Q&A with Jeff Haynie and IDC’s John Jackson On the (Near) Future of Mobile

Earlier this week, we announced the 2015 edition of our Mobile Trends Report, which surveyed nearly 6,000 mobile developers on the state of mobility, and the mobile readiness of their organizations. It’s a big undertaking and one we couldn’t do without our partner, IDC, and especially John “JJ” Jackson, the firm’s expert on all things mobile.

To tease out some of the findings a little further, we sat down with JJ and our own Jeff Haynie (Appcelerator CEO and co-founder) to get their views directly:

What are the biggest opportunities and challenges for mobile developers going into 2016? Are they different for indie developers versus commercial or enterprise devs?

JJ: The current mobile environment is full of multiplier effects for developers. As developers begin to target and code for more platforms and devices, they are encountering more design challenges and more opportunities to create net new value across workflows. The traditional cross-platform challenge of timing and resources remains, but it seems logical that as we move to mobile mainstream, developers—indie and enterprise—will continue to run into challenges inherent with wrangling backend data not optimized for mobile. The challenges revealed by this dynamic came through loud and clear in the new survey.

I’d expect enterprise developers to focus more on contending with backend access challenges than indie devs. There’s definitely less flexibility on the enterprise side, and devs working for larger companies will encounter the backend access issue more often. However, there is no single formula for mobile or tablet, or need for additional functionality. The resulting dynamic is challenging for the enterprise because many businesses take a while to shift approaches.

Jeff: An emergence of new devices and screens is creating opportunities and challenges for developers contending with more noise and peer activity every week. The Apple Watch is just one example of a companion device to phones and tablets with its own set of technology and design challenges. There are many more and they’re unlocking opportunities and creating challenges, pushing devs to learn and perform constantly.

The notion of “data-less” apps is over; including in the gaming and enterprise ecosystems. There are very few apps today that are not connected to some backend datasource. Typically, they’re connected to more than one. Apps with offline functionality are still reaching into data and sharing it through notifications, moving context. There’s a huge challenge for devs working with companies that haven’t yet invested in MBaaS. Meanwhile, there’s an emergence of cloud-only, full-stack apps and tools that help the mobile leaders while challenging laggards to keep up.

I certainly agree the cross-platform challenge isn’t going away. Each of these mobile platform owners—Apple, Google, even Microsoft to a smaller extent—are aggressive, deep-pocketed, and have a lot to win or lose in the push for ecosystem innovation. There are more and more opportunities to disrupt entire industries. Indie developers also have a constant stream of opportunities but they need to compete with other indie devs to own categories and deliver innovation. On the enterprise side, there’s a seemingly endless demand for more and better apps. We’re still in Day One in mobility as a whole, and within the enterprise, I think it’s still early morning of Day One.

What is the role of enterprise IT in a world gone mobile? App development? API enablement? Project management? Vendor selection? All of the above?

Jeff: I talk to a lot of global CIOs and CTOs. One recently told me that trying to control everything is not important or efficient to conducting business, especially IT. In fact, he said that the IT industry needs to get back to roots of automation across the enterprise. We’re seeing movement from the notion that IT’s going to govern everything—to say Yes or No to every conceivable project—to one where IT’s mission becomes enabling tech innovation around the organization. Exposing APIs—basically, easily accessible RESTful endpoints to backend systems—is one of the best ways of doing this, in my view. And I see plenty of organizations waking up to this as well. Maybe I’m overly optimistic, but I really believe IT’s relationship with the C-suite is changing in a positive way.

JJ: The evolved charter of mobile development is one of service to the organization, rather than a buck-stops-here/be-all-end-all destination for the lines of business. There is a fair amount of DIY going on in development, with a prevalence in backend with custom web services available. In fact, DIY initiatives were off the charts among developers surveyed—increasing from 40% in 2013 to 75% in 2015.

Given how much more complexity there is in frontend and backend development for mobile (versus, say, the days of web and Wintel), there’s been an understandable desire for some kind of silver bullet. HTML5 and more recently RMAD have attempted to fill that role on the client-side, neither entirely satisfactorily. Is there a silver bullet for mobile development?

JJ: There’s no silver bullet for anything ever, but there is a need for tooling to approach challenges from a full stack perspective. Focusing on the weighted average of release cycles of our surveyed developers unveils some interesting data points. In fact, when we parse leaders among the group, they are consistently investing aggressively in tools for all facets of the release process. In the 2013 survey, the average release velocity was every 104 days. In 2014, the number decreased to an average of one app every 92 days. This year, we’re back up to an average app release velocity of 107 days. While this may not appear to be a great trend, it can be attributed to more emphasis on tools. Leaders in 2015 are releasing apps every 55 days. Meanwhile, laggards are taking their time and releasing apps every 201 days. The gulf in release velocity difference and the correlation to aggressive use of frontend, backend, and analytics tools is not a coincidence.

Jeff: We’ve seen this picture before in the ’90s when devs were focused mostly on frontend and on static websites/presentation. In the early 2000s, developers moved on to the backend and spent significant resources to strengthen middleware; moving from mainframe to the mid-tier computing approach. Subsequently, it became all about pulling data to meet and make sense of the immense scaling of web traffic—moving from single to multi-source and the web. Today, we’re seeing a rise in platforms and apps centered around connectivity, dynamic play and backend data to bring forward into experiences. Developers across the stack are charting new ways to design for the supercomputers in people’s’ pockets.

When I was at CSX (one of the first customers for Java), we were building our own tech platform just as Netscape introduced its app server. That forced us to ask ourselves why we were spending time to build a native platform when there was now a commercial offering available. Each company goes through a similar cycle with every paradigm shift in the stack. There may never be a single, silver bullet tool that’s available, but that certainly doesn’t mean the answer is to try to build your own!

There’s more and more talk of “full stack” development and developers: what is this and what’s the reason for the trend?

Jeff: I’ve been using the term “full-stack developer” for a number of years, so I’m glad to see it’s becoming more common in the mobile development discourse. Historically, we’ve been operating under artificial distinctions between elements of development. If you look at process of each layer in a more traditional stack, there’s clearly overlap and complementary work being done. But those layers evolved mostly independently, each for its own realm and adopting its own preferred languages. This is the reason most companies have organized their technology teams around each layer.

Well, I think mobility is beginning to unwind this. The need for velocity is just too great—and the expense of carrying a bunch of dedicated organizations too high—to persist in a bunch of distinct “tribes” for each stack layer. In the future, there will continue to be delineations among developer models, but the “full-stack” developer will be significantly more common. Just as we’ve seen role compression between previously distinct identities like “designer” and “developer,” we’re increasingly seeing the same kind of compression between “frontend” and “backend” developers. Arguably the biggest enabler of this is JavaScript and the rise of Node.

This role compression—and the power and efficiency of running your stack on a single language—is a big reason we rallied around JavaScript from the outset. If you’ll forgive the CEO blowing his own company’s horn, we have patents dating to the mid-2000’s around JavaScript for the backend. And of course with Titanium, we pioneered the JavaScript-to-native cross-platform mobile development that so many others are now trying to adopt. Being able to leverage the same language and framework on front and backend provides tremendous capabilities. There’s less back and forth among silos and dedicated specialists; you can leverage the same people and skills, which drives down costs and latency in a huge way.

Based on the survey’s findings, what are your predictions for mobile development in 2016?

Jeff: On the full stack issue, the tech stack can now lend itself to greater process or role compression, but only with the right tools. If you have that in place, you can move fast. If not, you’ll need to leverage things like MBaaS—which increasingly is a subelement of a full microservices architecture—to address data in a more deliberate and optimized way. I think the full-stack push will drive languages like Swift to the server-side as well. It wouldn’t surprise me to see Swift become used for projects outside of iOS.

But generalities are a little easy, so I’ll give you a specific prediction. One year from now, I expect the number of folks targeting custom web services for backend data to fall by at least a third. [See page 4 of the survey.] That’s a legacy model that just doesn’t work in a world of connected “things.”

JJ: I’ll see your specific prediction and raise you one: IDC’s own data suggests that enterprise mobile apps will quadruple by 2018, and—this is the kicker—60 percent of those apps will have no PC antecedent.

Mobile development is the future and a knowledge of the full stack will be crucial to optimizing release velocity, pushing the boundaries of technology and advancing the value offered by apps to consumers.