Deep linking is hot, and for good reason: it can make mobile search much more powerful by making it more predictive. In a world governed by small screen sizes and large intent, that's a huge asset.
And it's why deep linking startup URX recently introduced AppViews, a product with the not-so-modest goal of "understand[ing] what a user is doing and what could make them happier."
To better understand not only AppViews but the innovative deep linking marketplace, and whether the increased intelligence buried in deep links will lead to privacy concerns, I sat down with URX marketing chief (and former Googler) Mike Fyall (@mikefyall).
Going Deep On Deep Links
ReadWrite: Why is deep linking important? And is there more to it than people generally think?
Fyall: Deep links help users travel directly inside of an app to the right place, similar to a URL on the web. They help users save time, marketers create better campaigns, and developers build cross app experiences.
Deep links will also usher in much needed new tools for app discovery and engagement. The deep link itself is just an address; however, as companies like Google, Facebook, URX, and others build an understanding of the content behind deep links we'll be able to recommend apps to users when they are most relevant and useful.
This is happening today. For example, Google is showing app content in search results, and companies like URX are building products (AppViews) that recommend relevant apps based on a users context. We'll see lots of innovation here in the coming years, particularly as daily transactions inside of mobile apps continues to increase.
Everything Is Connected
ReadWrite: What sort of data does URX glean from deep linking?
Fyall: When URX crawls webpages with deep links on them, we capture information about the page content—for example the headline, keywords, images, and other metadata—using similar techniques to how search engines index webpages today. This helps us understand where the deep link goes so we can recommend it at the right time.
URX can also check to see if a user has a given app installed before deciding which AppViews is most relevant.
ReadWrite: When announcing AppViews, you talked about building an understanding of the entities inside mobile apps and how they relate to each other and the physical world. What does that mean?
Fyall: In order to show relevant AppViews for a given context, we need to understand how people, places, things, and concepts relate across apps.
For example, the "Beyonce" that you can listen to in Spotify is the same "Beyonce" you can also listen to in SoundCloud, or buy concert tickets for in Stubhub. Without this understanding, "Beyonce" is simply a 7 letter word to a computer and you can't figure out what a user might want to do next.
We've have a team of data scientists working on our knowledge graph (short explanatory video) for over a year. We utilized the Freebase project to kick it off and continually enhance it as more data flows through our system.
Today, if a user is browsing a story about Beyonce we have a host of possible destinations a user might be interested in visiting.
ReadWrite: You talk about suggesting the "next action" to make users "happier," but how can you discern what they really want from deep links?
Fyall: AppViews are about giving users a relevant recommendation for what they want to do next. To them, we aren't offering a deep link: we are helping them discover content or take action. We named it "AppViews" as the goal is to give the user a glimpse inside other app before deciding to leave their current experience.
The higher the user engagement is with AppViews, the better job we've done for users.
ReadWrite: Who pays for this? Meaning, a developer might want to surface content to me in her app, but presumably you're not going to show it to me unless it will make me happy, right? How does this work?
Fyall: To date, we've seen the majority of developers implement AppViews to add functionality to their site and getting paid is secondary.
Nexercise, the developer of the Sworkit fitness app, told us they've seen a host of positive reviews and feedback when they added the ability for users to listen to music before they workout.
In terms of the business model, advertisers will be able to pay for "Promoted AppViews" which will appear alongside related content. For a promoted AppView, the developer receives a majority of the revenue and URX take a cut as well. Developers have complete control over the format and types of recommendations shown.
All The Pieces Matter
ReadWrite: With AppViews you're getting into big data-type applications, not merely mechanical "Click here to go to the right place in an app." Where does this lead? (And will it be a privacy problem?)
Fyall: Deep links are the required infrastructure to build the new mobile discovery platforms that are so desperately needed on mobile. For as powerful and amazing as apps have become, we're still stuck browsing through app stores and can't move intelligently between apps. URX is one of several companies trying to reinvent how we discover and engage with content on our phones.
Most mobile ads today, for example those shown on Facebook or exchanges like MoPub, are targeted toward users based on demographics or previous behavior where privacy is a strong concern. However, AppViews are shown purely based what someone is doing right now— it doesn't matter who you are, just want you might want to do.
So privacy shouldn't be a problem for AppViews now or in the future.
Lead image courtesy of PicBasement
Walang komento:
Mag-post ng isang Komento