Strategy
A team that can run experiments and a team that gets value from them are not the same team. The second one has answered a set of questions the first has not: which ideas are worth testing at all, what a hypothesis has to say before it earns traffic, how many tests can run at once without contaminating each other, and what to do with the large majority of experiments that will not win.
This section is about those decisions — the programme around the tests rather than the mechanics of any one of them. It covers prioritising a backlog, designing an experiment that can be interpreted afterwards, migrating from a tool that has been shut down, choosing between platforms on something firmer than a feature grid, and the qualification a test has to pass before it is worth anyone’s time.
The articles below assume you already know how to launch a test and want the programme around it to be worth running. Read them in any order; each is written to stand alone.
Articles in this section
Testing AI Models with Feature Flags: LLM Prompt Optimization
A/B test AI models like GPT-5.5, Claude Opus 4.8, and Gemini 3.1 Pro with Optimizely feature flags to optimize prompts, cost, and quality without redeploys.
A/B Testing Hypothesis Template and Examples for Optimizely
A reusable A/B testing hypothesis template with examples, plus how to store it in Optimizely native Hypothesis and Test plan fields.
Optimizely vs Amplitude: Experimentation Compared
A balanced, practitioner-focused comparison of Optimizely and Amplitude Experiment: experimentation models, statistics, analytics depth, and how to choose.
Industry Experimentation Playbooks: A/B Testing by Vertical
Vertical A/B testing playbooks for e-commerce, SaaS, media, and travel: the metrics that matter, test ideas, and how to pick your first experiment.
How to Prioritize A/B Test Ideas: ICE, PIE, and PXL
Rank your test backlog with the ICE, PIE, and PXL frameworks. Learn how each scores ideas, how to cut bias, and how to check you can power the winner.
Running Multiple A/B Tests at Once: When Overlap Is Safe
Can you run multiple A/B tests at the same time? Usually yes. When interaction effects matter, and how to set up mutually exclusive experiments in Optimizely.
Optimizely Experiment QA Checklist: Validate Before Launch
A repeatable pre-launch QA checklist for Optimizely experiments: validate variations, targeting, assignment, and events, with the right tool for each check.
Migrating from Google Optimize to Optimizely: A Complete Guide
A practical guide to replacing Google Optimize with Optimizely after the sunset: map every concept, rebuild experiments, audiences, metrics, and GA4 reporting.
Experiment Design for A/B Testing: A Practical Framework
Use experiment design to define trustworthy A/B tests: unit, treatment, control, metrics, sample size, analysis plan, and Optimizely configuration.
Website Personalization: A Test-and-Learn Strategy That Proves Incremental Lift
Build a website personalization strategy: choose audiences, match experiences, preserve a holdback, run tests, and measure incremental lift.
Price Testing: Running Controlled Experiments on What You Charge
Run a price experiment on revenue per assigned user, keep assignment sticky and the shown price honoured, and separate randomised price testing from personalised pricing.
Subscribe
Practical Optimizely tips, monthly. No fluff.