free html hit counter But Before I Go... | John Battelle's Search Blog

But Before I Go…

By - November 04, 2008

I just learned my next piece is up over at the Looksmart series. It’s a rumination of sorts….

In ten short years, Google has become our social glue – we all presume that two people, asking roughly the same question, will get pretty much the same answer, and that answer will be correct. For most of the past decade, that was a pretty fair assumption. Google has become a universal search resource, reliable, accurate, and … consistent.

But for a variety of reasons, that assumption is no longer true. The ongoing goal of all search providers has been to personalize search – to tailor answers to the individual who is doing the searching. Search no longer takes one signal – your query – and finds results against the entire web. Instead it takes many signals – your search history, your geographic location, things you’ve clicked on in the past, files on your hard drive (if you allow it), and many others – and processes those signals against probable sub sets of data that have a higher chance of providing *you* the best answer. And that answer, increasingly, will be quite different from someone else’s, even if that other person asks exactly the same question.

Along the way, I think, something has been lost. It’s the same thing my mother lamented as she watched my generation abandon the newspaper – common ground, common spaces – a common set of facts around which we as humans can gather, debate, and connect. And therein lies an opportunity, I sense, to create a new kind of search that is in fact *not* personalized, but rather socialized – shared and common to all.


Related Posts Plugin for WordPress, Blogger...

8 thoughts on “But Before I Go…

  1. JG says:

    And therein lies an opportunity, I sense, to create a new kind of search that is in fact *not* personalized, but rather socialized – shared and common to all.

    Isn’t “socialized”, rather than “personalized”, the way Google already works? For example, Google uses the clickstreams of all the people who have done my same query before, and then gradually adjusts the ranked lists to boost those results that receive a lot of clicks. And then everyone benefits, because everyone sees the same, aggregately-adjusted (“socialized”) results.

    Or are you talking about something different?

  2. AP says:

    That’s an interesting observation, but what if we found out that search results are optimized for monetization? Yes, maximize monetization as opposed to user satisfaction. Follow the link to see the evidence this is happening.

  3. wonderful idea! the ministry of truth!

  4. mrg says:

    speaking of universal search & personalization, noticed that GOOG finally got patent today for #7,447,678 for the user interface to universal search. So broad, it’d seem any search that incorporates verticals in the mix could encroach on their IP.

  5. Nothing has been lost. Some may choose to have narrow focus. That will likely always be the case.

    You said,
    “we all presume that two people, asking roughly the same question, will get pretty much the same answer, and that answer will be correct.”

    Regardless of correct or not, there’s a presumption even prior to this. That being… there was some impetus to asking the question in the first place. For personally specific projects, it’s fine that algorithms attempt using a variety of background. For the “common ground” issues your mom lamented, it’s more likely than not that the primary cause of the search initation had some spark based on something coming out of – or at least emergent – in the Short Head Media. (That’s what I call the opposite of the long tail.) Of course any blogger or Twitterer or whomever could chance to get BuzzedUp or DugUp or ThrownUp via ThumbsUp icon clickers and so on. But for the really big stuff, most people always have, still do and for the forseeable future will continue to seek out major credible sources. (Anecdotes regarding a) the occasional fallibility of generally credible sources and b) the few anecdotes of social media really breaking something big notwithstanding.)

    The shared zeitgeist is not much at risk. There is a select group of major outlets that are essentially ambient in our lives. And when something really big happens, we seek them out. (And then blog about them.) And will continue to do so. Regardless, you make a major assumption that Google in particular – or personalization as a common feature of any search engine – precludes concurrent use of a more general transmission channel. Or that a filter can’t be easily added to results that says, “Show Raw” meaning “Dump relevancy weighting vectors that include behaviorial history.”

    It’s kind of like Chris Anderson’s Long Tail theory vs. Anita Elberse’s Harvard Buinsess Review critique of it. Is one or the other right or wrong? Maybe. But neither has to be. There doesn’t always have to be a winner take all answer to all things. Different models apply to diferent needs. In this case, Information Foraging Theory, (and Marcia Bate’s Berry Picking anologies in particular), tell us humans have certain fundamental types of information seeking behavior. A new tool or ten isn’t going to change that. At least, not for awhile.

    So perhaps we can all lament a bit less.

  6. Erik says:

    I just started chapter 2 of your book ‘Search’. I’m truly enjoying this book.
    ———————
    You could add to the
    Five W’s of Journalism:
    (WHO, WHAT, WHERE, WHY, WHEN)…
    not only the question “HOW”….but also formulate the most probing question: “IF”.
    ======================

  7. For personally specific projects, it’s fine that algorithms attempt using a variety of background. For the “common ground” issues your mom lamented, it’s more likely than not that the primary cause of the search initation had some spark based on something coming out of – or at least emergent – in the Short Head Media. (That’s what I call the opposite of the long tail.) Of course any blogger or Twitterer or whomever could chance to get BuzzedUp or DugUp or ThrownUp via ThumbsUp icon clickers and so on. There doesn’t always have to be a winner take all answer to all things. Different models apply to diferent needs. In this case, Information Foraging Theory, (and Marcia Bate’s Berry Picking anologies in particular), tell us humans have certain fundamental types of information seeking behavior.

  8. Harisankar H says:

    Isn’t it a trade-off between what an individual wants to know and what the society expects her/him to know ? In that sense, extreme personalization is a step towards anarchy in media while “common space” tries to mould an individual to the society. The holy grail is the right trade-off.