Web First, Hybrid Second, Native Third
If you want to create a mobile app, one of the big questions you need to answer early on is: Web, hybrid, or native?
There’s no one-size-fits-all answer; there’s a lot to consider. But how do you go about making the decision? Which option should you consider first? I propose Web First. Only if the Web alone won’t do, consider Hybrid Second. Finally, think about Native Third.
The best reasons to go with the Web can be summed up as: portability, shareability and updatability (I may have invented one or more of those words, but you get the idea!).
Not all applications need to work on multiple platforms. If you’re setting out to write an iPhone game, it’s a valid choice to just target iPhones. But for most application developers, we need to consider multiple platforms. How do we cater for people with Android, Blackberry or Windows Phones? This is a problem that’s growing fast. In 2009, Android’s market share was 4%. Now it’s over 50%. And Nokia are expected to sell 37 million Windows Phones in 2012.
Avoid locking yourself into one vendor by using open, standardised technologies that can work for many. Avoid writing two or three separate applications with different codebases and having to maintain them all separately. Write one app to rule them all!
The counter-argument to this is that it’s not easy to get Web apps working perfectly across different types devices. For anything beyond the most simple application, you’re going to find things you need to tweak for each device you test. You’re likely to need various tools and techniques such as modernizr and media queries. You may even need to factor in devices that don’t support these very tools that should be easing the process. It’s likely to be painful for the foreseeable future. But in many cases, it should be worth it.
As well as different vendors, there’s also the question of different device sizes. We already have mobiles, tablets, ultrabooks, laptops, desktops, Internet televisions… and there’s no telling where the variety will end. The Web is the best way to reach all these different types. Once again, it won’t be easy. You’ll need different sets of styles and lots of tweaking. You’ll probably end up reading a lot about Responsive Design (and perhaps like me you’ll think we’re not quite there yet when you load the much-hailed Boston Globe on your mobile phone and find one very, very long column). But… it’s possible to cater for this variety and you can’t say that for anything else but the Web.
URLs are underrated; they’re the Web’s killer feature. They make it easy to share your application, or even a specific page or part of it.
Native apps are really missing out. I saw an advert on the tube today that advised those looking for their app to go to the App Store and search for “Parker Car Service Smarter Minicab Booker”. This is not a good way to point people to your app. Okay they could’ve used a URL (iOS devices load itunes.apple.com URLs in the App Store), but the reason they didn’t is that it would be pretty confusing for customers.
Say we bothered to tap in that big long search query and we’ve now found the App Store listing for our app. Now we have to download it. This isn’t a big hassle if we’re going to use the app a lot. But nearly 30% of apps downloaded are used just once. Compared to just clicking a link, that’s a lot of effort to go through if you’re going to use it once and throw it away.
As well as URLs, it’s the ubiquity of the Web that makes it so shareable. All smartphones have a Web browser. Not all smartphones have a particular native app installed. This is particularly important for sharing on social networks. I can easily point my friends to a Web link and they can load it up and consume the content within their Twitter or Facebook applications. Twitter and Facebook can embed a Web viewer within their apps because the Web is ubiquitous and non-proprietary.
Releasing updates to native apps is a pain. You have to go back through the app store release process. For iOS, that involves re-submitting to Apple and waiting a couple of weeks for them to approve it (or they might reject it).
With the Web, you can just push out the new version at your convenience. You can be more responsive to feedback, fix bugs quicker and generally keep your content up to date much better.
Native apps are a pain to update from the consumer’s perspective too. On my iPhone, the App Store always has a big red number next to it, glaring at me for not updating my apps more often. Scott Hanselman called it “feeding the update beast”. For big apps, you may need to wait til you get on wifi. Native apps do allow you to potentially stick with a specific version, whereas with a Web app you’re forced to update. But effectively you’re forced to update native apps too, because if you don’t, you’ll just have a notification glaring at you for eternity.
The Web alone doesn’t work for you? Okay let’s move onto Option 2: “Hybrid”. This is a kind of mish-mash of Web and native; basically, wrapping a Web app within a native app. Tools like PhoneGap are very popular and make this pretty easy. Some reasons for doing this are: extra features, payments and discoverability.
1) Extra features
The biggest reason to put your web app within a native wrapper is to add native features that you can’t implement with the Web. For example, integrating with the camera or the contacts book.
I won’t try and list all the things you can and can’t do yet with the Web, but it’s worth saying for the benefit of us future-gazers that the Web should catch up with a lot of these features. The Device APIs Working Group is working towards this, but unfortunately it’s been rather slow-moving. Mozilla are hoping to fast-track some of it through their WebAPI project, which we should see something from quite soon.
Of course, the Web will never be able to do everything that all devices can do natively. Proprietary features can be made available quicker. Shared standards evolve slower. So there will always be reasons to develop particular features with native code. I predict though that more native coding will become simply add-ons to Web codebases; less apps will be written wholly in native code.
One advantage that native/hybrid apps have, at least on Apple devices, is that people are quite happy to pay for them. Apple have made it as easy as it can be, with a one-click-plus-password method. It’s just the same process whether you’re downloading a free app or one that costs money, so there’s no extra steps to put you off.
I don’t think we’ve really seen this level of ease come to the Web yet. However, it’s not a complete win for the hybrid/native approach. It’s worth remembering that you’ll pay Apple a lot for the privilege. If you are able to roll your own subscription or payment method, you could save a lot by not having to pay Apple a 30% cut.
App stores do provide a great way to discover apps, but I think this argument can be a bit over-stated. Let’s not forget the Web’s powerful feature, the URL. An address like app.ft.com is easy to remember and share. The FT replaced their native app with a Web app, outside of the App Store. By breaking the million user mark, they’ve proved that this model can be successful.
You’re still here? Can’t do what you want with Web or Hybrid? Let’s explore the third option then: writing the app purely in native code. I’ve bundled some reasons to go down this route into: Performance and Other Considerations.
2) Other Considerations
Other reasons for choosing pure native? If you need a lot of native features or if they’re the core parts of the app, then maybe it would be messy, or simply unnecessary, to write some bits using Web technologies.
Offline capability is perhaps the most misleadingly quoted reason to choose native. People instinctively think of the Web as being connected all the time, but it is possible to store Web apps (both the data and the apps themselves) offline, using HTML5. However, it is true that native can provide more storage and greater capabilities in this area. Also, being early days, it can be a bit tricky to work with these HTML5 features at the moment.
I believe that the Web should be the default choice for applications. It’s the most portable, flexible and accessible option. It’s not easy, but it will get easier. If you can’t achieve what you want with a straight web app, the next choice is hybrid. Finally, there’s the option to go purely native.
To convey this visually, I’ve created a flow chart. It’s a very simplified view and shouldn’t be taken too seriously, but I hope you like it:
Have I missed any key points? Been unfair about anything? If so, please get in touch!