GnuMusiq/topic subscription

From HypertWiki
< GnuMusiq
Revision as of 14:30, 8 September 2013 by Woozle (talk | contribs) (extracted from design page, and expanded)
(diff) ← Older revision | Latest revision (diff) | Newer revision → (diff)
Jump to navigation Jump to search

A topic subscription is a means by which someone can be notified whenever there is a new release (something for sale) that fits a certain topic. The idea of "topic", however, is quite flexible and could be used to categorize releases in many different ways, including (but not limited to):

  • artist
  • label (especially small labels)
  • performer (an individual musician who worked on the project)
  • songwriter
  • recommendations by others (see #Ecosystem for some expansion on this idea)

For each topic, users could choose whether (and how) to be notified and whether to buy automatically, and could also set purchase limits either by frequency (e.g. no more than once a week, or not spending more than a given amount per month).

Benefits

The main benefit is that it removes much of the element of chance -- and hence much of the need to advertise -- from the business of finding the music that you want to hear (and are likely to enjoy).

Another benefit is that it lets the artist know how many "units" they are likely to sell, which allows much better planning of the release. Overprinting represents waste, in that the extra copies often have to be sold at a discount or else take up expensive warehouse space until sold; underprinting represents waste in that the money those missing copies would probably have made will have to be deferred to another press run, by which time the disappointed customers may have lost interest (or bought used copies). Carefully tracking audience interest helps the artist make the optimum press run and thereby maximize net income.

Furthermore, this concept works extremely well in other media, too. For example, I'll buy anything Lois McMaster Bujold writes – in hardback, if I have the money. Often I don't find out about a book in time, and it's only available in paperback. Or not at all. If I had a Series Subscription for her works, I'd know in advance and perhaps even have time to scrape the money together for a hardback.

Ecosystem

The obvious, simple way to handle recommendations is that any user (call these users "reviewers") could recommend any release, and any other user could "subscribe" to a reviewer's recommendations feed, and thereby either receive notifications or automatically purchase the items recommended.

However, this is only the iceberg's tip with regard to what could be done. I suggest the following refinements:

  • allow reviewers to rate each release on a numeric scale (as is commonly done now, e.g. on amazon.com)
  • allow users to pick a threshold for notification (e.g. only notify about releases with at least 3 stars)
  • allow users to combine multiple reviewers, so that no individual reviewer controls the user's notifications or purchases
  • allow users to rate reviewers numerically, and provide the user with weighted averages across all their chosen reviewers, so that the reviewers who best match the user's taste count the most towards what that user sees or purchases (this is basically weighted proxy voting applied to artistic preference)
  • allow reviewers to create separate topics by which to rate releases (genres, production, songwriting, performance...)
  • on the site's front page, display some combination of:
    • the overall highest-rated releases
    • this overall highest-rated reviewers
    • the releases rated highest, weighted by reviewer's overall rating
    • topics with the highest-topic-rated releases
    • ...and so on; experimentation will determine what works best

This will create a rich music-appreciation ecosystem that will allow the best music to rise to the top and find its audience, with a minimum of advertising or other promotional efforts required from the artist. This, in turn, will level the playing field, and make talent and ability the primary factors for success.