Multiple choice (all correct options required) · #635
An e-commerce company noticed that the repeat purchase rate among new customers within three months of their first order dropped from 38% to 31%. The marketing department concluded: "New customer quality has deteriorated, and we should reduce advertising spend." Using the five-step inductive method to evaluate this, which of the following approaches or judgments is most reasonable?
Show answer & explanation
Answer: ABC
- AFirst define the problem clearly: in which channels, product categories, and first-order amount brackets is the decline in repeat purchase rate most pronounced, rather than vaguely stating "new customer quality has deteriorated".✓ This corresponds to the first step of the five-step inductive method, "clarifying the problem." The more specific the object, time frame, and metric, the more reliable the inductive starting point. This approach is reasonable.
- BGroup new customers by channel, category, and first-purchase behavior for comparison, focusing on the segments where the decline is more concentrated, rather than only looking at the overall average.✓ This corresponds to the fourth step, "seeking patterns." Grouped comparison helps distinguish systematic trends from random fluctuations. This is reasonable.
- CSeparately report new customers acquired during promotional events or those who purchased due to one-time discount coupons, rather than mixing them directly with regular daily data when calculating repeat purchase rates.✓ This corresponds to the third step, "cleaning data." Removing or flagging incomparable data prevents short-term promotions from being misinterpreted as long-term quality changes. This is reasonable.
- DData on new customers accumulated over the past few years exceed one hundred thousand, so the sample size is large enough; therefore, the decline in repeat purchase rate directly proves that the ad targeting is imprecise.✗ This is a common misconception that equates large sample size with reliable conclusions. The key to sample design is covering key differences and excluding incomparable factors; size alone cannot replace data cleaning and pattern seeking.
Explanation:The marketing department's conclusion compresses a complex phenomenon into a vague statement of "deteriorating new customer quality." The correct approach is to first define the problem clearly—specifying channels, categories, and time periods; then group comparisons to find patterns of concentrated decline; and also separately account for incomparable data such as promotions and coupons to prevent short-term activities from skewing judgment. The option that claims a large sample size directly proves attribution overlooks the need for sample representativeness and data cleaning in the five-step inductive method, which is a common mistake.