My Photo

« The Larry and Shelley Story | Main | Just in Case »

March 12, 2006

Comments

I really don't get MusicIP or Predixis. I am a musician and music fanatic with over 17,000 digital music tracks saved and accessible on multiple devices (iPod, computer, Treo, etc.). Like many people I know, I generally listen to my music in Shuffle mode, unless I am in the mood to listen to an entire album or focus on a particular genre or artist. Clearly, Steve Jobs caught on to the fact that "shuffle" has emerged as the prevalent mode of listening to digital music.

So, with this as background, I really don't get the whole MusicIP value proposition. I also have have tried the Predixis Discovery Engine and have found it to be a total waste of time. It has never returned any meaningful results and frankly, I think it's pretty lame.

There are several music matching, personalization engines and social networking products that I totally love and have changed the way I listen to and relate to music. Prime among these is Last.fm. On Fred Wilson's recommendation, I tried Last.fm six months ago and it's a great product. Of course, there's also Pandora.

Rick, clearly I'm missing something with MusicIP. Why should I care or use this product and how will it complement Last.fm? Help me out here.

Hi David,

Glad to try and help with some answers. Let's start with the basic listener.

Let's say on your music device or desktop you have 2000 songs. You just got a parking ticket and are now pretty pissed. You want to play some music to calm down. With a MusicIP ready device or software, you press/hold the play button and our technology puts the player into scan/mix mode. The player will now go 30 seconds into a track, play 5 seconds, then go to a different type of music on your player, 30 seconds in and play 5 seconds. It will keep doing this until you press the play button again. Once you do, our technology automatically re-orders the all the songs (200 or 20,000) in the order you will most likely enjoy them based on selecting the song that suited what you were in the mood to listen to. We eliminate the skip button. Just pick a seed song and let us do the rest.

The MusicIP mixer product is designed to help you get more out of your existing collection. If you start up the current application and you press the little green button, it will go through your 17,000 tracks, in groups, 30 seconds into the track, then plays 5 second of the track. It will keep going through you collection until you hear a song you are in the mood for and press the the green button. Once you do that, the application automatically re-orders your songs based on the order you're most likely to want to listen to them.

So the, shuffle/random idea is taken to the next level, i.e. personal random/shuffle.

2.3 million users have the application either via directly or as an add on to WinAmp. So, for your shuffle mode, we should have an option that works for you. In addition, the ability to make various types of playlists, save various moods, etc, all seem, based of user feedback, to allow the listener a great way to maximize the enjoyment of the collection.

On the discovery windows. Today, that is in test mode. It gives you varying songs based on what you are listening to, and, for the most part, people have given the company positive feedback. it is limited to CDBaby and Amazon. We are expanding the collections over time.

The next release does a couple of more interesting things. First, it seeds free music into the mixes that you are playing. So, while listening to songs, tracks from other artists will pop up that match what you are currently listening to. That feature was designed by a bunch of musicians. The idea was this.

Musicians want better ways to get in front of fans, new and old. If I'm in the mood for a particular song and I've told the mixer what I'm in the mood for, based on picking a seed, the application can feed tracks to potential fans when they are most likely to want to hear them.

For the artist, the process is simple, I play like Kenny G so get me in front of fans that like Kenny G. MusicIP levels the playing field for musicians. An artist registers with us, tells us where his track(s) are, samples or not, etc. Once we add him into the system, his tracks are now available to the millions of people that use our software to listen to and discover music. When a MusicIP customer is playing a mix, he will see the new artist's tracks in our discovery window at the exact moment the song is most likely to be well received by the music listener. It gives the artist unprecendented exposure, and levels the playing field so they can get more fans, sell more music and be more successful.

The idea is to give the artist another way to get in front of the fans in addition to giving the artist some feedback on where the track is showing up as in number of times we've seen it, when it is looked at based upon what other track was being listened to, etc. All of this is done automatically and without humans.

For the artist: get more fans to hear the songs, ensure the metadata is correct and give the artist some feedback on how the track is fitting into people's music/listening.

For the industry, its a matter of longest tail. Recommendation engines that do those that bought this, bought that or gene based, don't give recommendations based on what the person is in the mood for at that moment. When you are listening to a song and you ask for more 'like that', we believe that the recommendations should be based on the track that caused you to ask for more in the first place.

MusicIP also has a number of embedded customers. The iPod shuffle is a great product. Press the play button and you get random. Sometimes, tho, you hit the skip/next button. With a MusicIP device, you press and hold the play button, the device will scan your collection by playing a few seconds of the track. When you hear what you are in the mood for, you press it again and the device automatically re-orders the tracks for your mood, based upon the song you just selected. As I mentioned, you can try that feature out with the desktop application.

Rick,

I appreciate your response. I must say, though, that it hasn't clicked in yet for me.

I can't imagine a scenario where I would sit there mindlessly for who knows how long and listen to five second clips until I found something that I was in the mood for AND wanted to hear more of that type of music. But, hey, that's me. I'm a shuffle type of guy. Maybe it works for others.

Personally, I'm interested in discovering new music and getting connected to others that share my interest in music. That's why Last.fm is such a great product. That's also why Amazon's referral engine and the new iTunes personalization engine are so popular and effective.

As for MusicIP, I wish you guys luck and hope for your sake that there are a whole bunch of folks out there that are constantly in the mood for a particular style of music and can't figure out how to do it themselves. Personally, I don't see it. I have never encountered any of my friends or musicians who have said that not finding the right mix of music is a pain point for them.

Hi David,

Ah the life of a VC. So far, knock on wood, we've gotten lots of great feedback from users, artists and music resellers in the ecosystem. Keep your fingers crossed.

By the way, the mindless part. The issue with shuffle is a good point. All I'm saying is that on a device with our software, that shuffle becomes a more personal random instead of straight random. Not for everybody tho.

Again, thanks for stopping by and the civil discussion, that part is ALWAYS appreciated.

Rick,

I'm interested in what were the key factors that led you to invest in Predixis. Looking over your portolio, it doesn't strike me that you or JLA had any real background (personal or firm) in the music industry or in mass consumer plays. Also seems like a long way for you to go (L.A.) to invest in a first-time music play.

Did you review Pandora, Last.fm, Music Match and other engines during your due diligence? Wondering what you thought of these companies and their approach to the market. Also, what do you expect from Amazon as they move to deploy an integrated strategy (device and music store) that will surely leverage their referral engine.

Seems to me like a lot of noise and a lot of stiff competition in this market. What did you see that made you reach for your wallet?

David,

There were a number of factors and, yes, we reviewed all the companies you mentioned.

It will be fun to watch it play out. Send me a link to your music.

Great stuff, Rick!

I came by to make a comment about the website, though -- I just went through most of it, and once I'd finished reading it, I still didn't feel like I had a good idea what the product actually did, and it seemed like a lot of the _really cool_ stuff was kinda hidden.

For instance, I get the impression that it analyzes tracks by "listening" to them, so that it will be able to match things by style to tracks I've personally ripped off an old LP even if the database has never heard of the artist. To me, that sounds really cool. I found out about it in the middle of the FAQ, and I'm still not sure how accurate it is.

Also, what you said in the comments, above, that it can match up freely-downloadable music to what I'm listening to, while I'm listening to it. That's really really cool. Combined with the above, it means that I can find more stuff like things I've got on old LPs (including oddball things like the Eagles' "Journey of the Sorcerer" or James Galway's version of "Dueling Banjos" that are very different from everything else on the album, where I want something like _that song_, not like the artist in general), and much of the stuff it would find is free.

There's one more thing I didn't see obviously on the site: Privacy policy. Does MusicIP get a list of what music I have on my computer? Is the info about what songs have been suggested to me, and which one's I've downloaded, stored somewhere? Does that info get shared with anyone? Tell me why I don't have to worry that the RIAA might use this to audit my computer. (I presume any company you'd sponsor is honorable and upstanding, but I still want to see the legalese.)

Hope that feedback's useful, and good luck with it.

Hi Brooks.

Thanks for the feedback on the website, it is a work in progress and more work needs to be done to get the messages a bit tighter. The privacy policy is crystal clear. We don't link you with your music in anyway. On finding music, yes. Once you have your music analyzed by, us, we will point you to tracks/songs that match the mood you are in based upon what song/track you are listening to at the moment.

It is targetted at the music you own and like. We help you get maximum enjoyment out of what you have and then show you more music based on the mood of the moment.

For the free music, the new release has a slider that allows you to make a mix on a track you are in the mood for and then have the application feed new music into the mix, the amount/percentage you can select.

It's fun stuff.

Rick,

One last comment (promise).

Is MusicIP exclusively a desktop application or do they have -- or have plans for -- a thin-client version for use on mobile, portable or alternative (e.g., Musica/Symphony from Olive.us) devices?

The reason I ask is that increasingly the management, delivery and listening of music is moving away from the desktop. Best example right now is the iPod. I listen to most of my music directly on my iPod when traveling or on the road and on my Bose Soundock when at home (the Apple Hi-Fi will certainly expand the market for non computer delivery of music at home). I almost never listen to music on my desktop anymore. How can MusicIP help me analyze my music or create personalized playlists or recommendations if I'm not using my desktop?

Second, how will MusicIP work in situations where users are downloading and managing music off the air on their Rockr phones or via Sprint's or Verizon's new wireless music stores?

It strikes me that there is a sea change in the way consumers are going to be purchasing, managing and listenting to music. Music recommendation and personalization engines will need to adapt by being deployed as embededded thin-client applets or as a real-time Web service.

Thoughts? As Mark Evans suggests, I would be very interested in Fred Wilson's take on MusicIP.

The comments to this entry are closed.