Similarly, why I have always been worried that the data center is a powerful medicine for enterprises. Because the data center overemphasizes itself and emphasizes the establishment of data capabilities for enterprises, but often seldom emphasizes the export and application of these data capabilities. Describe it as a master key, a panacea list of phone numbers for all diseases, a data hard currency that can connect to various demand scenarios. But the problem is that for most parties, they still urgently need to figure out what problems they want to solve At this stage, a large and comprehensive data center is not more targeted and better to use than a specific data system in a certain field, but it consumes more resources and has a greater risk of failure.
Ways to avoid pitfalls: Don't follow the trend. Before deciding to build a data system or data platform, you must do two things. First, before building any data system, figure out where the data is going to be used—do business consulting, whether it's internal "introspective" consulting or hiring external business consulting. Second, based on the business consultation in the first step, do a "data audit" to evaluate the existing data situation and the lack of data and data capabilities. Based on the two, it is possible to have serious planning and the possibility of establishing a successful data application system.
Data has many shortcomings. For example, data can never be comprehensive. There are some consumer attributes that you just can't get, and you won't be able to get these data just because you are on the so-called data platform. In many cases, it is the wisdom and efforts of operators who can really get the data, and the tool just keeps the records it should do well.
Data isn't always accurate, either. What it captures is data about people, and people are so fickle. Therefore, realistically speaking, data systems ultimately provide probabilities, and the good results you apply to these systems are also an increase in the probability of certain business results.
Data is not necessarily all so-called "assets". There is indeed too much data, too discrete, too random. Mining the value behind these data, there are many cases of no success.
Moreover, many, many times, the so-called big data is not more useful or better than small data at all. Big data is mixed and ambiguous, and small data is more accurate and high-quality; big data may have already passed its shelf life, but small data is more real-time and fresh.
The expectations for the application system should also be realistic. As mentioned earlier, the realization of many functions requires the satisfaction of many preconditions. The application scenarios for all functions are also very specific. Especially when dealing with the walled garden system, there are many more restrictions.
