What are the differences between the data service of the third party company and the data service of party a?

Abstract: Article source @ Weibo UDC Design Center is now a big data era, everyone's mouth is hanging data to create value, data mining and other hot words. Within the company also gradually recognized the importance of data, have set up a data unit, expect data can really

Article source @ Weibo UDC Design Center is now a big data era, everyone's mouth is hanging data to create value, data mining and other hot words. Within the company also gradually recognized the importance of data, have set up data departments, expect data can be truly business services.

In addition, there are a number of third-party companies to do data services continue to emerge, hoping to help generate data to share the burden of data analysis, mining more valuable laws to help party a continuously improve the level of business, and constantly find problems in the business. From this point of view, party A and third party data services company's original intention is the same.

So in the end of the Third-party companies to do data Services and party a company to do data services are there any differences? In the light of the experience of the third party company in the last few years to do the data, the difference between the two data is summarized briefly.

1, the pursuit of different

There are roughly several modes of cooperation between the third party company and the party A on data service:

1 long-term monitoring flow data-regularly provide daily, weekly and monthly reports;

2 temporary project-receive the party a company's demand, initiate the project, in the stipulated project cycle, summarizes the project research result in the form of the report.

No matter which way of cooperation, the third party analyst in the process of analysis, always lonely, always more dependent on data. Because not in the analysis of the environment, do not know what the operation is doing recently, do not know what the product has a plan, a word: do not know the other side of the point of attention in the end.

Party a company to do the internal data, in fact, the cooperation is similar to the above listed two kinds. But the details are different:

1 before the study will be detailed understanding of the project background and product or operator confusion;

2 The data can not be explained in the study, the relevant personnel feedback could be found at any time. Analyze the causes of data anomalies with your business colleagues.

3 After the study will report the results of the study, and business colleagues to discuss the next step of the improvement strategy and plan, the next time to agree on the study point.

Therefore, subject to the conditions, although the original intention is consistent with the company, but due to the lack of in-depth contact with the business, so the ultimate goal can only stay in the completion of a rigorous, professional data analysis report. As for the follow-up, How to use the report, how to improve the business, improve the business and so on a series of follow-up work is unknown. As a result, the output of third-party companies is always unable to form a closed loop throughout the business chain.

But this is certainly an unavoidable thing. Believe that most companies, especially Chinese companies, the company's internal data privacy is still more important to the Third-party company's attitude will not be completely open. Therefore, the cooperation between the two sides is limited to the relatively shallow level of co-operation. Third-party companies want to really go through the business closed loop, at this stage is completely impossible.

Over time, third-party data analysts are accustomed to making a sophisticated report the ultimate goal.

Attachment: True feelings during the transition phase

Just from the third party company to enter Sina Weibo to do the first project, the project report that day was the affirmation of the product personnel. I was very happy at the time and felt that my work was recognized. Thought that the product colleague all approved, certainly the leadership also will feel good. But the result is quite the opposite.

The group leader was puzzled by the fact that I didn't have any follow-up plans for the next step, and from our serious conversation I deeply realized that the company data in party A is really for the product improvement or operation service. If your findings only stay in the summary and collation of the problem, the work is basically only 50 points. The doctor took X-rays for the patient and then ignored the patient.

2. Different presentation

The value of the project is reflected in the report for the cooperation of the third party company and the party a company, so the report should be made in a beautiful and "value-for-good" way. Only a few pages of PowerPoint will never be the result of the final output. It would be more valuable to assist in the study by establishing a complex model to help illustrate the results.

And inside the company, everyone is eager to know what I can do to see the results of this research. If the model is too complex to be understood or overwhelmed by the product staff, it is meaningless. On the contrary, the study of a specific problem, even if only a few pages, a few data, will also make the product staff very excited.

For example, occasionally from the data to see a number of phenomena or problems, at this time to do a simple collation, print out a few pages can go and products, operating students to chat. To see how the line of business colleagues are looking at this phenomenon, is there some important operational strategies that affect some data results, resulting in abnormal data results, rather than real problems. If there are no other factors affecting the impact of this phenomenon, then we really need to project to spend time to find the cause of the problem.

3. A sense of achievement manifests itself in different

Third-party data companies, the core business is to collect data, analysis, and therefore responsible for the production of data reporting data analysts, in contrast, the results of the work can easily be seen. Therefore, also very much the company's attention.

And in party a company, data service is a function line, it is for product and operation Personnel Service. Perhaps the embodiment of work value can only be reflected after the whole business chain is completed. Even after the whole business chain, how to assess the role of data in the whole process is also a difficult task. However, as a data analyst, to see their own analysis, to help products or operations to identify problems, so that the product experience or operating mechanism has been improved, this sense of achievement will be exciting for the analyst, or will be enthusiastic to rush to the next project.

4. Different understanding of industry standards

Third-party data companies, due to a number of companies in the service industry, will put together the same business modules of the same company, an industry standard. Provide industry standard data to customers without disclosing their business secrets. So that each family can clearly understand their position in the industry, to understand which data looks good on the surface, but compared with industry standards, in fact, the situation is not optimistic. I think that's the biggest value of third-party data companies.

However, the current situation is that China's internet industry, we are still very close to the data, do not want too much to let the third party intervene. As a result, there is no industry standard. The results of the data can only depend on experience.

To sum up, the difference between the two data analysis work, give me the experience is: data analysis is not the focus of data, but how to really solve the actual problem. The ultimate value of a data analyst is not how many statistical tools are used, how many data models can be mined, but the real business. Seeing the results of the data can tell which businesses are having problems, and seeing business problems makes it clear what data can be learned about the cause of the problem. Of course, if, in large terms, companies are willing to work with third-party companies to build industry standards through third-party companies, that will ultimately make the data more valuable.