基于手机银行App的多层分流对比实验方案

Multi-layered Controlled Experiment Solution Based on Mobile Bank App

AB实验起源于生物医学实践中的双盲测试,后广泛应用于互联网行业,用于对多个方案进行比较,从而选出最优方案。提出基于手机银行App的对比实验方案,充分利用手机银行的渠道优势,通过多维度的标签组合以及多层级的流量分发,保证各个AB实验间互不干扰且实验内各组用户具备相同的统计特征,实现客户流量的多次利用。同时,结合前端组件,通过埋点记录客户行为数据,采用漏斗模型进行分析,得出各实验组的点击率、转化率等产品指标。该方案可用于产品推荐、用户界面(User Interface,UI)设计、广告方案以及产品功能等业务场景中,从而帮助产品经理选择出较优方案,提高流量的转化效率和产品研发的迭代效率,实现数据驱动业务增长。

AB experiments originated from double-blind tests in biomedical practice and were later widely used in the Internet industry to compare multiple solutions to select the optimal solution.We propose a comparative experiment scheme based on mobile banking app,which makes full use of the channel advantages of mobile banking and ensures that each AB experiment does not interfere with each other and each group of users within the experiment has the same statistical characteristics through multi-dimensional label combination and multi-level traffic distribution,so as to achieve multiple utilization of customer traffic.At the same time,combined with the front-end components,customer behavior data is recorded through buried points and analyzed using a funnel model to derive the click-through rate,conversion rate and other product indicators for each experiment group.The solution can be used in business scenarios such as product recommendation,user interface(UI)design,advertising solutions and product features,thus helping product managers to choose the best solution,improve the conversion efficiency of traffic and the iteration efficiency of product development,and realize data-driven business growth.