Optimizely started as an A/B testing company with a clear identity and has since assembled a large digital experience suite through acquisition, absorbing a content management system, a commerce platform, a data platform and a content marketing product along the way. The academy reflects that history and not entirely happily, because training a coherent discipline and training six loosely joined products are different problems and this academy is trying to do both. The experimentation content is where the original company's expertise survives and it is the reason to be here. It is good on the things that actually determine whether an experimentation programme produces value or produces noise.
Sample size and statistical power, why you cannot peek at results and stop when the line looks good, the multiple comparisons problem that arrives the moment you test several variants or several metrics, the distinction between a statistically significant result and a practically meaningful one, and the discipline of forming a hypothesis before rather than after looking at data. Anyone running tests without this knowledge is generating decisions from randomness with a confidence interval attached, and that is a worse position than not testing at all because it feels rigorous. The training takes this seriously and I would recommend it to someone running experiments on any platform. The rest of the suite is where it thins out.
The CMS, commerce and content marketing tracks are essentially structured product documentation. There is nothing wrong with them as such, an editor who needs to learn the CMS will learn the CMS, but there is no discipline being taught underneath, no equivalent of the statistical grounding that makes the experimentation content valuable independently of the tool. You finish knowing where things are in an interface. The orientation problem compounds this.
A newcomer arriving at the academy faces paths for Agent Platform, CMS, Commerce, ODP, Product Recommendations, Experimentation, CMP, Optimizely One and Analytics, with limited guidance about which of those they need or how they relate. Product naming across the suite has also changed repeatedly, which means older content refers to things by names the current interface does not use. Somebody who has just been told their company bought Optimizely and they are now responsible for it will have to work out the map themselves. The recent repositioning around agents and AI is the weakest part.
The messaging has moved fast toward agentic capability and the training has followed, and it reads as a layer applied on top rather than something integrated into how the products are taught. This is a common pattern in 2026 and it is usually a sign that a positioning decision arrived before the pedagogy caught up. I would expect this material to be substantially rewritten within a year and I would not invest heavily in it now. The pricing opacity is a genuine irritation and worth flagging as a signal rather than just an inconvenience.
Account creation is free and gets you into the eLearning catalogue. What certification costs, or what the instructor-led sessions cost, is not stated anywhere I could find, and the academy directs you to email an education address to find out. Vendors that publish their training prices do so because the prices are defensible. Vendors that route you to a sales conversation are usually pricing by customer size, which for an individual learner trying to decide whether this is worth it is a bad experience and an easy reason to walk away.
Three point seven. The experimentation training is worth four and a half on its own. The average is dragged down by a sprawling catalogue, unpublished pricing and a suite that the academy has not managed to make coherent.