Disney+ is the most-anticipated launch of a subscription streaming service since Netflix practically invented the category more than a decade ago.
More than two years after Disney first announced its streaming service plans, Disney+ is now out in the wild. But before today arrived, the company’s technical team was busy testing and retesting the platform to make sure it delivered a quality, consistent customer experience.
FierceVideo had a chance to sit down with Joe Inzerillo, chief technology officer at Disney Streaming Services, about one month before the Disney+ launch to talk about what the company was doing to prepare.
The following transcript has been editing for length and clarity.
FierceVideo: Disney+ is going to scale, probably really quickly. But you already have some experience through ESPN+ with handling big influxes of traffic. What are the key takeaways you’re applying to Disney+?
Joe Inzerillo: There’s a bunch of tech underneath that’s related to scaling, surge crowds and things like that. ESPN+ is a great example especially on a pay-per-view night with UFC or Top Rank where people are showing up en masse. You do whatever marketing you want to do, and we continue to get better at it but people are like, “I want to watch it now.” Because it’s a live event it causes this huge rally point where everybody, like 90% of the people are going to come in then no matter what you do. Which is not all that different from dropping an episode of “Game of Thrones” when we were running HBO. You drop that episode, everyone shows up.
The thing that those two things have in common is that they’re essentially appointment viewing, and they’re going to put demands not just on the log in but also commerce. Your peak time to view is also your peak time to sell. Everything really has to be humming there in order to get people into the product, whether they own it or not initially, and get them through the flow and into it.
Both of those things, and WWE pay-per-views, Major League Baseball Opening Day, NHL Opening Day, the Super Bowl, Eurosport, Bundesliga, EPL, things like that; all of those things have a lot in common. So, from a technology standpoint, we try to do as much as can always to make things scale horizontally. So, what can I do to limit the number of things that are either a single point of failure or a single point of scale reduction? I don’t want one big, honking database that everything has to go to because at some point I can only make a database so big.
Our system is designed to be asynchronous and very horizontally scaling, so we can bring on additional capacity and it just works. The system we have now, you’re always twiddling with it but it’s really our fifth-generation platform since 2002. We really have some impressive capabilities here, even beyond what we could do a few years ago.
FierceVideo: You’re surely expecting big peaks on day one for Disney+, but that’s something that could continually happen due to the service’s weekly episode drops.
Inzerillo: Totally. The time that it drops is a big aspect of it…But certainly the fact that there will be new content dropping weekly is going to create a lot of interest. How sharp the edge of that square wave depends on what time that you drop it. We’re prepared for all eventualities. When you’re going to get the next episode of “The Mandalorian,” I think that show could be as big as “Game of Thrones” for sure.
FierceVideo: That’s the one I’m psyched for.
Inzerillo: You should be! It’s really awesome!
FierceVideo: So, Disney+ has been in development for years.
Inzerillo: 18 months.
FierceVideo: Feels like longer.
Inzerillo: Believe me, it does.
FierceVideo: What are the priorities now that you’re within one month of the launch date?
Inzerillo: It’s a great question. I think part of it is taking feeds of data and learnings that we have from the Netherlands technical test that we intentionally put out there to get more data about how people are using the app, about how the technology platform is performing, about how individual clients are out there, about the state of certain devices that we might not see in the ESPN world since it’s a U.S.-only service.
There’s nothing like real data when you get out there. It’s like a battle plan. Whatever your plan is, as soon as it meets the enemy for the first time, plans are gone. There are so many variables out there of how people have things hooked up. There’s no substitute for real data. So, we’re taking all that data and we’re plowing it into tuning the last little parts.
I’d say overall in the Netherlands, we didn’t launch any worse in any category than industry average, and some much better. Now, everything’s much better. So, it’s just continuing to tune those dials to get it to where it’s the quality that we would expect from us.
Right now, all the infrastructure required for the U.S. launch is deployed already. So, that stuff’s out there but there’s still a lot of testing, a lot of chaos testing and simulated runs. You know, intentionally breaking things to make sure they fail over the way they’re supposed to. That type of operational readiness is really what we’re doing now going into launch along with a little bit of fit and finish.
It really is a cadence that you get into at some point in time. We’re very close to that time. We won’t be making any code changes. We’ll just be drilling operational readiness. We know exactly how the car is going to handle on the track. We know exactly what we need to look out for, if anything. We know exactly what surfaces we have to control once it’s out there. We feel really good about where we’re at.