A fun Long Tail learning tool

Mike Vicic, "Dr TV" of the interesting TrivialTV blog, emailed me with a clever way to explain the Long Tail and Filters to the uninitiated. Here's his note, posted with permission:

When speaking with non-techie or non-media savvy people, I sometimes have trouble explaining how filters and relationships allow people to access the long tail for useful information.  I've used word counts and a thesaurus to demonstrate these concepts quite nicely.

I use the Wordcount site to demonstrate the shape and contents of the long tail.  People relate to this quite easily since you can have them read some standard (pre-written) text to reinforce the fact that the words in the head truly occur quite often while less common words occur in the tail.

Next I have them choose a relatively common noun or verb that is repeated in the text (and appears near the head of the curve), and we discuss the utility of that repeated word. We then talk about how replacement words are more useful than the original word since the replacements convey the same meaning to the reader and eliminate repetition.

Now the goal is to find these more valuable replacement words.  First, I have the person look through the Wordcount long tail for a word by moving the bar to the right and browse the window for appropriate words. As you'd expect, people get frustrated pretty quickly.  Before you know it, they're asking for a thesaurus. I then describe a thesaurus as a book that lists relationships between words.

I then present the analogy between this and a familiar long tail example...

   words = music
   head = songs that sell at least one million units
   long tail = songs they may have have forgotten or never heard
   thesaurus = [recommendation] software
   synonym = similar

...and explain that they'll use software to find music in the long tail that is similar to the music they hear on the radio in the same way they used a thesaurus to find a word in the long tail that is similar to the word they read in the text.

Comments

[David]

Chris, this is useful as an illustration. Still I think it's a pretty big leap to go from looking for words with the same (or similar) meaning to looking for similar-sounding music. For one thing, the idea of what's similar in music is much more subjective. But even more important than that, just because I might like Dylan or Radiohead or Philip Glass, it doesn't mean I want to find more music that sounds like them. The thing I like about them in the first place is that they're different from other things I listen to.

[Spy Tools]

I disagree with David, previous poster. I think the music industry is proof that similar music is very marketable, though of course there will be exceptions. FOr the most part however, we all have our niche and follow it. I like Nirvana, I also like groups that sound like Nirvana. The Beatles started a whole new sound which was duplicated for decades. Some sounds I don't care for, I'm not a fan of rap, any rap, just don't care for it, same with country music. What is it I like or don't like? The fimilar sounds associated with the examples I gave.

[Darin London]

But the 'likeness' aspect is vague in a really useful way, in that there are many ways for one thing to be 'like' another. The concept of 'like' can be defined as:

'people who like x also like y'

'people who tag x with z also tag y with z'

'people who tag themselves with z like x, if you tag yourself with z you may also like x'

etc.

Some or all of these ways can be exploited by people to search the tail.

[Ashley Bowers]

Thats very cool of him to let you post this note and let the rest of us read it.