New (beta) Podcast Directory

[rick]

Hi folks,

I've just launched a new directory at http://podcastlikethat.com/. Some of you are undoubtedly thinking right now, "Another directory? What's the point?", but I'm going for a much different approach this (the Alley) and other sites like it. Instead, it's designed to link podcasts together by similarity, so by looking up your own shows and/or the ones you otherwise listen to, you can have other shows that match your interests recommended to you.

Being new, things are a bit sparse at the moment, and I'll be rolling out a lot of the best features in the coming weeks. That's why I'd like all you podcasters out there to send in your feeds! The more podcasts we get, the easier it should be for you to find other podcasters into the same topics, and for their listeners to find you.

If you have any questions, or if anything breaks (it is beta, after all), feel free to email, PM or just reply here, I'll be hanging around.

[BSquared]

Not to burst your bubble or anything but there is already amigofish whish is a ratings/similarity based directory http://www.amigofish.com/

Is there something unique about yours? The way the 'similarity' is calculated perhaps?

I'm genuinely curious.

[Soccer Shout]

Good news: I signed up.

Bad news: Doesn't iTunes do this already? And with a heck a lot more data.

I'm interested in understanding how you're going to collect your data.

I think iTunes is: If person A likes podcasts 1 and 2, and person B likes podcast 1, then person B will also probably like podcast 2.

I can see 2 ways of doing this: An user-association engine like iTunes, or a keyword association ("If you like "Soccer Shout", then you'll probably also like "World Soccer Weekly"....and "Right-Wing Shout Off")

I think user-association works better (it relies on people to understand subjects), but the data is hard to come by.

[agile]

Not to burst your bubble or anything but there is already amigofish whish is a ratings/similarity based directory http://www.amigofish.com/

Is there something unique about yours? The way the 'similarity' is calculated perhaps?

I'm genuinely curious.

I have found Amigofish to be pretty useless. The recommendations rarely match up to what I enjoy.

If you can get genuinely simililar podcasts grouped together then you will have a winner.

Regards,
Tim

[Shawn]

[quote=BSquared]If you can get genuinely simililar podcasts grouped together then you will have a winner.
I agree. I know that, in my case, my show is similar to other (dare I say it?) top-rated shows, and I think the fans of those particular programs might enjoy mine. But the problem is, getting those fans to give my cast a listen. So, if a directory could make these kinds of associations, it'd be helpful. And not just for me, but for everyone.

[rick]

I can see 2 ways of doing this: An user-association engine like iTunes, or a keyword association ("If you like "Soccer Shout", then you'll probably also like "World Soccer Weekly"....and "Right-Wing Shout Off")

It is, indeed, a keyword association where relevance is calculated based on the tags that two shows share and how often a tag is used. Everything is taken from the feed, so data collection isn't an issue - that said, it currently relies heavily on the podcaster providing good <itunes:keywords>, which many don't, so I'm working on extracting keywords from the description and show notes to improve the reliability of matches.

I think user-association works better (it relies on people to understand subjects), but the data is hard to come by.

Not necessarily. People can like or dislike things for all-sorts of strange, subjective reasons that marketing types get paid lots of money to analyse, but computers are poor at identifying. You might like a podcast because you think the host is attractive, I might hate it because I think the host is a self-absorbed tit. Neither of those have anything to do with the subject of the podcast, and don't necessarily reflect on what someone else will think. Highly relevant shows content-wise can get buried due to being unknown or unpopular, so better to keep the basic comparison clinical, and let people weigh user reviews themselves, if they care to.

Not to burst your bubble or anything but there is already amigofish whish is a ratings/similarity based directory http://www.amigofish.com/

Not at all :). I never claimed that what I was doing was completely unique, merely that it isn't the same lists/categories/features formula as the Alley and many other directories. There are a few (eg, PodShow) which have a superficial "Similar Shows" system, but I'm aiming to be much more complete and accurate.

[Soccer Shout]

Well, all the best to you, Rick. If you can pull it off, we'll all benefit.

I'd encourage people to submit their feeds.

[EclecticMix]

Best of luck to you, though I think that this will be a hard sell. For instance, I'm not sure how one would match my Eclectic Mix up with anything. Sure, I play music, but one week I may play Classical, the next Punk (yes, I've done this), so anyone looking at keywords will be disappointed in this case. OTOH, my One Minute How-To is similar to other educational podcasts, but that begs the question as to whether or not a simple breakdown by "type" of podcast might serve the listener better.

I did try entering "http://Feeds.Feedburner.com/EclecticMix" into the form and got a 500 error. It would not allow me to try again because it said it had "already been fetched in the past hour." My other podcast apparently worked, but since you update only weekly and mine comes out three times a week, it will regularly be out of date (I seriously doubt that I'll want to remember to manually ping it following each change, so a feature that indicated frequency might be useful).

Cheers -

george

[rick]

I did try entering "http://Feeds.Feedburner.com/EclecticMix" into the form and got a 500 error. It would not allow me to try again because it said it had "already been fetched in the past hour." My other podcast apparently worked, but since you update only weekly and mine comes out three times a week, it will regularly be out of date (I seriously doubt that I'll want to remember to manually ping it following each change, so a feature that indicated frequency might be useful).

This seems to be a repeat of the problem (http://podcastlikethat.com/blog/2006/10/03/114430#comments) I had last night with some enclosures not providing the information I'd expected. Sorry for the inconvenience, it's been fixed and you can find it here (http://podcastlikethat.com/podcast/5956b81d1183f841be5f573a443fb7d8d748cbd5).

Best of luck to you, though I think that this will be a hard sell. For instance, I'm not sure how one would match my Eclectic Mix up with anything.

As I've said, things are a bit hit and miss right now depending on how good your <itunes:keywords> are. As soon as I get it to trawl through the show notes, it should in the very least match you up to other podsafe music shows with the genres listed in the description. You do have a point that it's going to struggle given your show notes don't explicitly mention which genres are being played each episode. It could well be worked around, though, so I'll see what I can come up with.

[Bucket]

Same thing happened to me, and it appears we are not listed. Help?

The feed is http://feeds.feedburner.com/blogspot/AsPB. Thanks.

[rick]

Same thing happened to me, and it appears we are not listed. Help?

Odd, I just pinged it now and it worked fine. If it was yesterday that you tried, it could have been something fixed since.

http://podcastlikethat.com/podcast/4c9df4dff7a90a4855d7c9a69240338ebe44e522

[Soccer Shout]

Regarding the 'ping', see if you can figure something out with mr 'triyo.com'.

I DO use his 'ping' service everytime I release a podcast.

[Craig]

I'll be very surprised if you can accomplish your goals based strictly on keywords. Just because I like one photography podcast, for example, doesn't mean I like others. It seems like you'd also have to factor in listener ratings and come up with a much more complex algorithm in order for the recommendations to truly be useful. After all, Netflix wouldn't be offering $1,000,000 right now for ideas for a better recommendation algorithm if it was that easy.

Craig

[frankmcma]

Rick one huge thing you have done right is the submission process.

Paste in your feed, hit submit. Done.

It sounds logical, after all, all the data is already in the feed, but I have seen countless directories that have pages of forms for manually typing in every possible show detail.

Thanks at least for making the submission process quick and painless..and good luck!

Frank

[BSquared]

I'll be very surprised if you can accomplish your goals based strictly on keywords.
Craig

I agree. For one thing most shows have far too many keywords. But, far more importantly, there are far more factors in what make up my list of favourite shows than whether they share any tags.

The best "if you like this then we think you'll like this" site that I use regularly is Amazon. Not that I always agree with their recommendations (some are way off) but they come closer than anything else that I use inlcuding any of the podcast directories...I think because amazon use a combination of controlled genre/keyword/tag plus user ratings/reviews plus user purchase habits. Of course they also have huge numbers of users which adds to the statistical validity. I'ts much more likley you'll get statistically meaningful results when you have millions of customers/users than when you have only a handful.

So, while I wish this new enterprise all the luck in the world I can't see it working with (a) no control at all over the keyword/tags (some podcasts assign multiple irrelevant tags to their shows 'cos they can and 'cos they think it will get the more exposure) and (b) no human input into the process.

I hope I'm wrong.

[agile]

I'll be very surprised if you can accomplish your goals based strictly on keywords.
Craig

I agree. For one thing most shows have far too many keywords. But, far more importantly, there are far more factors in what make up my list of favourite shows than whether they share any tags.

The best "if you like this then we think you'll like this" site that I use regularly is Amazon. Not that I always agree with their recommendations (some are way off) but they come closer than anything else that I use inlcuding any of the podcast directories...I think because amazon use a combination of controlled genre/keyword/tag plus user ratings/reviews plus user purchase habits. Of course they also have huge numbers of users which adds to the statistical validity. I'ts much more likley you'll get statistically meaningful results when you have millions of customers/users than when you have only a handful.

So, while I wish this new enterprise all the luck in the world I can't see it working with (a) no control at all over the keyword/tags (some podcasts assign multiple irrelevant tags to their shows 'cos they can and 'cos they think it will get the more exposure) and (b) no human input into the process.

I hope I'm wrong.

I seriously can't understand all the negativity from some people in this thread about the project.

Hats off to you Rick for trying something innovative. Sure it may not work out perfectly at the first attempt but keep at it and you could have a huge slice of the podcast directory market.

Best of luck with it,
Tim

[WyethDigital]

I added our podcast and the process was smooth on our new one (yet to be officially announced), and our existing show was not so smooth. I received a error from the server. It eventually took.

I like the clean layout, the easy submission process, and have no opinion about the "what other people like" feature -- it's only interesting to me as someone that is curious about what other members of my audience like. As a consumer of a podcast, I could really care less what my fellow listeners like (unless it's presented to me in a review from Tim or Bernadette ;) )... But lots of people do like it, so if you can get it working the way you like it, I'm sure it'll add great value to the site.


Eric

[Hittman]

Rick one huge thing you have done right is the submission process.
Paste in your feed, hit submit. Done.

The other day I visited a directory I haddn't seen before. I searched for my show, and it wasn’t there. So I clicked on the “Enter a podcast” link, and was told I’d have to register to continue.

I said “Goodbye” and moved on to something else.

[Craig]

I seriously can't understand all the negativity from some people in this thread about the project.
It's called constructive criticism...people are taking the time to offer suggestions that will hopefully help someone succeed where they might otherwise fail. Rick can choose to ignore it if he wants or (as he seems to have done in his previous post) take it into consideration while developing his site.

"This sucks..." = negativity
"This would be better if..." = constructive criticism

Craig

[agile]

I seriously can't understand all the negativity from some people in this thread about the project.
It's called constructive criticism...people are taking the time to offer suggestions that will hopefully help someone succeed where they might otherwise fail. Rick can choose to ignore it if he wants or (as he seems to have done in his previous post) take it into consideration while developing his site.

"This sucks..." = negativity
"This would be better if..." = constructive criticism

Craig

I understand what you are saying Craig. I just think that if I were receiving some of the comments that Rick was getting I wouldn't feel too encouraged.

-Tim-