Customer Login  |  

Robotics guru Masahiro Mori’s “uncanny valley” theory posits that as automated activity increasingly mimics human behavior there is a point just before the mimicked behavior is completely accepted where it triggers a strong negative emotional response. As our personal and business online lives become more dependent on an Internet that tries to divine our wants and needs and as we generate more and more information about our actions and personal preferences I think we are all collectively on the verge of reaching this “creepy” valley.

The most newsworthy indicators of this approaching “valley” are the backlashes to ad tracking software and those children’s game apps whose functionality is based on accessing “personal” information. However, an enormous number of other common Internet application functions are triggering a similar response.

For instance, LinkedIn and Facebook’s recommendations for people you might like to connect with always includes people you are close to, but this doesn’t mean that you don’t despise many of them or find the recommendations disturbing (the boss who fired you, the dreaded competitor, the ex-girlfriend, the deceased relative, etc.). Other types of personalization can backfire, too, like Amazon’s recommendations to your friends that you might like something when some of those things might reflect activity you prefer not to make public (e.g., books and videos on far too personal political or sexual topics).

Some of these things can be mitigated via some pretty simple functionality tweaks (i.e., LinkedIn allowing you to flag people with negative attributes or Amazon asking “Do you want to save this transaction to your personal preferences?”), but there is a far bigger question here: Because nobody ever actually reads the “terms of use” agreements to which they blindly agree, wouldn’t be a lot easier to use vendor relationship management (VRM) to make the results based exactly on the way you want your personal or corporate activity data used?

Doc Searls’ latest book, The Intention Economy, was just released last week and he makes a strong case for just this. The implications for going down this road are extremely profound for those of us in the information business. For one thing, company directories could disappear or mutate into permission-based publicly accessible databases. When will this brave new world emerge? The answer, like most things having to do with our networked world, is once again “sooner than you think.”

{ 0 comments }

posted by Shyamali Ghosh on May 14, 2012

The buzz about crowdsourcing continues to get louder every day. The crowd can now, it seems, handle a range of tasks including journalism, design, and even filmmaking.

If you’ve been asked to look into crowdsourcing as a way to stretch your budget and haven’t tried it before, however, you’ve probably asked yourself whether the results of crowdsourced projects are close enough in terms of quality to justify the pursuit of the cost-savings they can potentially represent. Put another way: “Are the results of crowdsourced projects good enough for consumers with low expectations, but not up to snuff for discerning professional audiences?”

The short answer is that, while not all crowdsourced projects are created equal, projects for B2B audiences do typically require an experienced professional to manage the process and perform quality control on the results, even if the crowd did the “heavy lifting.” Crowdsourcing is indeed a very effective tool for collecting certain types of data, but it takes an experienced manager to determine which particular data points in a data collection project are good candidates for crowdsourcing. It also takes know-how to draft the worker guidelines and set up the checks and balances needed to yield quality data. Without this, a project can require so much quality control work that any cost savings from crowdsourcing are wiped out. Done right, the process can yield a perfect data set collected at a very low price. Done wrong, it’s easy to end up with a data set of dubious quality that needs long and expensive clean-up to get it up to minimally acceptable professional quality standards.

Do you have a data project you’re interested in crowdsourcing? Contact IEI for an evaluation and a ten-page white paper on our crowdsourcing methodologies. Both are free of charge but available to qualified inquiries only.

{ 0 comments }

posted by Shyamali Ghosh on May 7, 2012