We needed to actually validate two different propositions, with two different apps – one for drivers and one for passengers
Other platform considerations When you launch your app, you need to test it out on the smartphones that will run it. This means you need only two devices to test your app on. Apple has kept it simple, because it builds both the phone itself and the operating-system software that runs on the phone original site. If you want to be thorough you can test your app on older hardware such as the iPhone 4 or 4s as well. Android, on the other hand, is more complicated to get right. This is because it is open-source and has therefore been adopted by hundreds of companies to run their smartphones – so you’re looking at a slightly bigger selection of handsets to test on. In 2012 there were about 4,000 unique devices running Android; in 2013 it was around 12,000. About 600 different companies manufactured those devices.6 Hmm. Not sure you can fit that many handsets in the office. What does that mean? It means if you focus on Android first, you’ll need to support a lot more than two screen sizes from the start. It also means that everyone needs to think about screen size all the time – from your product managers, to your designers, to your engineers. It also means you need people with rich experience of working with Android – you don’t want to have an inexperienced designer cut their teeth on your Android design. It also means you need to have at least 10 of the top Android handsets in your office so that you can test that your app works on them as expected. And, yes, there are big differences between how your app can or might behave on an HTC ssung one. If you don’t test thoroughly, you risk annoying users unnecessarily. We saw a wide range of issues with the Hailo app on Android. Supporting Android added significant overhead. Another consideration is which versions of the operating systems your app supports. When we talk about these two operating systems – iOS and Android – it’s worth noting that they both have a substantial update about once a year. It can be the equivalent of upgrading your desktop from Windows 95 to Windows NT to Windows 8. There is a good chance that your app will require development work to make it function perfectly – or at least a lot of testing.
And it may not just be something superficial that goes wrong: it could be something a lot more operational
Recent changes have made updating the OS a bit easier – in late 2011 Apple revamped its operating system to update itself over the air. This is what happened with the latest upgrade from iOS Version 6 to Version 7. Within a few weeks of its release, about 75 per cent of all Apple devices (that’s about 230 million) were running the new version. 7 In early 2014, there were 600 companies running a version of Android on their smartphones. It means there are eight different versions of Android running in the wild.8 In reality, it’s not quite that bad: the latest three versions of Android account for the vast majority of devices. But it does mean you have to make some hard decisions about which of your users to support – and which ones you need to let down.
You’ve made your decision to build on iOS or Android. Now you need to get your head around what you’re actually going to build. Everything starts with Version 0.1 – it’s the very first iteration, the prototype, of your app (Version 1.0 is reserved as the first version shipped to the public). In this version you want to focus on the most basic set of features that will make your app unique, useful and different. It’s often called the MVP (or the minimum viable product). At this point you want to focus on only the parts that are absolutely necessary to show why your app delivers something new and novel – something that wows your users. For Hailo it was focusing on how a user could see nearby taxis on a map, then hit the ‘Pick Me Up Here’ button and have a driver accept the hail. We also added in the ability to see the driver come towards you. That was enough to make people feel a wow moment. Since we had limited time and resources we ignored everything else – there wasn’t a proper way to register an account, you couldn’t add your name, you couldn’t interact with the map, and you could see only the taxi’s number plate. It was bare bones. But it delivered a feature that no one else could offer! But it wasn’t quite that simple with Hailo. All on a tiny budget.