Business Intelligence: real challenges

Business Intelligence: real challenges

During business operation, enterprises have been accumulating data in digital mode for more than 40 years. While accumulating enterprise data, tools that organize, classify, and analyze data and convert them into useful information for enterprise decision-making are constantly developing and improving functions. As for the supplier cases we talked about today, we can let everyone in the company have the information they need and make faster and better operational and strategic decisions, so as to satisfy customers and shareholders. The customer is satisfied because we know each of them and can predict what they need and what products and services they want. Satisfying shareholders is because customers are satisfied with the company's products and services. Therefore, the company's profitability will be enhanced. The reason we say this is based on reliable facts, such as good enterprise cases. Information-driven enterprises do exist, and their practices are also effective. Some of them do even better than other competitors in the market. In this case, should we make full use of all the data in our data warehouse and achieve the same high performance as these enterprises in information utilization? Is it time to realize "Enterprise Business Intelligence? It would be nice if it was so simple. For more than half a century, since Peter Drucker and others proposed the idea of an information-driven business, as data continues to accumulate, many practical problems have gradually emerged. For universal and effective business intelligence, three of the obstacles are particularly obvious: O few enterprises have consistent definitions of all the data they have accumulated. Therefore, data must be coordinated and organized before data analysis. This "primary data management" process lags behind business intelligence, and the tools it uses are not all so effective. The reason is that this is a very difficult problem. To eliminate the inconsistency between data accumulated for 40 years, the cost is very high. Due to lack of interest in the required professional knowledge, the vast majority of us (or our corporate sponsors) come to the conclusion that data is generally not worth cleaning and the computing cost is relatively low, even if there is inconsistency between the data, we need to process it. Analysts can always find these inconsistencies or inconsistencies. Second question: audience • We can come up with the argument that when business intelligence can be used by frontline staff and business managers at any time and is integrated with their work, enterprises can obtain the greatest value from business intelligence. People who contact your customers and major business partners need to make a lot of decisions every day, and their decisions will be better if there is better information for support. On the contrary, business intelligence is usually concentrated in a smaller group of specially trained "data analysts" who pass through complicated (and often less user-friendly) analysis of business performance, and then submit the "Report" to the business manager to take corrective measures afterwards or notify the strategic formulation process. Although this practice has some value, we do not intend to implement universal Business Intelligence along this route, because there is evidence that, this approach is totally different from the principle that we provide real-time decision-making support individually based on better information, which leads to our third problem: Information comprehensibility. • Ideally, information should be provided based on the actual background. The information provided should be directly related to the decision-making process and can be immediately used by the decision-making process. In this way, intelligence can have a real impact on such a decision. In today's business intelligence platforms, we rarely do these two things. The first requirement is that our business intelligence platform must have the "situation understanding" function in order to filter information relevance dynamically, in a complex and unpredictable manner. The second requirement is that the information provided by the platform must be customized in part based on the cognitive preferences of the information users. The information users can be anyone, including those who provide managed services for your enterprise, this type of person does not belong to your company, so it is not under your control. These are difficult problems for business intelligence platform suppliers. However, it is more difficult for enterprises to solve these problems in terms of organizational structure design, role design, and human capital management processes. Even if we can solve these problems, there will be more challenges. Although the data we aggregate and filter is what we need to support decision-making, the data may not be understood by users immediately. We may need to dig deep into the original information to fully understand the situation and make better decisions on this basis. For today's business intelligence tools, this is theoretically easy. Dashboards that provide summary information are directly linked to underlying basic data. However, in practice, this feature brings about some problems. How high is the computing and storage costs for unlimited "deep mining? If I publish all the content in the data warehouse, can I still manage security and privacy? If I find errors, can I fix them? (And are they really wrong ?) What is the impact of its views on the "correctness" of other people's data? Will the changes I made invalidate the decisions that I thought were made based on the correct data in the previous several seconds (or hours or days? If I want to Revoke Previous operations, how long do I need to record them? How can we avoid conflicts in editons with widely accepted truths in wikis? How can we mediate disputes? When everyone wishes to obtain the same information at the same time and this sudden information demand may result in a surge in information usage, how can I handle "hot spots" that are inevitable in the database and will happen on a regular basis? Therefore, Universal Business Intelligence is not easy. It requires a structured method. It needs to provide a better tool for the vast majority of users. It requires more than just a bunch of data and some analysis models. It requires us to ask ourselves: if everything in the business is determined by information, what role does the business manager and decision makers assume? When can people make decisions over information-based decisions? Should we monitor the decision-making process to better understand what information actually improves our decision-making and why it can improve our decision-making? Should we measure the ability of our business managers and colleagues to use information to better complete our work? Is "Business Intelligence blind" a new obstacle to career development? It is true that enterprises have made a commitment to rely on information (intelligence) to promote business development.