One-click install
npx skills add https://github.com/delta-and-beta/braze-agency --skill analyst-testing
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: analyst-testing
Source: https://github.com/delta-and-beta/braze-agency/tree/main/skills/analyst-testing
Command: npx skills add https://github.com/delta-and-beta/braze-agency --skill analyst-testing

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes references (resource) components.

What problem does it solve?

Braze-based experimentation is complex: teams need structured design, robust analysis, and reliable interpretation to drive trustworthy campaign improvements. This skill acts as a guide to design, execute, and reason about A/B and multivariate tests, including how to choose optimization strategies and how to relate results to platform-wide holdout concepts like global control groups.

Core Features & Use Cases

  • End-to-end experimentation lifecycle guidance: pre-test design, test configuration, during-test monitoring, and result interpretation.
  • Support for Winning Variant and Personalized Variant optimizations across multiple channels.
  • Synthesis of topic references to load atomic knowledge at runtime, keeping the knowledge graph navigable.
  • Global control group planning, random bucket usage, and race-condition awareness to contextualize lift and significance.
  • Analytics-driven decision support: leveraging Braze's statistical tests and confidence metrics to drive action.

Quick Start

Outline a basic A/B or multivariate test plan for your Braze campaign and configure the initial variants.

Frequently Asked Questions about analyst-testing

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I design an A/B test in Braze to optimize campaign outcomes?

To design an A/B test in Braze, outline your campaign variants and apply experimental design principles to configure pre-test plans, optimization strategies, and conversion tracking for reliable results.

How does statistical significance work when analyzing Braze A/B test results?

Braze A/B test analysis evaluates results using chi-squared testing and 95% confidence metrics to determine statistical significance and drive actionable, data-backed campaign decisions.

Can I run multivariate tests across multiple Braze channels like push and email?

Yes, multivariate testing in Braze supports single-channel campaigns across push, email, webhooks, and canvases, allowing you to compare multiple variables and optimize outcomes simultaneously.

What is a global control group in Braze and when do I need it for experimentation?

A global control group in Braze is a platform-wide holdout used to contextualize campaign lift and significance. You need it when measuring true incremental impact across random buckets.

What is the difference between Winning Variant and Personalized Variant optimization in Braze?

Winning Variant optimization sends the best-performing variant to all users, while Personalized Variant optimization tailors variant selection to individual user preferences and behaviors.

What are the limitations of Braze A/B testing regarding race conditions?

Braze A/B testing requires race-condition awareness during global control group random bucket usage to prevent overlapping exposures that can skew incremental lift and significance measurements.