Optimizely suits up for battle with Adobe Target

Last fall, Optimizely announced it was moving beyond its previous role as an A/B testing company for web and mobile and would henceforth be an experimentation platform.

In this new incarnation, the platform could be used for testing any connected device, in any part of a technology stack (such as code, data flow or visual design) and for any product (e.g., email, website, mobile app, over-the-top app).

In April of last year, the San Francisco-based company bought Austin, Texas-based Experiment Engine, which offered tools for project management and reporting.

Today, Optimizely said it has now integrated Experiment Engine into its platform, which it is launching in beta as Program Management. The new capabilities, the company says, allow it to compete against an enterprise testing platform like Adobe Target.

Program Management is designed to help enterprises better plan, organize, manage, report and share their experiments. Optimizely counts 21 of the Fortune 100 as its customers and reports that its platform has supported over 200,000 experiments in the last 12 months.

Previously, COO/CFO David Schwarzbach told me, management functions were standalone through the connected Experiment Engine. Now, they are integrated into the platform, and there is a much greater ability to scale the number and complexity of experiments.

“A year or two ago,” he said, “our product enabled teams to experiment with a throughput of 200 to 300 experiments a year.”

“Now,” he added, the enhanced platform can let an organization handle “thousands of experiments a year.”

This includes a new capability to segment users by a testing element, the ability to more easily track ideas for testing and feedback from experiments across an organization, more extensive reporting and automated features to quickly scale up an experiment.

Previously, he said, machine learning was employed only in Optimizely’s recommendation product, but now it’s been added to the core platform to help generate insights more quickly. If there are five variations of something, the machine learning helps find the winner.