You are a data analyst tasked with building a comprehensive test success tracker.  Create a chain of prompts designed to analyze the success of A/B tests, focusing on key metrics and providing actionable insights. The chain should consist of three distinct stages:</p>
<p>**Stage 1: Data Aggregation and Cleaning:**  This stage should prompt for the input of test data, including metrics such as conversion rates, click-through rates, and A/B test variations.  The prompt should specify the expected format of the input data (e.g., CSV, JSON) and instruct the AI to clean and standardize the data, handling missing values and outliers appropriately.  Include an example of the expected data format.</p>
<p>**Stage 2: Performance Analysis:** This stage should use the cleaned data from Stage 1 to calculate key performance indicators (KPIs) for each test variation.  The prompt should specify which KPIs to calculate (e.g., lift percentage, statistical significance), and should generate a summary table showing the performance of each variation.  The prompt should also include instructions for identifying the winning variation based on statistical significance and practical importance.</p>
<p>**Stage 3: Actionable Insights and Recommendations:** This stage should analyze the results from Stage 2 and generate actionable recommendations for future tests.  The prompt should focus on identifying patterns, explaining the reasons behind the success or failure of specific variations, and suggesting improvements for future iterations.  The output should be in a concise, easily understandable format, suitable for presentation to stakeholders.  Include examples of the types of insights and recommendations the AI should provide.</p>
<p>The overall output should be a comprehensive report summarizing the test results, including data visualizations (charts and graphs) where appropriate, and providing clear, actionable recommendations for future experiments.  The entire chain should be designed to be easily repeatable and adaptable to different types of A/B tests and data sets.