I will take "the Web" to mean, essentially, that the rendering of the front-end is done in a browser. Then, what you have to do (I learned this the hard way) is stick to things that get rendered well in browsers on all the platforms you plan to target (which usually includes Android, and that means no fancy scrolling). Develop your aesthetic around that. You can put out native versions later, when you have the budget to support all the platforms.
Now, this has several advantages. One is if you want your app to become popular. If your app involves inviting friends in any way, realize the friends don't have the app, and the invitation will probably happen via SMS/email/facebook/twitter. When that link is clicked it takes the friend TO THE WEB. Yes you can have the friend look at an app in the store, download it, launch it, confirm their phone number again, and then finally engage with whatever they were invited to. Or you could have a WEB EXPERIENCE ready for them and create an account with 0 steps!
And since you've already coded this web experience, why not wrap it in Cordova/PhoneGap and add bridges to the underlying OS APIs via plugins. This keeps separation of concerns and actually maintains much better than many native apps. And you'll also instantly have access to many cool components like Mobiscroll, which have been tested and work everywhere. Compare that to downloading some native lib for iPhone, like Layer, only to realize it doesn't work on other platforms like the web, and you have to wait or roll your own.
So now you've got a unified codebase which renders in a browser, you can wrap it in MacGap and other wrappers, hire specialized developers on each platform and pay them to maintain ONLY the part they need to. Not the interface, not the server interaction, backend etc. Which all has a standard and RICH environment to build in, including localStorage, local database, sockets, even WebRTC for peer to peer time topologies! All that stuff is good enough in 99% of the cases to build an app.
Well I'm a little biased. My company has been building a plaform for years to make apps successful and viral, a sort of open source alternative to Facebook (which just shut off app access to people's list of friends, something that was the biggest selling point at the very beginnng and the reason you shouldn't build your business on someone else's centralized platform). The idea was to have a standard platform finally to power SOCIAL apps, that was all well designed and take care of everything for you, so you can focus on write your app, not getting cordova, user signup invites and privacy / security right etc. It also supports a growing ecosystem of reusable widgets and plugins that you can just plop onto a page, like chatrooms or docs. I thought it would be done in a year but it's been four years and we're still at 0.8!