游戏用户获取推广策略效果评估与优化数据集合

数据清洗规则：采用3σ准则剔除曝光量、转化率等指标中的异常值，对缺失的渠道成本数据采用同类型渠道均值填充，确保数据完整性；归因算法：改良Shapley值模型，引入“转化贡献权重系数”（依据游戏行业转化周期设定），解决传统归因中“最后触点独占功劳”的偏差问题，精准计算各渠道贡献占比；效果评估规则：构建“CAC-LTV-留存率”三维评估矩阵，将策略划分为“优质（低CAC、高LTV）”“待优化（高CAC、中LTV）”等4类等级；数据更新规则：渠道实时数据每小时增量同步，日度生成转化与留存汇总表，周度更新竞品对标数据。Data Cleaning Rules: The 3σ criterion is adopted to remove outliers from indicators including exposure volume and conversion rate. For missing channel cost data, the mean value of the same type of channels is used for imputation to ensure data integrity. Attribution Algorithm: An improved Shapley value model is developed, which introduces the "conversion contribution weight coefficient" (set based on the conversion cycle of the gaming industry) to address the deviation issue of "the last touchpoint monopolizing attribution credit" in traditional attribution models, enabling accurate calculation of the contribution proportion of each channel. Effect Evaluation Rules: A three-dimensional evaluation matrix of "CAC-LTV-Retention Rate" is constructed, and strategies are categorized into 4 grades such as "High-quality (low CAC, high LTV)" and "To-be-optimized (high CAC, medium LTV)". Data Update Rules: Real-time channel data is incrementally synchronized hourly. Daily summary reports of conversions and retentions are generated, and competitive benchmarking data is updated on a weekly basis.