in the warm autumn, a blizzard will suddenly fall, it almost wiped out all my enthusiasm. In this case, let's stay awake and think calmly: what are the challenges for enterprises to go to bi? What conditions must be met before the benefits can be realized and the benefits can be more valuable? The following is a brief analysis. The summary is not comprehensive. I hope you can enjoy it easily.

1, Influence of the company system.

Many companies, including private enterprises, collective enterprises, and some foreign-funded enterprises, have more or less problems in the company. Do more and do less, and do worse, more mistakes, more mistakes, more bad luck; otherwise, less mistakes, less merit, but not too much. Wages and bonuses are based on seniority. If you spend a year or a year, you will have a day in your early stages. In this way, why Bi is more than one thing. On the contrary, if Bi has a history track of the company, it finds how low the previous efficiency is; how rugged the previous Road is; how absurd the previous decision-making was. Leaders can't hold on their faces. Do you still have to mix them up? Save your time and enjoy yourself by stealing food and chatting ". Bi is cool, where to stand.

In the company, the style of the Parent-Teacher System is prevalent. The words of the superior, especially the middle-and lower-level leaders, are the sacred purpose. The instructions of the leadership are a beacon on his way forward. In the long run, leaders feel better and better about themselves, and experience is becoming more and more intense. No matter what tools you use and how thorough the problem is, they still feel that they have no direct experience, they do not feel trustworthy. In this way, scientific decision-making is often replaced by the leadership consciousness. Do you still say that Bi is useful?

2, The impact of data quality.

The first is data accuracy. The data accuracy will directly result in the accuracy of the analysis results, leading to the correctness of future decisions. Although many enterprises have been implementing informatization for many years and many information systems have been introduced, how high is the data quality? How high is the data accuracy? As the saying goes: a thousand miles away. In the Bi system, although there is not such a big impact, his data volume is generally relatively large, there are differences between some data, it is inevitable. However, if the difference is too far away, it is too dangerous and may bring disastrous results to the final decision-making. For example, is an enterprise investing in a real estate market in the current environment? Investing in stock markets? Investing in real Industries? Or deposit in a bank to buy a national debt? Internet, TV, newspaper, and overwhelming information. We recommend this information today and tomorrow. We recommend this information for the east. The East says that the housing market has a large appreciation space. The West says that the stock market is low and has potential. But do you believe this? Is your collected information comprehensive? Is it reliable? The front page of Sina often shows "it's better to work as a small boss". Is that true? Such information is a test of the accuracy of your data. I think that CNPC was the most profitable company in Asia in the first place. If you get a piece of cake at the opening, I don't know how long it will take to turn over?

The second is data integrity. The information system of most companies now records the final results of some things, such as the sales quantity, sales amount, financial income, and cost. It can be said that it is comprehensive and accurate, but how is the result data obtained? Why are the benefits of this year better than those of last year? Why is product A sold more than product B? Why is a's market better than B's? What decisions and actions have the company made for these results? What are the effects of the current market environment on these results? Many companies may lack this part of "Process Management". As a result, the value of the analysis results is greatly reduced. As time passes, personnel changes may occur, the scenario has changed. Over the past few years, the performance was good. Why? What are the benefits? Why was the initial performance poor? Almost unknown. I will continue to feel like I am making similar decisions in the future. For example, in a large store, its information system records the product types, names, quantities, unit prices, amounts, and information of some customers they sell each day. However, many marketing activities have been carried out in this process, such:

1) store decoration improves the shopping environment.

2) training and competitions on "Enhancing service awareness and improving service quality" were conducted.

3) Various advertising activities, such as street signs, websites, newspapers, televisions, and broadcasts.

4) various promotions including discounts and gifts were held.

5) The current economy is in good form, the stock market is soaring, the effect of making money is obvious, and the CPI coefficient is rising.

And so on. The time, cost, scope of impact, and duration of these activities are not recorded in the system or not detailed enough. Think about it, if the data is recorded in detail in the system, and there is still a lot of quantitative data to show the effect of each activity, then analyze and mine, the final value of the results I have come up with will be greatly improved. In the future, we will launch activities and advertising, which will be more targeted, so we will not burn money without any need.

3. difficult requirement sorting.

Bi is an analysis and management tool for the company's middle and senior leaders. Its needs naturally come from the middle and senior leaders of the company. These groups not only need to know the company's internal information, such as whether the indicators are achieved, whether the company operates well, and whether the company's business strategy needs to be adjusted; they also need to pay attention to information outside the company, such as the overall international economic situation, the development prospects of the industry, and the dynamics of competitors. Therefore, their needs are much more complex and easily changed. Often, a demand model has not been built yet, and new ideas have emerged, in addition, every leader has his own unique ideas and ideas. It is very difficult to unify these ideas.

With the promotion of Bi applications, the demand of leaders is also increasing. They not only hope to explore new business growth points from historical information, but also open up some new channels. They also want to turn this platform into a battlefield for their sandbox deduction. They can simulate various situations, analyze the change trend of the final results, and try to find an optimal mode. For example, for a manufacturing enterprise, the ratio of inventory to inventory is the most appropriate, the capital turnover rate is the highest, the capital volume is the most active, and the benefit is the best, which may be learned a lot, there are also many factors related to procurement costs, production capacity, sales status, and current economic form. If we can simulate various situations, analyze the results, and then use the dregs to extract the essence. The quality of the decisions made in this way will be much higher. If we do not take every step, we will cross the river by feeling the stones. Of course, this is another realm of Bi. Relying on the writeback function in cube, You can implement some requirements and make some preliminary deductions. If the number is too large, the performance is poor and the efficiency is too low, it is not a good method. We are still exploring this part. If anyone has a good idea and shares it with us, thank you first.

4. high cost.

1) hardware and software investment. The investment in general information systems is tens of thousands, and the investment is hundreds of thousands. If the investment is tens of thousands, the investment in the information system is very small. However, BI may be much larger, with millions and tens of millions being even more common. This investment generally includes two aspects: software and hardware. Software: it includes data integration (ETL), data analysis, data mining, and front-end display tools. If it is a mature system, it will be almost millions, our company was originally preparing to introduce Oracle to the company's company, because the price was too high, and finally had to give up. In terms of hardware, if the company has accumulated about 10 years of experience and the number reaches TB level, it may not be too easy to use two to three 16-core CPUs and 64 GB memory servers, coupled with cross-regional network costs, this is also a huge expense.

2) labor costs. Includes three parts:

A) The implementation cost of software vendors is really worth their technology. After a system is implemented, the total cost is much more than the software cost. If the business is very wide, the business is very complicated, there may be more.

B) Investment costs of the company's IT staff. Most companies basically have their own IT teams. When the suppliers implement Bi, they closely follow up, one is to help them provide some data interfaces to introduce business processes and other work; the other is to learn about machine learning. When the implementers leave, there will be minor changes to their needs, you can also handle it temporarily. It is also a common practice to implement bi for a year and a half. Although it is not worth the money, it is also a cost.

3) Cooperate with business personnel. As long as Bi involves the business, it is generally necessary for the business director or lead to cooperate, from organizing the requirements, confirming the requirements, Results acceptance, they have to participate in each step. Leaders are the pillars of the company. They are busy and time-consuming. It is not easy to let them sit down and discuss it over and over again. However, I have to say that Bi is intended for middle-level supervisors. It is their future crutches. They are not responsible. Who is responsible! They have to work hard.

Having said so much, it seems that the road to Bi is really a long way to go! Comrade Xiaoping said that some people are allowed to enrich themselves first. Conditional and determined enterprises can take action first and finally drive everyone to take action together. There may be still hope.