This post is part of a new series called “The Future of Modern Software” and is brought to you by New Relic. Read the series here. As always, VentureBeat is adamant about maintaining editorial objectivity.
While some burgeoning companies may have the opportunity to build with a mobile-centric focus from the ground up, there are still those out there that began back when building Web-first was the way to go. For these companies, the move to mobile brings about a whole new set of issues when it comes to monitoring application performance and integrating thate data into their development practices.
When social shopping site ModCloth launched back in 2002, the term “Web 2.0” had yet to become popular, and the first iPhone was still five years out. Ten years later, ModCloth had been named one of the fastest growing private companies in the country, and the move to mobile was an obvious next step. Early in 2013, it released apps for iPad, iPhone and Android.
ModCloth CTO Udi Nir says that going mobile meant rethinking some things.
“In transitioning to being a fully mobile-first organization, we had to really think about how it would affect the rest of the business and how we approached new features and experiences,” says Nir. “Our customers have a very personal connection with their devices, and the initial experience after installing our app has to be excellent. To that end, we are very deliberate about testing prior to releasing our apps so there can be a slower frequency of updates, with each release going through much more testing. Monitoring how the apps behave once they get out in the wild among consumers is also an integral part of our release cycle.”
Application performance management (APM) for mobile applications is a relatively new development. SaaS APM provider New Relic, for example, just released mobile monitoring for iOS and Android in early 2013. However, organizations are already finding the capabilities very important for success with mobile apps.
“We use APM to inform our decisions about capacity planning and where to target our optimization efforts,” says Nir. “Identifying bottlenecks is a basic part of running high traffic services. But identifying these bottlenecks is made much easier with tools that collect the right data and allow us to analyze and visualize that data.”
Nir explains that monitoring mobile applications is an entirely different ballgame than an app run on their own servers.
“When releasing mobile apps to users, the apps runs on their devices as opposed to the ModCloth servers. Our code is running in environments that are variable–for example, available memory, network bandwidth, latency and more,” says Nir. “This makes it more challenging to understand how the application is performing and what the experience is for our users. It’s critical for us to have visibility into the user experience and the root causes of any issues, and while we can get some information from our API servers, that’s only part of the equation. We need to know what’s happening at all levels of the stack: DB tier, API, and client applications.”
According to Nir, the move to mobile has affected ModCloth in more ways than one. Going mobile has made the company rethink how it approaches its customers and forced them to interact in new ways.
“ModCloth is built on a social shopping model that engages our community both with the brand and with one another, so we’ve had to think about how this social experience could flourish on mobile and offer a rich experience. Our goal is to provide the best shopping experience for our customers on whatever platform they choose to use, and in optimizing for mobile, we’ve been able to expand the social features that we offer to users,” says Nir. “For example, we recently launched a feature on iPhone called Fit for Me, which enables a customer to input her measurements and then sort through real customer reviews to find products most likely to fit her. Our mobile development has enabled us to create a better user experience for our customers, with more ways to engage.”