Why Apple is about to build, buy, or partner on a web search engine

Apple's Siri could change the game in search and put a lot of pressure on Google, but first Apple needs to deal with web search integration.

Apple's iPhone 4S was a disappointment to all of those who were expecting a redesigned iPhone 5, but in the grand scheme of the things the launch of the iPhone 4S may turn out to be Apple'sChamber of Secrets.

Forgive the Harry Potter reference, but Chamber of Secrets is the second book in the seven-book Harry Potter series, and while it's generally the least favorite of the books among Potter fans, by the time you get to the final book you realize that Chamber contained critical plot information that foretold important future events.

The fact that the iPhone 4S was an incremental hardware upgrade and lacked a new design has largely overshadowed its one revolutionary feature that could shape Apple's future: Sirivoice commands and voice-activated search.

Apple has limited Siri to the iPhone 4S to start, but that probably has less to do with Siri needing extra computing power on the phone and more to do with Siri still being in beta. Since Siri requires a cloud connection, limiting Siri's spread at first gives Apple the opportunity to stress-test its data centers and scale up for the future.

Even with its beta quirkiness, Siri is impressive. While Google Android and Windows Phone 7 both had a jump on the iPhone in terms of voice control, Apple has zoomed past both of them with thepurchase of Siri and its integration into the iPhone. The big deal for Siri is that it understands natural language and it is standardized across a lot of different applications on the iPhone. The user doesn't even have to be aware of which apps to use. You can simply give Siri a natural language command and she automatically interacts with the right app to execute it. That's a nice step forward for voice user interface (VUI).

The Siri experience hearkens back to the launch of the original Macintosh in 1984 when Steve Jobs climaxed his unveiling by saying, "I'd like to let Macintosh speak for itself" and it did (using Macintalk software), which blew the minds of techies at the time. Of course, in a larger sense, the whole thing also points back to the computer in Star Trek and its VUI. In other words, Apple has been entranced by the idea of integrating speech into everyday computing for a long time — almost from the beginning of the company.

However, as fun as it is to bark orders at your phone and have it obey your commands in real time, the revolutionary piece of Siri is what it does in Internet search. It's early and Siri is still imperfect, but there are moments when Siri drastically streamlines the search process and gives us a peek at the future.

For example, I recently asked Siri for "the closest Mediterranean restaurant" (right) and got a list showing 11 restaurants, their user ratings, and their distance from my current location. Clicking any of the selections in the list immediately took me to a map.

Another time, I asked Siri, "How many calories are in a kiwi?" She came back with 46 calories along with a full chart of all the nutritional information for a kiwi.

Last week when I was doing research for my article iPhone and Surface: The moment Apple and Microsoft diverged, I got frustrated trying to find historical data on the market cap and revenue of Microsoft and Apple going back to 2007. In desperation (and half-jokingly) I asked Siri a question about Microsoft revenue in 2007 and surprisingly got an answer, based on data from Wolfram Alpha (which was also the source of the kiwi data). That eventually led me to Wolfram Alpha on the web (from my computer) to do a full lookup of the data, but the fact that Siri led me there was a big "ah ha" moment.

Siri can also help you find nearby physicians, lookup movie times, and pull up weather data when you ask questions like, "is it going to rain tomorrow?" Siri still has a hard time understanding normal speech at times and it's limited by its access to freely available data sources like Google, Wolfram Alpha, and Yelp. But, Apple has shown us what's possible with a much more approachable VUI than anything we've seen so far in the consumer market. Siri is almost like an IBM Watson for the masses.

One of the important things to notice about Siri is how it disintermediates search results pages in general and Google specifically. Instead of giving you a page of possibilities to choose from, Siri tries to give you a single authoritative answer to your question. Since Google makes all of its money by allowing advertisers to place their ads next to the items listed on the search results pages, it's easy to see why Google Chairman Eric Schmidt is talking about Siri as a competitive threat.

The next step

Now that Apple has opened the door to a natural language VUI and demonstrated new possibilities, the game really begins. Google and Microsoft will undoubtedly take cues from Siri and bring similar functionality to Android and Windows Phone, since both companies already have a lot of engineers working on voice technology. That means Apple is going to have to rapidly improve and innovate Siri if it wants to be a leader in VUI. Siri has two areas that need the most work: 1.) it needs to keep improving voice recognition, and 2.) it needs more data sources to feed Siri and integrate into its equation.

Currently, if Siri doesn't have an answer to something, the fallback is to throw the question to a standard mobile web search. That's not going to suffice for long – especially when you consider the level of integration that Google and Microsoft will be able to do since they both own search engines. Siri needs a web search that is tightly integrated into the service in the same way that Wolfram Alpha and Yelp are today.

That leaves Apple with three options: build, buy, or partner.

Build

Siri itself is already a bit of a search engine, and with all of the searches that are now happening through Siri and running through Apple's servers, the company is amassing a treasure trove of data about the ways people are using voice search. Plus, all of the Siri data is tied to specific users and that will give Apple an excellent opportunity to do personalized search in the future.

Last year at the D8 conference when Steve Jobs was asked about Apple buying Siri and going into the search business he said, "They're not a search company. They're an AI company. We have no plans to go into the search business. We don't care about it — other people do it well."

While Jobs has famously denied lots of things that Apple eventually went on to do, it's hard to see Apple building its own web search engine from scratch based around the core team it acquired from Siri. That would take years and lot of resources. Just look at how much money Microsoft has had to throw at building Bing, with only moderate success and no hope of turning a profit any time soon.

Buy

The faster on-ramp for Apple would be to buy one of the smaller players in web search, integrate it with the Siri team, and put most of its resources into customizing a VUI that feeds Siri. There are a few decent candidates that Apple could gobble up: Blekko, DuckDuckGo, Yippy, Dogpile, and even good old AskJeeves.

Apple has $80 billion in cash reserves so it has plenty of resources to buy any of these search engines. The best options would likely be DuckDuckGo and Blekko. Both of them already do some things better than Google, but don't get much attention because they're so small.

Partner

If Apple were to partner with another company in search it would have to be Google, Microsoft Bing, or Yahoo (which has mostly abandoned its own search for Bing). Google is an obvious "no" since it's Apple's archrival in mobile. Bing might look like it makes sense in the short term, since Microsoft has fashioned Bing as a "decision engine" rather than a search engine and that fits pretty well with what Siri is trying to accomplish.

But, Microsoft is destined to want to do something similar to Siri in Windows Phone and that will be enough to scare Apple away from a doing a deal with Microsoft.

Sanity check

With Siri, Apple has lowered the friction on search and turned it into a mellifluous experience. But, to take it to the next level, Apple is going to need much tighter integration with web search. Building a search engine would take too much time and there aren't many good options for Apple to partner with in search, so the most likely scenario is that Apple will buy a smaller player and integrate it into Siri.

Siri clearly has tremendous future potential for Apple across its entire product line. By the end of 2013, I expect that we'll see Siri on most iOS devices and Macintosh machines. Nick Bilton evenbelieves Siri is the revolutionary interface that Steve Jobs wanted to bring to television sets.

The bigger and more entertaining question is if Apple does jump into search with both feet, will the company freely release Siri on the Web and challenge Google directly? I doubt it, given Apple's affinity for hardware/software integration, but it's fun to consider, especially as we look at Apple's new VUI as arguably the most important new development in search in the past decade.