13 Comments

Summary:

Google’s array of search product announcements today were impressive, but not unheard of. Imagine how many startup CEOs are at this very moment drafting blog posts welcoming the monolith to their territory and spinning competition from Mountain View as a good thing.

David vs. Goliath or tide lifting all boats?

Google’s array of search product announcements today were impressive, but not unheard of. Imagine how many startup CEOs at this moment are drafting blog posts welcoming the monolith to their territory and spinning competition from Mountain View as a good thing (just as OpenDNS did last week). On the plus side, if the upstarts don’t get marginalized by Google’s 65 percent search market share, they’re probably some of the ripest acquisition targets around. Here’s our initial crib sheet of which companies are seeing their core products challenged by new Google features:

Real-time: Like just about every real-time search engine, Google (for now) depends on Twitter updates as its main source of data. Impacted: OneRiot, Topsy, Wowd, Collecta. Perhaps the most significantly affected may be Twitter itself, which should probably establish itself as the best search engine for its own data — but hey, cutting deals with all the giants may be an OK short-term strategy.

Location/local: Google said it would be compiling its own reverse look-up data, challenging startups such as SimpleGeo and GeoAPI. It’s also including ratings data directly in its “What’s Nearby Now” mobile feature (at first only available on Android), which will more directly challenge local review offerings from folks like Yelp. Especially in the mobile environment, if users can get all the information they need from one app, they won’t leave it.

Products: Google said it hopes to weave in information about local stores’ inventory with its product search. This is not a new concept; companies like Krillion have been attempting the same thing for some time now.

Voice: Like Vlingo, Google wants to make search more accessible by taking more voice queries. (See also: “How Speech Technologies Will Transform Mobile Use” from GigaOM Pro (sub. req’d))

Images: Using image recognition to match and identify similar pictures is a hard technical problem that plenty of companies have thrown Ph.D.s at for years. Some startups in the space include GazoPa, Like.com and Pixsta.

Image via Flickr user foxtwo.

  1. Liz,

    I like the direction you are taking here. But I can’t help feeling that Google’s real-time offering is seriously flawed. It involves the user choosing to look at “recent” results, so many won’t see it at all. And when it comes on automatically it seems to be entirely devoid of any real filtering or ranking within the stream. Hence… lots of low value results. Not very imaginative execution imho.

    Share
    1. @Keith – But I think most real-time results are not the best response to a search term. Some good situations for real-time results: when it’s the answer to a very specific status question, when news is breaking, or when a topic is trending because lots of people think it’s interesting. There’s not always something super interesting going on in real-time about every topic.

      How would you execute instead?

      Share
  2. If I told you I’d have to kill you :-)

    But…. having said that. There are some obvious things.

    1. Google needs a version of page rank that makes it possible for important recent stuff to beat old stuff in regular search. Page-rank itself is SOMETIMES the problem.
    2. Google shouldn’t show recent stuff unless it is more important than older stuff. At least not in the same interface. Universal search – of which this is just another feature – is already starting to look very confusing. This just confuses things even more.

    The above 2 things would deal with the standard UI.

    THEN.

    It should have an entirely different real-time interface for those who ONLY want real time results. In that interface it needs to go beyond keyword search/matching. For that it would meed much richer metadata and be able to rank based on link or click metrics. Of course Twitter and bit.ly and others would be in a better position than Google to deliver on this given the data sets they have.

    Share
  3. Its a innovative step by google ..all google services must be real time ..
    Nice tip… :)

    Share
  4. Soon, Google will have no friends left in the Valley. They’ll be known as MS was known to be, a predator chasing other peoples innovations.

    Share
  5. In the “Products” you mention Krillion as a company that could be threatened. May be true, we’ll see. But recently-funded startup Milo.com has gotten a lot of press lately as a Krillion competitor. Seems odd that you didn’t mention them as being threatened. I’m sure it has nothing to do with the fact that your boss Om is an investor in Milo, but to avoid even the appearance of that conflict, may be worth pointing out.

    Share
  6. Pixsta has developed a level of sophistication unseen in its Empora.com Search portal when it comes to visual product search. Clearly Google will and tries to catch up. However it demonstrates that this is a very attractive market. The level of visual matching on Images.google.com reminds me of the level of sophistication Pixsta had 3 years ago.

    Disclosure: Straub Ventures, http://www.straubventures.com is a seed and early stage investor into Pixsta and Empora, http://www.empora.com

    Share
  7. @Ryan.

    I agree. I can’t help the feeling there’s a huge backlash coming when web site owners finally realize Google only wants their web site data to build a competing offering on. Duh!

    Share
  8. Liz, in similar-image search, there are two more worth mentioning – Retrievr http://labs.systemone.at/retrievr/ and
    TinEye http://tineye.com/ Thanks.

    Share
  9. this cycle happens every year or so now with Google. They jump into everything, fall off the map with 75% of it, rinse and repeat.

    Share
  10. Full disclosure: I work for Wowd (www.wowd.com), which is mentioned in this article.

    From the article:
    “Like just about every real-time search engine, Google (for now) depends on Twitter updates as its main source of data.”

    That’s not true for Wowd. Wowd uses the Twitter API to get attention data, but Wowd does not offer a Twitter search solution.

    Attention data describes what pages people think are important. Tweets that contain links (appropriately filtered using measures like follower count, re-tweets, etc.) can be used to generate good attention data. Approximately 20% of Tweets contain URLs, so there’s good data to be mined there, if you know how to get it.

    But Wowd has its own, proprietary, and potentially much richer source of attention data that comes from the cloud architecture used. (More at blog.wowd.com.)

    The bottom line is that Wowd uses attention data to rank the web itself, in real time. Pages on the web are what a search at Wowd returns, not Tweets. And these pages are re-ranked in real-time, according to how people are referring to them.

    Share

Comments have been disabled for this post