The Fatal Flaw of Deeplinks And How To Fix It
I’m a big fan of deeplinking – it takes the best of the web, the ability to seamlessly jump from site to site, and brings it to mobile apps. Many such as Facebook, Google, and URX are trying to move deeplinking forward by making it easy for developers to implement them. Here’s an example of how it works: I’m in the Google Maps app and I tap get an uber, which launches uber. Another example might be opening a link to a tweet on the mobile web and instead of it opening the mobile web version of the tweet, it opens the Twitter app. As you can see deeplinks need to happen or apps will continue to sit in silos. In short deeplinks solve a very simple problem: letting apps talk to each other and open one another just like webpages
There is one fatal flaw: The user must have the application they are being deeplinked to installed on the device already.
If they don’t, they’re taken to the mobile web version or simply asked to download the app. This is a terrible user experience. Imagine it if the web worked like this for the past 20 years: You could only access a youtube video or this very article you’re reading, only if you had the app installed for my domain. If you didn’t, you’d have to wait a good bit for it to install and then redirect you finally. You would eventually repeat this process hundreds, if not thousands of times, until your system was overloaded with applications you didn’t really need. Luckily, that’s not what happened and the web allowed us to instantly access what we were being linked to without the need to “install the site”.
This works great for the head, applications such as Facebook, Twitter, Pinterest, Spotify, etc., but it doesn’t work for the torso and long-tail. As anyone will tell you, that’s where the magic happens and why the web was able to grow. The web combined with search/social allowed us to instantly access hundreds of millions of smaller places instantly. We weren’t relegated to having the site installed. Until we fix this problem, we’re going to be in a pre-netscape and pre-pagerank era. There needs to be a way to instantly access apps when we are driven to them via links. So what’s the answer?
The answers is simple: cards. Cards will allow us to instantly start utilizing an application or a piece of content we are linked to on mobile without having to install the app. If we have the app installed, the corresponding deeplink will bypass the card and open the application. Most won’t fit into that use case. Most won’t have the app installed and they should be able to get value from the app without having to go through the friction filled process of installing it. Cards allow us to link applications together and make their utility instantly accessible. For example, when I was linked to Uber for the first time a few years ago, I shouldn’t have had to download the entire app. I should have been linked to a card where I could order my first ride to get started. From there, I could install it if I wanted to use it further. If I’m linked to an commerce product on an obscure e-commerce site, I don’t want the mobile-web, I want the app experience without downloading it and I want it instantly. Cards give me the frictionless access provided by the web and the experience of a native app.
Cards are useful for first time users of software applications that should eventually end up on the home screen. Cards are for the: I don’t use this yet and don’t want to install it quite yet, but I want to give it a go. Cards are also useful for the content that we never will install. We will never install the millions of websites and e-commerce sites out there. <10% 2 10 of their traffic is through the homepage, but driven by links on google and facebook. we access content via referral traffic. instead accessing mobile website, i should a card for that article or product. cards solve pre-netscape problem “i have to install an app in order use it" not being able instantly software like could web. shouldn’t just be linkable, they embeddable. why did youtube soundcloud grow so fast? were embedded across can’t embed uber, great from flipboard, dark sky forecast, etc. thing is: totally to. think as application video card. if embeds it would never taken off. embedding lets us page rank ability apps. seeing who linking applications, can start measure where apps are them. more also installed, higher rank. this going crucial bring next billion smartphones developing countries there sensitivity data usage. existing web provides webpages often too big size. at even 500kb per webpage, means pages alone consume 1% person’s data. other hand, using all nothing proposition. requires downloading entire app, eating up lot data, all. most go latter. With cards, we can let people try and use an app for a fraction of the bandwidth. 10%>
Lastly, what’s interesting is that on Android Lollipop, the operating system treats each tab in chrome just like an app in the app switcher. If cards start to become prevalent, they will be treated by the Android OS like apps in the hierarchy as well.
So here’s the recap, every app and every site should have an associated card for when they’re linked to from other sites and apps. If the user has the app installed already, they should be deeplinked into the app, bypassing the card. If not, which is the majority of the time, show the card. The card should let them have a simple interaction – book a table, book a car, buy an item, etc. After that first interaction, they can then install the app and add it to their home screen. Cards are how we bring mobile into the post page rank and post netscape era. We can now let apps link to each other and be instantly accessible.