Glueing eyeballs to apps - profitably
Shakespeare drew in his theatre audiences with low humour and high poetry from 1594 onwards. In the 20th century, soap operas became a television staple, watched by millions of viewers. More recently, sales and marketing types discovered potential in extending enterprise applications to mobile devices.
For all these audiences, a satisfying personal experience is the main reason they keep coming back for more, be it comedy, drama - or finding the closest hamburger served up by a global brand. Also, audiences have always been fickle. Consumers can write off a clunky enterprise mobile app by flicking a finger. Even business-to-business or internal apps (say, to apply for leave and have it approved) need to be up to scratch.
Making canny decisions about extending enterprise applications to mobile require business strategy, technology strategy, including application lifecycle management, a modern IT architecture, and, above all, a well-designed user experience.
To create those experiences on mobile devices, Web developers are flocking to an emerging standard that also promises minimal amounts of coding. That standard is called HTML5, regarded by many as the next revolution in rich mobile Web development. It is currently being adopted by most desktop browsers, and RIM and Windows on mobile, but Android and iOS are lagging behind.
A big reason developers swoon for HTML5 is the concept 'code once, use on many mobile devices'. Another is enabling a mobile Web app to figure out what kit a device has - the screen size and resolution, for example - in order to select and format content in a user-friendly manner for that device in real-time, even if its brand and model are unknown.
The best thing enterprise can do is wait and see while the garage developers innovate.Richard Pierce-Jones, 3Fifteen
Opposite to the HTML5 mobile Web approach is coding a 'native app'. Here, developers code a native app for each mobile device, using its software development kit. Doing this requires the developer to know each device's capabilities really well. Generally speaking, this approach is required to get the app into a proprietary mobile app store.
Due to HTML5's immaturity, however, developing several native apps can currently be far less work than a HTML5 mobile site, depending on business requirements.
"We tend to be anti-HTML5 for mobile, even though general consensus is HTML5 will kill the native app," says William Brander, architect at solution developer BBD. "It's not because our customers don't want to pay less for application development. It's because their users are driving the market. The client the mobile app is written for does not dictate if it will be successful - it's the guy with the phone who does."
HTML5 will take some time to mature, as will adoption by major stakeholders. While that happens, new points of exploitation in security will be created (unwittingly) in different operating systems. Finding these unknown security risks will demand extensive testing of HTML5 developments, says Richard Pierce-Jones, GM for mobile at solution developer 3Fifteen. Meanwhile, native apps coded by developers who know the devices are far more likely to be secure, even though more coding effort is involved.
"The best thing enterprise can do is wait and see while the garage developers innovate," continues Pierce-Jones. "These guys will have all the pain. Enterprise can then jump in at the top of the R&D curve, buying, owning or adopting technology.”
"The role of enterprise is not to innovate," he continues, "but to cherry-pick innovation and adopt it. The decision whether or not to pick this cherry is still a year away. HTML5 compliance is not mature enough for enterprise to be seriously considering it for application development now."
Businesses will have to trade off the richer user experience and better security of native apps against the broad device coverage and fast time to market that HTML5 mobile Web enables. They'll have to weigh up less expected development effort using HTML5 versus the emerging standard's dependence on consensus from many stakeholders, and its vulnerability to unknown security exploits.
The organisations need a modern, loosely coupled IT architecture to even get started on extending enterprise applications to mobile in an orderly fashion. But entrenched, modern enterprise applications generally lack Web-generation interfaces, so Web developers encounter older interface standards when extending to mobile.
Getting a profitable 'wow' from enterprise applications' end-users extended to mobile will remain a strategy balancing act for some time.
As good as Angry Birds
On his rounds to customers, Michael Chaize from Adobe starts by asking three questions to figure out how ready they are for extending enterprise apps to mobile devices. Chaize is the company's developer evangelist in EMEA. Key sections of the HTML5 standard are being developed by Adobe and other industry players.
The client the mobile app is written for does not dictate if it will be successful - it's the guy with the phone who does.William Brander, BBD
"Firstly, how do you access your information, the content for the mobile app? Are you using a modern Web architecture? For a large organisation, a service oriented architecture and Web services need to be in place.
"Secondly, how do you structure the information in the back end?" asks Chaize. This becomes critical when a desktop app's mobile version needs to present far less, more focused information to the user on a small screen.
Chaize describes a global hamburger franchise using analytics, finding out that mobile users just wanted to have one question answered quickly: 'Where is the nearest place to get my hamburger?' The user experience on the main Web site was changed, so if the Web site is accessed from a mobile device, the user is asked if he wants to be geo-located and pointed to the nearest outlet.
"To do this, you need well-structured content," continues Chaize. "This is why you need a content management system. For a particular mobile user, you just want to send one specific piece of content, because it makes sense in his context. But for desktop users, you want to send lots of content to the bigger screen.
"Sometimes, large organisations don't structure their content in small pieces of data with relationships between them. Sometimes they just push a whole bunch of content to mobile devices, and that's very hard to manage and display on small screens."
The third question is something Chaize always asks when he visits customers: "Do you have designers?"
"By that I don't mean visual designers, but do you have people in your team who are in charge of the user experience?" says Chaize.
"Unfortunately, in large organisations, designers are used at the very end of the process. I always try to push a design-driven methodology: define the whole user experience on mobile and tablet and desktop devices, and then from the visual experience, create your architecture and start coding.
"Lack of user adoption is very dangerous, especially on mobile and tablet devices. User expectations are very high in B2C (business to customer), and for B2E (business to employee) apps. Employees won't understand why they have to interact with a bad user interface on a B2E app if they can have an excellent user experience on the next icon, like Angry Birds.
“I always make sure that large organisations create a team of user experience consultants for internal applications, not only for the corporate Web site.
"Large organisations can take a year to choose a technology, but you can choose the best technology and still fail if you don't have user experience designers involved," concludes Chaize.
Sniffing for tools
Picking mobile app technology was certainly on the minds of local companies at a recent global event.
South Africans asked lots of questions at the February 2012 Mobile World Congress in Barcelona, says Ken Parmelee, senior director of product management at Antenna Software.
The company's mobile enterprise application solution (MEAP) was rated as 'visionary' in an April 2011 Gartner Magic Quadrant, ahead of SAP-Sybase (also visionary) on completeness of vision, but a bit behind Research In Motion and Syclo in terms of ability to execute. Currently, Antenna deals only with global companies that have a presence in SA.
"There was a tremendous amount of focus on SA and other African states that I have not seen before: from start-ups in SA that want to spin up an applications business, through to mobile carriers that want to provide services to African and South African markets.
"In a lot of regions, company execs say, 'I want to build a whole bunch of applications'," continues Parmelee.
"But if you don't think about how to make the business grow, both in your choice of tools and the strategy you go to market with, it becomes, 'I just built a whole lot of apps,' not a sustainable business. The intelligent thing about the South African approach is that it avoids the app-build mindset, and goes for building a whole mobile practice encompassing a go-to-market strategy," he says.
That strategy has to deal with many fluctuating technology variables, one of which is the recent revolution in Web development using HTML5. Web developers are using HTML5 in droves, although it is very much a work-in-progress.
HTML5 will be the latest standard-by-consensus for Web site development, once it is more fully developed and more widely accepted by major stakeholders. That will take a few years, due to factors typical to standards and consensus scenarios. Currently, the standard is incomplete for critical mobile functionality such as video and geo-location.
HTML5 aims to provide a rich interactive user experience on mobile devices within the Web browser. That means a mobile application only needs to be developed once, instead of many times for different mobile operating systems and devices. This also happens to be one of the main reasons why solution providers urge companies to adopt MEAPs.
If by some powerful magic HTML5 was complete and consistently adopted as a standard tomorrow, what would be the business case for Antenna's solutions, rated as so forward-looking by Gartner despite the company's industrial history, rather than others on the MEAP market?
"That goes to the core of strategy versus app-building," answers Parmelee.
"Up to now, so much of the mobile market has focused just on, 'Which toolset do I select to build my app, and how do I secure my enterprise data?' Some take the additional step and ask, 'How do I deploy the app?', and they can use some available point solutions to do that.
"The AMPChroma solution is about full-on support of a mobile strategy. It's not just the tools that build the apps, it's a lifecycle infrastructure that supports publishing of applications- and metrics-gathering such as in-app analytics.
“On the market side, the solution supports a branded storefront the customer company owns. So you're not one of a million apps in an app store. Device management is included as well."
Antenna provides development support for fully native apps as well as HTML5 and hybrid app development, as any number of MEAP vendors claim to do.
Says Parmelee: "Our position is that HTML5 is the future."
Beyond SOAP and Web services
When developing strategy for mobile app development, an enterprise should take an architectural approach, and look at interfaces first, says Gartner analyst David Mitchell Smith.
"We advise organisations to first focus on the APIs," he says.
In general, software applications use APIs (application programming interfaces) to communicate with each other.
"From that, you can develop a multi-channel strategy that can deal with any kind of end-point," continues Mitchell Smith.
"But these are relatively new, and most enterprise applications have not been written to take advantage of these," says Mitchell Smith.
"Many enterprise applications contain hardcoded interfaces. If those are standards-based, they're more likely to be based on older standards, such as SOAP (simple object access protocol) and Web services.
"SOAP and Web services are okay, but these are not the way that people developing mobile apps are going to expect to interface, and not the way modern apps on the public Web expect to interface. Traditional enterprise developers will be fine with these, although these earlier standards will cause delays and issues when getting into the new world of mobile and Web."