Many students are confused when analyzing problems: how to do deep excavation? How far will you dig? A bunch of indicators have highs and lows, how do you see them? Today, I will share with you the system, how to dig deeper into the problem. 1. Level 0 Depth The simplest reason analysis is to make a subdivision comparison in a single dimension. For example, after analyzing why the performance did not meet the standard, and subdividing it, it was found that 2 of the 5 branches did not meet the standard. Then the analysis conclusion is: "Because the two branches did not meet the standard, the performance did not meet the standard"! Of course, at this point, you can do one more step: quantify the impact of each branch that does not meet the standard. For example, see who has a low compliance rate and who is a drag on the broader market.
After this step, the single-dimensional subdivision comparison is complete. 2. Level 1 depth Since single-dimensional segmentation can be done, multi-dimensional segmentation can also be done. The most common is to pull the cross table. For example, why is the performance not up to standard? Pull out the dimensions of products, users, branches, etc., and cross them with the performance goals, and then list the differences in each dimension one by one. Many students are doing problem analysis, but they are actually doing this... Note that one more step can be mobile number list done here, that is, to discover the connection between several dimensions. Because it is very likely that the three branches of A, C, and E do not sell well, and there are the same reasons behind them: they are the main sales places for A's product, and A's product is rotten.
This connection can be through two dimensions Cross-comparison findings. 3. Level 2 depth However, the matter is not over yet, we will still ask: Why do nail products fail? Is the congenital quality of this product not good enough, or is the later operation not in place? Notice! At this point, the problem changes. Whether it is "innate quality is not good enough" or "later operation is not in place", there is no way to directly measure it with an indicator. At this point, we need to convert the hypothetical reasons described in language into quantifiable and verifiable indicators. For example, "innate quality is not good", there are two points here: what is quality why not Quality can be measured from the perspective of hardware such as product performance and configuration (requires second-hand data collection), or from the perspectives of user experience and word-of-mouth (requires research/public opinion data collection).