Grokr & Expect Labs: More Mobile Search Insurgents Appear

Just when you thought competition in search was effectively over, it starts to get interesting again. Over the past two weeks I’ve spoken to no less than four companies that say they are trying to reinvent “mobile search” for smartphones or tablets. I would almost guarantee there are more out there as well. These companies […]

Chat with SearchBot

Screen Shot 2012 12 12 At 6.37.51 AMJust when you thought competition in search was effectively over, it starts to get interesting again. Over the past two weeks I’ve spoken to no less than four companies that say they are trying to reinvent “mobile search” for smartphones or tablets. I would almost guarantee there are more out there as well.

These companies either believe Google is vulnerable in mobile or say they’re trying to bring a fresh approach to content discovery that goes beyond what they believe has been done to date on mobile devices.

A More Visual Approach

Last week I wrote about KickVox, a new mobile search app for iOS and Android that tries to take a much more visual approach to initiating queries and presenting content. Even though others before it (e.g., Taptu and Do@) have tried and largely failed with similar approaches (in the face of the “Google habit”) KickVox believes it has created a richer and more satisfying mobile search experience.

There’s another company I recently spoke to that is doing something similar for tablet-based search. However it has yet to launch its product.

There is also another version of “visual search” that uses camera-centric apps to obtain information by taking pictures or scanning products or barcodes. Indeed, augmented reality can be a form of search. Google Glass falls into this category too.

The Search Assistant

Aside from the more visual approaches to mobile search there are, broadly speaking, two metaphors taking hold as the next generation of search (on mobile devices) starts to unfold. One is the personal assistant, exemplified by Apple’s Siri but also by startup-apps such as Speaktoit and others. Google has created its own version of this with its recent voice search upgrade that came with Jelly Bean for Android.

In these cases structured data sits behind a voice-enabled personality or avatar that employs natural language understanding (to varying degrees in each case) to deliver requested information: “Where’s the nearest gas station?” or  “When is Skyfall playing this weekend?” In certain fundamental respects this approach is quite a bit like query-in-a-box traditional search with a speech-to-text overlay. There’s a user query and an “answer.”

Predictive Search

Another approach, represented by new mobile search app Grokr, employs a kind of passive or persistent search. It gathers information from various sources — your location, time of day, social networks, your implied or stated interests — to provide content that dynamically changes throughout the day. It will suggest you search for breakfast places until noon, and so on.

Screen Shot 2012 12 12 At 9.05.43 AM

This approach is generally being called “predictive search” because it tries to predict and suggest content to users and doesn’t require the user to input a query for established categories of information. Follow certain sports teams? Grokr will show you their scores without requiring you to ask for them. However Grokr also offers a conventional search capability and results, as the image above left indicates.

Grokr has done a very nice job with its UI. It also offers a great deal more value and information than simply “mobile search.” As an aside, if I were Yahoo I would buy the company and make it the guts of a new, richer Yahoo mobile search app.

Is Mobile Search Really “Broken”?

While we “in the industry” talk about the limitations of “ten blue links” and other perceived Google mobile shortcomings it’s not at all clear that the public believes that mobile search is “broken” or that it needs alternatives to Google on mobile devices. One potential counterpoint to that assertion is the fact that smartphone owners often bypass Google or mobile search and go directly to trusted apps to get the information they need.

On the PC most people start at Google, even to navigate to known sites. This behavior has been bitterly described by some publishers and brands as a “Google toll” or “Google tax” because users often click on paid ads for the intended brand or site. Publishers are compeled to pay for customers and users they already “own” in many cases. That navigational query behavior is almost typically absent in a smartphone context where users have dedicated apps installed.

Google’s adaptations to mobile devices have included the creation of a Siri-like voice search experience, as mentioned, more “answers” in the form of knowledge graph content and information “cards” (e.g., weather, sports, stocks). Google Now is the company’s own version of “predictive” or persistent search, which presents these info-cards in response to temporal or environmental cues. In fact Grokr is being described as “Google Now for iOS.”

As more “big data” and “ambient” information are collected by smartphones and utilized by companies and apps, predictive search will get better. Google Now can be very impressive but it can also be wrong. Expect much more development in this area.

MindMeld Offers Real-Time Predictive Search

Another interesting variation on predictive search is Expect Labs‘ MindMeld iPad video chat app. In partnership with Nuance, which also front-ends Siri, MindMeld uses natural language understanding to offer search results and content in real time in response to the content of conversations. On a more sophisticated level it’s like what Google is doing by matching ads in Gmail to the content of emails.

As people chat on the Face Time-like app the machine mines the content of the conversation and offers information and suggestions. For example, a conversation about where to go to dinner in San Francisco would yield data about San Francisco but also produce a list of restaurants. None of this involves the formal input of a query into a search box.

The Expect Labs technology is very interesting and has lots of enterprise applications (think customer care). MindMeld is merely one expression of the underlying technology. Below is a video of MindMeld and how its “search” function works.

Where to from Here?

As you can start to see from these examples (mobile) search is evolving. That evolution will continue and perhaps accelerate. Some of these innovations will make their way back to the desktop — as we saw with the recent changes to Google’s OneBox to mirror mobile search results.

Mobile devices (smartphones in particular) have limitations and capabilities (e.g., precise location awareness) that both enable and require companies to push the conventional boundaries of PC search. In many cases these innovations also raise privacy questions. There will be a lot more to say about privacy in 2013.

Whether these mobile search startups can gain mainstream usage among consumers is another question. However it’s great that they exist and have the chutzpah to take on Google.

Putting aside “social search,” which until further notice has stalled on the PC, mobile is pretty clearly where the action is in search. It’s where the majority of the innovation — and potentially competition — will come from in the immediate future.


Opinions expressed in this article are those of the guest author and not necessarily Search Engine Land. Staff authors are listed here.


About the author

Greg Sterling
Contributor
Greg Sterling is a Contributing Editor to Search Engine Land, a member of the programming team for SMX events and the VP, Market Insights at Uberall.

Get the must-read newsletter for search marketers.