Aug 02, 2022
In Welcome to the Forum
For this kind of question, the judges generally ask more divergent questions, and some will ask questions about a certain part of the information, such as asking why the data in the middle of the promotion materials fell, what happened, and some more email list general questions, such as "I encountered a problem during the process. What is the problem and what action was taken to deal with it.” There is a pit that is particularly easy to step on, that is, after the previous goals and methods are set, the later execution gives people a feeling that the results are easily obtained. All the process data performance is very good, and the results are not bad. Is it right? Pass me? In fact, if this happens, the jury thinks that there are two possibilities. The first is that the logic of what you are doing is too simple and there is no challenge. The second is that this is not easy, but you are not using the correct one. A data-driven and fast iterative approach to doing things simplifies the process. Both of these situations are not good. The former will make the judges feel that things are too simple and it will be difficult to judge your personal ability, and the latter will make people feel that you are not using the correct working methods. To put it more seriously, you can obviously use the correct method to get a bigger result, but because you didn't use the correct method, even if the result you get now achieves the goal you set before, it shouldn't, because you missed it. for greater value creation, which you are supposed to do for the company in your place. Therefore, the key to answering such questions is to reflect the process of "scientific iteration". In actual work, we have no way to make a comprehensive prediction of the project beforehand. It is often the way of "preset + experiment + data feedback + adjustment". . Let’s take the previous example, assuming that the goal is to increase the conversion rate, and subsidizing SKUs and reducing prices is one of the strategies to increase the conversion rate. Then, around SKU subsidies, it is unlikely that we will know which SKUs are subsidized and how much is to meet the goal. And it is efficient, which is gradually "optimized". Another example is when we make a page revision, there are several designs, we also test different styles with small traffic, and decide which revision form can be expanded according to the data feedback. Therefore, whether it is promotion materials or answering questions, it is very important to be able to reflect some "scientific iteration" process, the current premise is that the actual work is done.