Re: [podcasters] Voting and podcast stats
- Evo sez-
> I'm not disparaging anyone from casting votes nor asking for them onMur sez-
> their shows. "To each his own". I've been guilty of it myself. In fact,
> we plugged it pretty hard on SoSF a few months back and hit #1 on PA.
> Then I grew disenchanted with the whole idea of constantly reminding
> people to vote, and now we're no where to be seen in the rankings. Our
> audience size (both subscribers and downloads) have increased since that
> time, which shatters the idea that a show with more votes has more
> listeners. It's not anecdotal for me any longer; it's fact.
I'm with Evo. I ran through the "Important" things the other day, and
I'd have to say "vote for me on podcast alley and subscribe through
iTunes and Odeo and yahoo and add me to your favorites on podcast
pickle and... and... and..."
Eventuallly we'll have to make an entire podcast based on all the
places you can vote. Might be something for our Keeper of Directories
to consider, eh, Rob? :)
Geek Fu Action Grip
I Should be Writing
- Podcast pickle origianlly had a 5 star rating method when we first started.
The problem with it was that people would rate a podcast with a bad rating
to lower it. Then by lowering it, they passed it and became a higher rated
On 11/2/05, alex_nesbitt <alex_nesbitt@...> wrote:
> We've been having a similar debate about how to rate/vote on
> We have used overall rating score, number of votes, number of
> inclusions in My Podcast favorites and outbound hits ( all with ip
> limits) to create lists of top ranked podcasts. I think we have two
> (and maybe more) problems with our rankings. First, ratings are
> on a 1-5 star scale and it's hard for a podcast that's been in the
> directory awhile to maintain a perfect score. Second, none of these
> measures take into account the number of days a podcast has been in
> the directory. Old podcasts really get an advantage over new
> for total click throughs and total votes.
> Here's how I'm thinking of changing our overall ranking - we need
> something that considers the curiosity factor (click throughs),
> quality (ratings/reviews), quality of ratings (number of votes ) and
> inclusion on favorite lists. Then we need to make sure we normalize
> for the number of days on the site to take listing age out of the
> So something like clickthroughs/day*weighting factor
> +votes*ratings/day*weighting factor +favorites/day*weighting factor
> What do you think about this idea?
> Digitalpodcast.com <http://Digitalpodcast.com>
> Yahoo! Groups Links
[Non-text portions of this message have been removed]