Robots, Autopilot, and The Holy Grail of WebOps

Today, I’m thrilled to announce that Pantheon has acquired StagingPilot, creating a clear pathway for us to offer automation for website updates through our WebOps platform. This acquisition includes not just StagingPilot’s outstanding technology, but also founder Nathan Tyler and his brother Phil, who have created the gold standard of automated CMS maintenance.

Currently, StagingPilot automates over two million test steps that would otherwise be done by humans every month. Impressive. We can’t wait to get to work integrating this amazing capability with the rest of Pantheon.

Robotic Eyeballs To Check Your Website

StagingPilot is pretty incredible. It monitors your sites for updates to WordPress core or any plugins. It then creates a pre-production copy of the site, applies the updates, and uses a sophisticated visual regression testing process to determine if the updates are safe to deploy. All with zero upfront test-writing required.

The value of visual regression testing is huge. We’ve been making the case for it for years. It’s a great way to catch flaws in details that human eyes can easily miss, and also key to hitting the turbocharger of continuous deployment for websites. The challenge is that visual regression testing is complex to set up and often costly to maintain, which means that only a rare few projects can really leverage it.

StagingPilot changes all that. It creates test plans programmatically, on the fly, including functional tests on top of simple page comparisons. It also handles common false positive results, and can automatically set tolerance thresholds so that human operators are only engaged when there’s something meaningful to look at. Otherwise, updates deploy seamlessly.

A more hype-inclined person might call StagingPilot AI for website testing. We’ll get there someday for real (really, we will), but for now, I’m happy to call it zero-configuration visual regression testing. That’s groundbreaking in and of itself.

StagingPilot has already proven it can scale. Together, we’re going to take it to the next level. Integrating this service with Pantheon will allow us to democratize access to visual regression tooling, free developers all around the world from the toil of testing websites for routine updates, and save untold thousands of site owners from snafus when untested updates go live.

Upstream Autopilot: The Holy Grail of WebOps

One of the special things about Pantheon is the way our Website Operations Platform helps customers who are managing lots of sites. We have unique tools like Multidev, Upstreams, Quicksilver, and Terminus which can make developers up to four times more productive when working across a large number of sites. You don’t have to take my word for it: Forrester Research recently confirmed this in their independent research report on the Total Economic Impact™ assessment of the Pantheon Platform.

That’s huge. Making it sane and sustainable to manage CMS at scale provides a way out of the megasite vs. multisite catch-22 that confronts so many folks when thinking about their web presence. Customers can choose to adopt WordPress and Drupal as CMS platforms, versus having to think about “one big install.” This is enormously liberating.

But there’s still a key piece missing. Once you have a hundred sites (or heck, even twenty), the grind of keeping up with routine updates can be daunting. Our existing WebOps tools let our customers automate a lot of that maintenance, but building and managing that automation is on them.

That’s annoying. We want our users to automate their operations, not operate their automation. I’ve always personally considered providing a turnkey autopilot feature to be a “holy grail” for our product. With StagingPilot, that’s within reach.

What’s Next: Executing the “Tesla Playbook”

Our roadmap for StagingPilot integration follows what I like to call the “Tesla Playbook.”

The first phase will be to leverage their technology for our existing Managed Updates capability, where we directly guarantee that customer sites are kept up-to-date. This is a high-end, white-glove service—the “Roadster,” metaphorically speaking—and integrating StagingPilot will mean extending their technology to support Drupal as well as WordPress.

Once this is ready, we’ll move on to the “Model S” phase: using economies of scale and optimizations to bring the cost of Managed Updates down to where it makes sense for a larger group of customers. This will still be a premium service, but we’ll be on our way to scaling it broadly.

Things really get exciting when we fully integrate StagingPilot with our existing upstreams and organizational features and are able to offer it to customers directly, and as part of the partner program. StagingPilot has already proven they provide great value to agencies, and we’re confident that this will also be true for others managing significant site portfolios. This is “StagingPilot for the People”—the “Model 3”—which we expect to ship early next year.

Further into the future, there’s an opportunity to connect real user behavior—plus Google’s applied machine learning capabilities—with the power of StagingPilot in order to create something you could truly call “AI-driven testing” (with a straight face). We’ll also keep working to scale the service and offer it to more and more customers. I’d like to eventually include a baseline level of service as standard with every site plan because there’s really no reason to run an out-of-date CMS.

We’ll have more updates to come as we make progress, but it’s really exciting to be taking off on this journey. Maintenance risk and toil is probably the single biggest impediment to open source adoption, and I can’t wait to slay that FUD and make the open web more peoples’ go-to platform for delivering results.