For functional requirements, operations need to organize internal and external category email list user crowd testing, and focus on testing whether the process works. Feedback needs to be collected so that issues can be resolved in a timely manner to avoid issues with the official version. For requirements involving operational strategies, such as algorithm tuning. More category email list rigorous experimental tests, small-flow A/B tests, are needed to verify whether the effect is as expected.
Of course, the measurement of A/B test results also needs to be category email list comprehensively considered, such as whether the traffic distribution is random, whether it covers the entire population, whether the test time is sufficient, whether the results meet statistical significance, and so on. It is only after confirming that the demand realization meets the expectations that the closed loop of the demand is completed. Issues that the current version can optimize, that is, scheduling category email list optimization processing.
Problems that cannot be solved, if they have little impact on the category email list business, can be regarded as a legacy problem, and the next iteration will be completed first. Second, the demand sorting method Requirements There are three mainstream theoretical methods for current requirements ranking, KANO model, four-quadrant rule, and ICE model ranking method. The KANO model is more theoretical. The four-quadrant rule is a very practical method and is used category email list a lot in our daily work.