Interviewer: Hello everyone, today we have invited Song Libao to join our CDA licensee interview! Say hello to all of you! 

Guest: My name is Song Libao, I graduated from Beijing University of Technology, and I am currently working at Yiren Hongsheng Technology Development Co.

Question 1:What does your position do specifically? Can you introduce an actual case to us?
Guest: 1 output daily operational reports, such as daily report of self-contracting performance, daily report of team ranking, daily report of seat performance ranking, etc.; 2 output daily analytical reports, such as daily report of incentive tracking, daily report of performance commission warning, daily report of single month assessment, etc.; 3 target management, according to different data sources, participate in the development of incentive programs to promote marketing, and follow up the output results and target achievement for review and analysis; 4 project data support: various types of stage data, such as the data of the project, the data of the project and the data of the project. 4 Project data support: various types of phased project data support.

To give an example of data support for a project: one project needs to clean a wave of newly arrived data sources, and the goal is to determine whether these data can be used for daily marketing. To provide data support, the very first step is to communicate with the project side, which needs to understand the entire project operation process, what the goal is, what data the project needs, what indicators need to be observed, and finally produce data reports containing various process indicators and target indicator values as required. The report is based on the business requirements, and the report needs to be updated and improved.

Question 2: usually too many operational data reports, how to batch capture data?
Guest: Every day we need to contact a lot of reports, of which 30/40% of the reports data temporary reports, is the CDA to explain the topic of the briefing type of reports, there is no fixed cycle, his demand point from the business sector or leadership of the temporary needs. This kind of data capture data is mainly through SQL from the database to find, excel processing. About 50% of the reports are the daily notification and periodic review reports explained in the CDA, of which no urgent time-sensitive reports can be directly run manually by SQL, and for the higher real-time requirements directly to technical colleagues to give the demand for them to develop online automatic reports, which can meet the real-time update observation. The last 10% or so of the data is non-data data, for this some database data through excel for processing.

Question 3: What are the management difficulties and key points of telemarketing call out, and what problems can be solved by using data analysis?
Guest: Difficulties: the list strategy is more complex (user operations team of data for a variety of categories, pre-credit / loan / closing, new customers / old customers, according to the time dimension for classification, the operational process dimension for classification, data classification is still relatively cumbersome.) For the case of a larger volume of data, you can use Pareto's Law to find the top few data lists that account for eighty percent of the output performance, and focus on their observation. If the problem is on-site, you need to track the process indicators of the team's seats and analyze the reasons for the problem, whether it is caused by the dialing strategy or by the seat's conversation. If it is a strategy problem, you need to communicate with the data upstream to see if there are some changes that lead to the overall change, and what kind of impact this strategy adjustment will have on the overall performance.

Question 4: What data analysis methods do you use in the telemarketing process?
For example, the sales funnel model (from the customer being marketed to the customer entering to the app operation and finally to the successful release of funds, which involves several operation processes), from the different stages of conversion rate, the funnel model can help us to recognize the type of customers more clearly, and also help us to recognize the opportunity point, which link needs to be further improved, so as to ultimately improve the performance. further improvement, so as to achieve the ultimate purpose of improving performance. For example, if the credit rate of a data source is higher than other data sources, it means that the customer has demand, but his lending rate is lower than other data sources. By comparison, we find that the piece rate of this data source is higher than other data sources, so we assume that the customer has no desire to withdraw money because the piece rate is too small. By A/Btest to see the results, if the results and predicted the same adjust the logic of this data source of dialing. For the sales team, we can use target ratio analysis, same ring analysis to find the problem point, and according to the root cause analysis to further discover the problem point and output problem solving solutions.

Question 5: If you want to receive an offer from your department (position), what abilities or skills should you have? 

Guest: In fact, for my position, the skills need to be excel, not that I have used it can be, in which the use of data visualization to be skilled, some common functions to be very skilled (find class vlookup, match, index, etc., summary class sumifs, countifs, etc., logic if, iferror, etc.). sql (basic add, delete, change, check) to the database data tables, table relationships have a certain understanding, otherwise the number of runs will be inaccurate, will python or BI words better. Need to have some statistical foundation, need to have some data insight and good logical thinking skills. There is a certificate of data analysis class, high you in front of the interviewer good feeling. In fact, the skills can be learned in the CDA course.