November 25, 2012

The future of search

I believe there are two ways search will improve significantly in the future. First, since talking is easier than typing, speech recognition will allow longer and more accurate input strings. Second, search will be informed by much more persistent user information, with search companies having very detailed understanding of searchers. Based on that, I expect:

My reasoning starts from several observations:

In principle, there are two main ways to make search better:

The latter, I think, is where significant future improvement will be found.

So how does a search engine understand what you want? It can listen to you directly, parsing your search string. It can ask for more clarity, through some kind of disambiguation interface. Or it can make inferences, based on — well, based on just about any kind of information that might exist about you and your online behavior.

Search strings are short, typically four words or less. That doesn’t leave room for a lot of innovative parsing. Not a lot of progress can be made until search strings get a lot longer, and that is unlikely except perhaps through the convenience of speech recognition.

Faceted/parameterized selection has its place. For example, when I search on Amazon.com, the site encourages me to also select a department from its dropdown menu; otherwise, it refuses to rank the search results. And when I buy shirts from Land’s End, I just click through and never search at all. Still, Google’s been around for 15 years, and about all its successes in searcher-does-the-work disambiguation boil down to is:

Whatever the user attitudes and behaviors are that constrain Google’s or its competitors’ success in this area, I can’t imagine them changing much — except, once again, in the event that speech recognition leads to richer human-computer conversations.

I’ve now highlighted two different ways in which there’s a search-interface challenge that will be tough to beat without turning to speech recognition. But the case for speech recognition is even stronger than that. We’re moving to small, mobile devices, and:

And so I conclude that speech recognition is a big part of the future of search.

What will that allow? Since talking is easier than typing, speech is a way to get longer text strings as search inputs, or more of them. It’s plausible that people might speak queries as complex as:

Getting search engines to the point that they can handle such queries will be difficult but straightforward — but even more progress is needed. Search results for various queries will be greatly improved if the search engine “knows” things like:

And that will cement internet search squarely in the world of — for once I approve of the term — big data.

Comments

4 Responses to “The future of search”

  1. Diane Berry on November 26th, 2012 6:22 pm

    Hi Curt,

    Thanks for the interesting thoughts on enterprise search. I have been remiss in not keeping you informed about Coveo. Here are a couple of recent articles, in Fast Company (http://bit.ly/FastCompanyCoveo) and Forbes (http://bit.ly/ForbesCoveo), that can give you an idea of what we’ve been up to. It’s far beyond basic enterprise search–in fact it is close to what you are writing about–presenting information based on knowledge of the user’s needs/context. As an example, we’ve recently launched Coveo for Salesforce which instantly assembles and presents data, directly within the Salesforce UI, that is contextually relevant to what the user is working on. Here’s a link to a video demo: http://bit.ly/CoveoVideo. Coveo brings intelligence from anywhere, to any context. We’ll reach out to schedule a briefing soon.

  2. Curt Monash on November 26th, 2012 9:49 pm

    Thanks, Diane. A few quick thoughts:

    1. Am I correct in reading that as more of a knowledge-management focus for you, only you perhaps call it something like “360 degree customer view?”

    2. Congrats to your CEO for graduating from college when he was 20. I got my PhD from Harvard when I was 19. 😉

    3. Are those links sponsored pieces, like so much else that is ascribed to “Forbes” these days?

    4. It’s tough to reconcile 80% growth, 100 employees, and ~400 customers with the figures you gave me in 2008 as per the link above. Is that article some years old? Why has your customer count outright declined?

    5. Fixed a URL typo in your comment for you.

  3. B2C internet software | DBMS 2 : DataBase Management System Services on November 27th, 2012 1:04 am

    […] So is search. […]

  4. Diane Berry on November 27th, 2012 9:56 am

    Hi Curt, Thanks for responding! Yes, Coveo’s business model has indeed changed since 2008. We’d be happy to give you a briefing to walk you through our evolution from a downloadable, point search solution to an enterprise-level solution provider. I’m impressed with your PhD at 19!! The articles are not sponsored – pure editorial. I joined Coveo in 2009 – and it’s entirely my bad for not updating you. We will reach out to schedule a meeting.

Leave a Reply




Feed including blog about text analytics, text mining, and text search Subscribe to the Monash Research feed via RSS or email:

Login

Search our blogs and white papers

Monash Research blogs

User consulting

Building a short list? Refining your strategic plan? We can help.

Vendor advisory

We tell vendors what's happening -- and, more important, what they should do about it.

Monash Research highlights

Learn about white papers, webcasts, and blog highlights, by RSS or email.