The Problem with Suggestion Engines

The pervasive, and growing, number of suggestion engines concerning topics of all sorts have become a de rigeur aspect of social networking pages. I intentionally didn’t say social networking sites, because, as we discussed on the podcast lat week, social networking is becoming more a regular part of ALL sites as opposed to platforms in themselves.

Whether it’s film, television, music, books, websites, cars, etc., the ability of web users to rate and comment on anything under the sun is facilitating the aggregation of opinion. And popular opinion, right or wrong, often gets confused with fact… but that’s another issue.

There is an inherent danger in recommendation engines that is often overlooked and has nothing to do with the truth or fallacy of the opinions used to generate the recommendations. Recommendation engines homogenize choice. Discovery of the new, unique, and radical is, by nature, stifled by numbers.

If 1000 who liked movie A also liked B, C, and D, the “push” aspect of adhering to such an engine would force me down a road that reinforced a opinions of a cultural demographic instead of a spirit of exploration (excuse my esoteric flight of phrase). You’re taking recommendations from a website, with a certain demographic of users that is even greater constrained by the subgroup of users that actually contributes their opinions (via checks or stars or thumbs). That subgroup becomes a further subgroup of the particular afficionados of whichever particular genre you’re searching for. The model breeds laziness in exploration and a funnel-like status of recommendation cramming that, unless abandoned, will perpetuate a vicious circle of “now 1001 users who like movie A also like movies B, C, and D.”

I only have so much time in the day… it’s a shame that, while, movies J, Q, and U are fantastic, I won’t have any time for them after D.

Check for the new podcast next Monday, where my website of the week will be related to a new spin on music recommendations.