In this interview, Jason Arbon, the CEO of Appdiff, digs into the best way to build your mobile app quality strategy. Whether you have a web, hybrid, or native app, building a quality and testing strategy means knowing what data and tools you have available to make agile decisions.
Jennifer Bonine: We are back with another virtual interview at STARWEST. I am so excited. Jason's one of my favorite folks to have on our interviews, so we're lucky he's back again this year. Thanks, Jason, for being here.
Jason Arbon: Thanks, Jennifer. It's good to see you.
Jennifer Bonine: For those of you that don't know out there, why I'm so excited for this is because Jason started a company called Appdiff. If you out there have not heard of Appdiff, you need to listen up and hear what Jason's doing and what his company's doing, because it is so cool. If you have mobile apps that you're running out there for your organizations, this is someone you should be aware of and be talking to. Jason, maybe give us a little overview of Appdiff.
Jason Arbon: Thanks, Jennifer. Yeah. Appdiff was really painful to build, but it's super easy to use. Basically we're tackling the two biggest problems that mobile apps have today, well app teams have. One is measuring the performance of their application. That's really difficult. Most people barely have a handle on that. The other big thing they want to do is speed up their agile cycles, their lean kind of ... that CI and that kind of world. They get more builds than they can deal with.
What we do is we basically build a bunch of little robots that live in the cloud. All you have to do is upload your app and that's it, or you can tell us where it is in the app store and we even work with just that. We do a complete performance analysis of your application and show you where the slowest pages are, where the fastest pages are. We basically spoon feed it to you. There are no charts and graphs. You can get those, but really we just show you the picture of the slowest part of your app to motivate people to actually fix the stuff.
Secondly, we do a basic regression automation suite. What happens is the robots look at everything inside your application. We'll check between ten and thirty thousand data points on average for every build. Then the magic happens when you give it the second build. In the second build it does the same thing, the robots spin up and they run around all through your app. We choose machine learning and fancy things like that to make it look more like a human going through the thing, to click on smart things, not dumb thing. Still not brilliant yet, but they're good. Basically they're a five or seven year old kid. That's what we describe them.
Jennifer Bonine: Yeah. You can have your kids test this stuff. Sometimes they're smarter than adults now with tech.
Jason Arbon: Right. I don't know. I'm comparing it to my children, so it may be.
Jennifer Bonine: Your children are probably advanced.
Jason Arbon: No! I don't know. I don't know. They're advanced in bugging me and stealing my iPhone. Basically the idea is that it runs through the second build the same way and with ten to thirty thousand data points, and then it does the difference. The cool thing is it tells you if anything got slower or faster in your app. It also finds out if anything new's been added. The worst thing as a tester is a new build comes out in the morning and you don't know if there are things to test. If something is missing, it'll also tell you that. It'll catch regressions. They don't know what the right behavior is. I don't pretend that they do. Machines are not taking over the world yet, but they just at least point testers within fifteen minutes of a build to stuff that changed.