I’ve been spending a lot of time thinking about how deep-linking is going to play out. For those who are not familiar, deep-linking is a technology that allows app-to-app communication to function very similarly to the way that web pages work. In the same way that a hyperlink to a website need not drop you off on the front page of that website, deep links hold the promise of allowing app developers and advertisers to deliver users directly to specific sections or portions of a mobile app, provided that the consumer has that app already installed. If deep-lining takes hold, long gone will be the days of clicking on a link to an app and either being delivered to the front page of that application or to a subpar mobile web application experience. That would be a significant improvement over the current state of affairs for consumers, advertisers, and developers.
There have been a number of somewhat recent announcements, including Facebook’s AppLinks announcement at f8 and the large round that URX raised from respected VC firm Accel. There is clearly a lot of attention being paid to this space.
Given how fundamentally powerful the concept of deep linking is, I can imagine a number of ways that deep linking could play out on in the market. My fundamental question about the space is whether a singe company or constellation of companies can control the technology or whether it will become an Internet standard. I can imagine three scenarios for how the market plays out in the short term:
1. Universal web standard service wins out – In terms of ecosystem simplicity, this would be my preferred solution. Some universal, well-understood and widely adopted convention and infrastructure for how web apps talk to each other regardless of the underlying OS would make life much easier for all parties involved. There would be no need to worry about whether the target user was on an iPhone, Android device, or desktop computer. Imagine a world in which you had to rewrite webpages to make them link properly for Google, DuckDuckGo, Bing, and every other web search engine. That would be a nightmare – thankfully we have standard conventions for the web. I would love to see this happen for mobile, particularly for those companies whose strategies are rooted in cross-platform development.
2. OS-specific solutions from Apple and Google – This seems to me to be the worst of all worlds and the most likely thing to happen. I don’t see any compelling reason for Google and Apple to collaborate on a common way to make this work in the short term. I would expect that each would focus on what works best for its respective app ecosystem. The set of companies that should have the most vested interest in a common set of protocols is the Pinterest / Twitter / Facebook crowd – large applications with meaningful audiences that span both platforms and who have larger platform ambitions. Having a common framework would make it much easier for developers utilizing their respective platforms to support deep links with less friction.
3. 3rd party solutions carry the day – There is a handful of companies, including URX, a number of ad networks, Facebook, Twitter, and others who have their own agendas for building deep linking solutions that are not deeply tied to any specific OS and advance their own specific interests. The potential benefits to developers are a) those solutions can be cross-platform and b) they are likely to work today as opposed to standardization deliberations which can drag on and c) they could be a useful abstraction layer as the OS and app store specific solutions morph and change over time.
My sense is that the third party solutions will have an opportunity to establish themselves before the OS-specific deep linking frameworks really take hold. And if you are an iOS or Android specific application, the need to support cross-platform frameworks is greatly diminsihed.
As always, comments are open below or you can send me your thoughts on Twitter @chudson.
Good post. Certainly a lot of fragmentation out there so far. I agree that a universal standard would be the best for the whole ecosystem. But I have doubts that both Apple and Google would get on board with a common one at this point.
I think a 4th potential way this could play out is an open source framework. The goal being to provide an open source industry standard not owned by one particular company.
Not sure if you’ve seen http://mobiledeeplinking.org/ yet but a lot of companies have contributed there. (Disclosure: I work at TapCommerce and we co-authored it). But a lot of leading mobile advertising companies including Criteo, MoPub, HasOffers, Flurry, etc. are pledging support.
You can access the open source spec & technical resources here: http://mobiledeeplinking.org/
Great comment, sir. I agree that open source will play a role in this as well. I think that’s what Facebook wants to do with AppLinks as well.
Hey Charles, found you via twitter and quibb. Nice blog!
I’ve been seeing stuff in quibb about deep linking as if it’s the next QR code and it reminds me of the princess bride. “You keep using this word. I don’t think it means what you think it means”
As an app developer we build deep links into every app we make. That doesn’t necessarily mean there CAN be more of a standard to it. Each app registers a custom URI handler (I.e. Facebook:// instead of http:// or ftp://) and then processes the rest internally.
It sounds like what your implying is a way for an external source to FORM the deep link without access to the target app’s source code, which won’t work. The best bet would be something like a REST API where the app could offer an interface declaring what deep links are possible.
The problem is there is no standard for deep linking even on the web, where everything speaks the same language! To do so is much harder in an app because apps aren’t just collections of addresses the way the web is. The deep link has to instruct the app on how to return to that execution state and display that content– a state that is not a fixed bookmark but more like a train of thought. You can’t get there without recreating the entire thing from the first step.
Think of it more like asking a jazz band to play that part again that went “doo de doo de doot” and you’re getting close to the problem.
That’s why I am confused as to how it got to be a buzzword in the first place. There is no standard to how apps are coded and experienced, so there is no way to standardize “deep links”. Maybe within social apps that are collections of objects (like Facebook) or MAYBE games that have defined levels like Angry birds. Those apps can all accept deep links provided they are coded for it in their own unique way
Noah,
Thanks for the thoughtful comment. I want to be a bit more clear on what I’m talking about. As you mentioned, there are ways for developers to implement deep linking in their own apps today. Some are doing it themselves, some are using 3rd party SDKs / toolkits to deep-link enable their own applications.
What I’ve heard from talking to developers in my portfolio and outside of it is that they do want a more consistent way to talk to other applications that I, as a user, have installed on my phone. So while they can control the way that they implement deep links on their own apps, they have virtually no control over how other developers have chosen to implement it. So it makes it harder to have good, consistent app-to-app experiences. I though the FB demo of what they want AppLinks to enable at F8 to be fairly compelling. And in talking to folks, it seems like there is some fear that the way in which apps talk to each other to enable deep linking could become balkanized.
I think you highlight exactly why apps and the web are different, but I still hear from marketing teams that they want the ability to drop users into specific portions of their own apps. For ad campaigns, they can supply the ad network with a custom link that can deliver that user to a specific page / section / level of the app because, as you said, they had full say in how it was implemented.
So I’m hoping that the ability for apps to deliver users from other apps to specific sections or sub-sections of their own apps becomes easier. That’s what I think of as deep linking.
As for the web, I’d argue that deep linking works relatively well on the web. It’s not perfect, but I think if apps worked as well as the web on that front it would be a vast improvement.