Apple Inc. (NASDAQ:AAPL)’s new Siri and Spotlight search services, which deliver results including deep in-app searches, appear to be the biggest unveiling at WWDC.
In-app search is highly significant
The first IOS app store, released in 2008, created an atmosphere like the early days of the Internet: you could only search by catalog, with the only search ability being by simple description or a rare recommendation.
Soon after, Apple added collected lists of popular titles and hierarchical groupings by subject, making the App Store experience similar to the early web era of Yahoo listings, where you could search for websites, but only expect to get simple results based on metadata labels or name matches.
The new iOS 9 in-app search is novel: a Google-like index where hidden data is found by a meritocracy of relevance and popularity.
How it works
In-app search is based on a combination of two mechanisms. One is CoreSpotlight, which offers a way for developers to index content within their app. After a developer indexes their in-app content as public, decided by enough users performing local searches matching the indexed content, a given result is popular enough for Apple to publicly recommend to users searching via Spotlight, or to automatically offer as Spotlight Suggestions, even for users who haven’t installed the app. By sharing popular search results publicly, a new level of exposure for valuable content within developers’ apps is provided, an idea greeted with tremendous applause by WWDC attendees.
In addition to public content indexed by the app developer, Apple will also incorporate hidden content from apps defined by NSUserActivity. It is the exact same mechanism Apple introduced last year as the foundation of Continuity’s Handoff feature. It identifies an activity in progress on one device that can be replicated on another, such as transferring a web page the user has on their iPad to their desktop Mac.
For in-app search, the same “App History” process captures a user history which other users may find useful as well. For instance, if you found a nice pair of shoes in an app, NSUserActivity could return you to the same pair of shoes in a Spotlight or Siri search. When enough users have performed the same search, finding the same pair, Apple’s global index would understand that app’s pair of shoes as a valuable result for others.
In addition to a pair of shoes, other results might include a concert ticket in Eventbright, or treatment for a sprained ankle in WebMD. In-app search means that rather than searching for an entire app, users can search for just the content they are looking for, and get specific results; tied to an app they can download. Further, the user isn’t just left to search within the new app, but is taken right to the content once the app downloads.
In-app search allows for much more intelligent Spotlight Suggestions that can predict what a user may be interested in by context, location and patterns of behavior. Further, it provides developer tools to make search results rich and actionable, supporting result listings that include images, audio, video, ratings, and prices. Examples include dialing an phone number (for an Airbnb listing), presenting a map (of a business contact), or playing media content (for a movie trailer within an app, or a song within iTunes).
Building off of these two index populating mechanisms is a third: Web Markup. It allows web site administrators to reference their iOS app from the web, allowing Applebots to pick up deep links to apps and also add these to its index. Users can search for content in Safari and see results that can open natively within an app.
Another benefit of in-app search is that Apple can more easily suppress “Search Engine Optimization” and the harmful content that comes with it. This is because iOS 9 is indexing in-app content and relevancy is tied to user engagement.
“Google Now.”, part of Android 4.1 and the predictive search features introduced by Google Inc (NASDAQ:GOOG) in 2012, has a striking resemblance to the new search-related features Apple Inc debuted for iOS 9 and OS X El Capitan.
Since then, Google has introduced “app indexing,” a feature designed to make search more relevant to mobile users by delivering results that can open within local apps. For example, a recipe might open within a cookbook app, rather than just presenting the same information on a web page or dumping users into the app to find the recipe on their own
Apple does not monetize its search with ads, and therefore has no need to capture and store users’ data and behaviors for future profiling, like Google. Craig Federighi , Apple’s head of software platforms said that the new predictive search services would be anonymous, not related to an Apple ID or to other Apple services, use a randomized identifier, and not share data with third parties.
The dangers of in-app search
An article by Kif Leswing for International Business Times worried that there was a “distinct possibility” that Apple’s new search services unveiled at WWDC “may turn out to be inferior to Google Now because of the simple reason that Apple doesn’t collect as much data as Google.” The article cited IHS technology analyst Ian Fogg as saying, “Essentially what Apple is aspiring to do is deliver an experience as smart as the opposition while having one hand tied behind their back.” Nonetheless, iOS devices have plenty of access to private data that they can use to shape search results, without ever linking that data to an individual.
Two roads to profit
Google’s income is directly related to its appetite for identifiable user information linking across all platforms the same way Apple’s desire to make faster, thinner, more elegant hardware is directly tied to it income. Apple’s efforts to erase the persistent links between advertising identifiers, the identity of end users and the unique hardware identifier of their devices fits into their business model.
Google outlines that it collects and links lots of personal data. And the entire intent of Google+, an initiative the company has aggressively forced upon its users, is to build a persistent profile of as much personally identifiable information as possible. It’s not for relevant search results; it’s aimed at optimizing extremely targeted advertising. This fits into their business model.
Apple’s search services are showing is that robust, rich, relevant search results don’t need massive data collection programs, unless they sit upon a business model of advertising. Apple is also showing that even effective advertising doesn’t require massive data collection of private data through their iAd service.
Further, Apple is showing that it is not dependent upon Google for its search, maps, cloud storage or advertising services. That’s a big problem for Google, because its 75 percent of its mobile advertising revenue directly relies on Apple.
In the end, much depends on how many users can be convinced that their privacy doesn’t matter. Especially when the overt lack of privacy on Android increasingly offers few unique advantages over the enhanced privacy Apple secures for its customers on iOS.