Faster, easier, and more scalable: meet Optimizely X Web Experimentation

Silvia Amtmann
Design @ Optimizely
5 min readNov 9, 2016


By Silvia Amtmann and Zach Leach, Senior Product Designers

Over the past year and a half, we’ve been hard at work on the newest version of Optimizely’s A/B testing product, which we introduced as part of Optimizely X: the Experimentation Platform. It’s called Optimizely X Web Experimentation. To mark its introduction to our customers, we’re sharing the reasoning that went into redesigning Optimizely to be a flexible framework that grows with your experimentation organization.

The need for a change

Web experimentation has changed dramatically in the 6 years since Optimizely was founded. Instead of individuals managing optimization programs on top of their day jobs, the companies at the forefront of experimentation are now dedicating teams of people to running bigger, more complex experiments than ever before. While Optimizely Classic worked well for early testers, it didn’t work as well for teams, which are forming more and more of our user base.

We leveraged 5 years of user research and customer feedback to establish three principles for building Optimizely X: scalable, streamlined, and reusable. Whether you’re an individual or part of a big team, these changes help you experiment faster and collaborate better.

Save time by streamlining your workflow

Early experimenters spent a lot of time in the crown jewel of Optimizely: the visual editor. As the product grew, the editor felt like a natural place to house all of the different features for customization and management of experiments, but the editor interface struggled to contain them all. As lone optimizers became teams that increased the number, velocity and complexity of experiments, users started telling us that managing all aspects of an experiment through the editor wasn’t ideal. Critical aspects of an experiment, like metrics or audiences, would get forgotten, and team members would trip over each other to get their respective changes done.

Optimizely X has tackled this problem in two ways: we designed an experiment creation workflow that allows you to configure the critical elements of your experiment up front, and also created a new variation management view to handle experiment configuration before you ever enter the editor.

Configure the critical stuff upfront (if you want to)

In Optimizely Classic, all you needed to start an experiment was a name and a URL. That’s all you need in Optimizely X, too, but we give you the option to configure audiences and metrics up front as well. Whether you’re an individual or team, having these commonly-changed properties top-of-mind during creation should help prevent them from being forgotten.

For people collaborating with others on experiment creation, there’s an additional benefit to this new design. One person can easily create and configure any number of experiments and have them ready and waiting for another person to actually build the experiment variations, a pattern we saw in large teams’ use of Optimizely.

The right UI for the task at hand

The architecture of Optimizely Classic was simple: the home screen, the editor and the results page. But as Optimizely grew, we realized that the architecture we had wasn’t easily extensible. In Optimizely X, we’ve created a separate area for experiment management and experiment setup so that we can give you the right tools at the right time.

Separating management from editing has some added benefits. The list of element properties on the left-hand side of the Variations Overview doubles as checklist. Did you remember to set your goals? How about that one-off experiment integration? Just run through the list and check to make sure everything looks good before publishing.

The new experiment manager makes it easy to make changes without ever loading the editor.

Another benefit of removing all that management UI from the editor is that we have room to give you a list of changes you made to a variation. It’s an easy way to make sure you made all the changes you intended, and you don’t have to load a page in the editor for anything except actual editing.

The editor is all about editing now, which leaves us room for great features like the variation changes list.

Whether you’re a lone experimenter or a big team, the new experiment creation interface supports you throughout the creation process, and is always tailored to the work you’re doing when you’re doing it.

Establish consistency and scale

To help you do your work more efficiently, we wanted to find ways to reduce the amount of repetitive setup required to set up experiments and manage them across your team. Reusability became a key theme of Optimizely X.

As you might expect, audiences and metrics are some of the most common experiment properties that people set up and track in Optimizely Classic. The URL on which an experiment runs, however, had to be specified from scratch every time. It’s not that big of a deal if you’re experimenting on simple URLs, but when you have several variations on a URL, a list of URLs, or a multi-page experiment, repetitive configuration becomes annoying and can lead to mistakes.

In Optmizely X, URLs are now created once, and can be added to an experiment like any other reusable component. We call these URL templates “Pages.” Just like audience and metrics, there’s no more one-off configuration and tedious multi-page test setup for you to manage every time you make a new experiment.

Reuse all those URLs you’ve configured by making them a Page.

These are just some of the changes we’ve made to put greater emphasis on the instrumentation of your website. Overall, our goal is to make running experiments faster, more consistent, and easier, even for large teams.

A flexible framework to grow with you

As experimentation at your company becomes bigger and more sophisticated, Optimizely is planing for the future before you get there. Scaling Optimizely X means we have to handle more experiments, more channels and more kinds of users, and leave enough room for the product to grow and change with you.

Ultimately, Optimizely X exists to help you be brave and experimental, to learn from your customers, and to use data to make quick and informed decisions. We look forward to seeing the many experiments you run using Optimizely X Web Experimentation, and to your feedback as you transition to the new platform.

To learn more about Optimizely X, visit Optiverse and ask us your questions.



Silvia Amtmann
Design @ Optimizely

Product Designer, book-club organizer, and data vis nerd at Optimizely.