So yesterday it was announced that Google has purchased Songza. Congratulations to Elias Roman and his colleagues. They build a great product and sold it to a great company.

But I’d like to take a second to tell the story of Songza as I know it. I am sure there are lots of parts of this story that I don’t know but the parts I do know make for a great story and now is a good time to tell it.

A few Brown University students had a great idea in 2006. They felt that mp3s should be priced based on demand not on a fixed price. So they started a company called Amie Street and built that service.

I first met them at some point after they had graduated from Brown and moved to NYC. I liked the idea a lot but was hesitant to invest. Others were not and they raised some money and chased that dream.

At some point Amazon got involved, I think as an investor. The Amie Street model ultimately did not pan out and in 2010 it was sold to Amazon. I don’t know the terms of that transaction but it did allow the team to stay together and work on a something else.

Long before the sale to Amazon, in October of 2008, Amie Street acquired Songza, a music app that was built by Aza Raskin and Scott Robbin.

After the sale of Amie Stree to Amazon, the team focused on Songza and iterated on it for a few years until they landed on the concierge user interface that helped popularize Songza.

I started using Songza in early 2012 and have been actively using it ever since.

I have three modes for listening to music and a primary services for each.

Passive – Songza, Intent Based – Rdio, Discovery/Social – SoundCloud. I use Songza the way most people use Pandora. And I use it mostly on my various Sonos systems.

But back to the story of Songza. Over time Songza built a popular music service and they raised some more capital in the fall of last year. We spent some time with them during that process but we were already knee deep in online music with Turntable (RIP) and SoundCloud.

Every interaction I’ve had with the Songza team has been fantastic. They are great people. And every interaction I’ve had with the Songza service has been equally good. Which furthers my view that great people build great products.

I wasn’t surprised to see that they sold to Google. The streaming music business is hard. And the big platforms understand that music is a great audience builder and retainer. And Google has been a great home to great products (YouTube, Android, Nest, etc).

So that’s the end of my story. It has a happy ending.

If there is a moral to this story it is that tenacity pays off. The Songza team graduated from college eight years ago and worked on two separate services over that time with a fair bit of success and failure. They hung together and built something that is very good. And they got a good exit. As JLM would say “well played.”

#Music#VC & Technology

Comments (Archived):

  1. aweissman

    It’s funny, alot of us have old Amie St/Songza stories. I met Elias and co back when you did (and back wken Long Island City was very different from how it is now) and fell in love with it – in fact almost 7 years ago I wrote this love letter. I was wrong, but I wrote it

    1. fredwilson

      yeah, putting your opinions on the internet for posterity comes with the need to say “i was wrong” a lot!

      1. LE

        Back in the day when people made the wrong decision and things got totally fucked up we were cool with it as long as they thought it through and had a reason for what they did, and the decision they made, and it appeared they had contemplated the pros and the cons.That’s one of the things I hate about Monday Morning Quarterbacks who act as if the coach making the decision didn’t have a reason for doing what he was doing at the time. Criticizing people for decisions they have made is what writers do from the comfort of their home office.

    2. Ana Milicevic

      Arnold, that’s exactly how I felt when I first came across Amie. Still hoping that model has a future.//Of course, I meant Andy and not Arnold, although not a bad thread to summon Arnold to, wouldn’t you (all) think? @awaldstein @JimHirshfield

      1. JimHirshfield

        There’s Andy @aweissman:disqus And then there’s Arnold @awaldstein:disqus Iceberg, Goldberg…who can tell the diff?

        1. Ana Milicevic

          … And then there’s jet lag and much blushing and backpedaling on my part.Thank you @JimHirshfield:disqus for making me laugh while I apologize profusely.

          1. JimHirshfield

            Happy to help any time. 😉

  2. William Mougayar

    Great story. You’ve had a few love affairs with music startups :)What does this say about the future of this segment? The margins are thin, due to licensing.And if $15 million is the true reported number, that seems relatively low vs. the Beats acquisition.

    1. fredwilson

      well i can tell you that Songza is 100x better than Beatswhich produces another moral to this storythe better product doesn’t always get the better price

      1. fredwilson

        whatever the price it was a bargain compared to Beats which is a vastly inferior product

        1. JimHirshfield

          Never used Beats, headphones or streaming service. So, yeah, not sure what Apple was going for there.

        2. Cam MacRae

          … and yet a marketing triumph.I won’t be hanging up my beyerdynamic cans any time soon, but I have nothing but respect for a brand that liberates the kids of $300 for a pair of shitty plastic ear-horns.

        3. LE

          whatever the price it was a bargain compared to BeatsIf so, relates to a failure (or success) of negotiation strategy as well in either transaction.It is a game. Some people play it better than others. Some people are also better gamblers.Case in point: Facebook paid 8.5 million dollars for they bought it from the American Farm Bureau Federation (see Not from General Electric and not from some gambling domain guy or portfolio owner. But from a non profit group.In transaction, who do you think got taken for a hay ride and who made out like a bandit?

          1. awaldstein

            I think a waste of money even if those dollars are irrelevant to Facebook.Names matter. URLs do but honestly not as much anymore and fading.Sure some people disagree and that is good for your business but in a world where brand not URL rules, it isn’t a game changer. Important. A hedge but not a business answer in most cases.I’m sure there is another side of this of course.

          2. LE

            think a waste of money even if those dollars are irrelevant to Facebook.For the record I own a 2 letter name that has a “b” in the 2nd position.That said the market for that name isn’t people reading AVC so I will give my honest opinion here.1) The fact that they have a boatload of money does matter. They piss away money on many things, there is no question that if they had only $1,000,000 in cash this would have been more foolish.2) Facebook has close to 7000 employees as of the end of 2013. That’s a lot of email. It’s nice to be able to do rather than “not that much anymore”. – Totally not what I’m seeing in the market.4) If “some people disagree” then why do you think that is?5) Some people spend outrageous amounts of money on wine and art. Almost the ultimate dutch tulip in a way. In contrast, people aren’t bidding up domain prices thinking they can sell them to a greater fool (at least not at this level it does happen is some aspects of the business at a lower level). So anyone parting with the money has a reason and a strategy for doing so usually.6) “Almost never worth what you pay for it.”. I don’t even know how you can even begin to say that. That would be like me saying “almost always worth the money”. It depends on specifics of each situation obviously. Or like me saying “doesn’t make sense to buy a coop in NYC for $20,000,000” etc. People have all sorts of ways they look at value. I have traded in cars that are only a year old. To me, for what I get, it “makes sense”. For my mom, it wouldn’t. And you don’t even own a car.7) “but in a world where brand not URL rules”. The business world is quite large. Not everyone who does business is a brand in the sense of what you are talking about. MSC Shipping (huge company) just tried to get legally (by UDRP) the domain and lost. (They are 2nd to Maersk in containers). While they might have a brand reason for the name perhaps they are just tired of their fucking email which now goes to at bouncing because people are used to typing at And that does happen I run mail servers and I constantly get people’s email errors.8) And so on.Lastly, my comments relate to the negotiation and what happened there, not as to the value of the name. Meaning knowing how these things work there is a good chance they could have gotten the name for much less that what they paid.

          3. awaldstein

            If I had a name I wanted to sell you would most definitely be the person I would hire to get the best price for it.With all the rest I respectfully disagree.

          4. LE

            With all the rest I respectfully disagree.Paradoxically if everyone thought like I did I would have never been able to get the domains that I got in the 90’s for next to nothing. [1]And there would not end up being an opportunity that I was able to exploit.To me it was crystal clear at the time [2] that they would have value in the future. The only thing I was wrong on was how much value they would have.[1] With practically no downside I might add. How many gambles have no downside?[2] When I was a kid my Dad had a phone number that ended in 1000 for his small company. I think this is what planted the seeds.

    2. John Revay

      Hi Wiliam, I saw that #, I thought it was a typo. If true I reconciled it was a better outcome than our friends over at had

      1. ShanaC

        That was a serviced I loved, but I had no idea how they would monetize

        1. John Revay


    3. John Revay

      Perhaps the # was $1.5B…..$15M seems too low of a number for Google to even think about doing a deal

    4. jason wright

      is there a regulatory process that examines valuations? Beats seemed like a form of money laundering on a gross scale.

    5. Guest

  3. Eliot Pierce

    I love how the Songza team embedded Walking on Sunshine in their acquisition blog post. It brought to mind Jack Black dancing around the record store in High Fidelity… Congrats Songza – looking forward to seeing Google integrate Songza with Nest. 🙂

  4. pointsnfigures

    They are young. This won’t be the last you hear of them. They had a good enough exit to keep them in the game and start thinking. What’s next?

    1. jason wright

      Russia 2018

      1. William Mougayar

        Good one.

  5. Chris O'Donnell

    I was very disappointed when Aime Street shut down. I spent many evenings deep in the rabbit hole of unknown bands on that site, and it was always such a thrill to stumble into something I loved there, and then buy the album for 99 cents because I was one of the first to find it. I hope somebody takes another shot at that model.

  6. Andrew Kennedy

    Great story. Love the moral and believe it. Thanks for posting this.”Tenacity pays off”Well said

  7. JimHirshfield

    Songza is great for those times around the house when you want to “set it and forget it”. Just play me some music that fits the mood. Congrats to my friends at Metamorphic Ventures – second acquisition in as many days (TapCommerce to Twitter).

    1. fredwilson

      yup. that’s how we use it. its great for dinner parties and such

    2. John Revay

      Yup….I was just a little frustrated this past weekend, when I wanted to set it and forget it….and I wanted a simple channel…and I found myself going down a path of two many choices/options.

      1. JimHirshfield

        Choice is hard.

    3. Ana Milicevic

      Metamorphic is having a great week, and it’s only Wednesday! 🙂

      1. JimHirshfield

        Hmm…you think something else is going down? cc @MetamorphicVC

        1. Ana Milicevic

          Well, their portfolio is pretty amazing.Also given this week they need to update the ‘Exits’ portion in real time 😉

  8. Robert Heiblim

    It is a good story and a good ending, especially as this space gets more crowded and serious. As you put very well, music is a great attractant and a difficult business. Whither the independents? Whither Pandora in this?

  9. kirklove

    Great story, great product, and very good to see this team getting a chance to further music discovery (which is a tough, tough road as you know).One quibble – you can’t really compare this sale price with Beats. Beats was mainly for hardware, not the streaming service which I agree was an inferior clone of Spotify and Rdio.

    1. fredwilson

      great pointi guess if we could know (we can’t) how much of the $3.2bn was allocated to the music service, we could have a convo about that

      1. kirklove

        Yeah tough. You also have to assume that 3.2B was for the “brand”. Beats can suck in a lot of celebrity, musicians, and athletes that Apple probably couldn’t do on its own. And as you also know athletes and celebrities can lift a platform into the stratosphere a la Twitter and Instagram.

        1. LE

          You also have to assume that 3.2B was for the “brand”.In “I can form an opinion when n=2” my stepdaughter (not even 10) and stepson (2 years older) both own beats. I think one was +-$200. They insisted on it as a graduation gift or something like that. Personally I hate the look and the colors. I would never buy it. I tried to get to the bottom of why they wanted it. Was very clear it was about what their friends had. [1] I tested them against my Sony $30 headphones and they did sound slightly better. But I can’t see a kid being able to tell the difference though. At least not enough for the cost of the headphones. I guess this is obvious.One thing about grabbing kids early is that they get easily sucked into brands (or events) because they haven’t been flooded with as much advertising and brainwashing. And it’s something they can relate to and call their own. (Reason they like Mickey Mouse easily identifiable in a confusing world).On the other side “live by the sword die by the sword”. Just as easy for something new to come along which everyone switches to (I mean a particular age group). So Beats was wise to sell when they could and cash out. Because their isn’t that much of a differentiator in the actual product to justify the cost. If Beats can get celebs to make their product take off so can someone else.[1] They actually didn’t even say this. They actually had a hard time explaining why they even wanted it.

          1. Alex Dunsdon

            As an aside…..IMHO Brand is undervalued in tech world. Just as tech is undervalued in brand world. People build brand associations like birds build a nest – over time…and the shortcuts they provide in making choices (when we’re evolutionarily programmed to think as little as possible) are huge. Especially to most people in the world who dont give a damn about tech.

          2. awaldstein

            The truth is that most in the tech world that build real brands understand they are not selling tech at all.You are correct, it’s few.

          3. ShanaC

            Well, what do you think would make it easier for tech to adopt branding

      2. Marty Gottesman

        $500MM was allocated for Beats Music according to WSJ ( That’s $2k per subscriber! ($500MM/250k)Spotify’s latest round ($4B valuation) was at $100 per active user (40MM) or $400 per paying subscriber (10MM).Songza has 5.5MM active users, which puts this at about $7 per user ($39MM/5.5MM).The Beats numbers don’t make sense that way. It was clearly a celebrity acqui-hire.

        1. LE

          As I say “when something doesn’t make sense there is probably something that you don’t know about it”. In this case my guess goes to tax and accounting reasons.

          1. Marty Gottesman

            Yes, the 2 divisions were backed by different investors. So the music investors (major labels included) got a great deal.

    2. CJ

      Lots of people seem to think Beats acquisition was more about the music service and talent(Dre and Jimmy) than the hardware. Hardware has value of course, but the consensus that I’ve read is that it’s secondary in the deal.

      1. kirklove

        Right, wasn’t clear about that part. Elaborated below. It’s about the “brand”. Shouldn’t have listed only hardware. Jimmy, Dr, etc, and the market and audience, experience, and voice they bring.

        1. CJ

          Connections, They know everyone in the music business. They bring artists, status, etc. I commented below that this could be a label 2.0 play by Apple to shake up the biz.

          1. kirklove


      2. LE

        Beats acquisition was more about the music service and talent(Dre and Jimmy)From a business standpoint I wonder what whopping type of key man or life insurance you buy to cover your investment if that is the case.

        1. CJ

          A metric shit tonne.

      3. ErikSchwartz

        I don’t believe either of them will be Apple employees in 24 months.

      4. Salt Shaker

        Beats does roughly $1.4B in sales (predominantly headphones) and a lot of that admittedly was driven by Dre and Lovine. Apple understands the importance of image more than any other company, both in and out of tech. Yeah, you can put a dollar value on talent, but at the end of the day it’s an expense on a P&L. “Talent” is potential and speculative, while hardware sales are reality. Money talks.

  10. Jorge M. Torres

    Happy for Songza this morning. Super happy for LIC and the future of tech here. We need more of these.

    1. ShanaC

      hurray for queens!

  11. Jon Miles

    Google seems to be making smarter acquisitions than Apple. It seems like to me that Apple is drunk on it’s cash. 1.5 billion for a lame streaming service and access to Best Buy that they already have? I don’t want to assign intelligence to what seems to be a bit of a blind strategy to me, but unless they are going to take all those people wearing headsets and give them glass as well I’m not sure what the heck Apple is doing. Meanwhile this guy is filming underwater 4k footage of his dog… and Apple’s flagship product is no where to be found.

    1. CJ

      Maybe. But the possibility exists that this is about much more than the headphones. They just acquired world class music industry connections, If they want to reboot online music this is how I’d go about it.

  12. Bruce Warila

    I had the pleasure of working with Songza when a startup I was involved in created a music preference-based ad targeting product. Google is fortunate to have Elias on board. He’s super smart and passionate about the space. I am looking forward to seeing what they do as part of Google.

  13. ErikSchwartz

    Can any music service be a big, profitable, stand alone service or will streaming music always be a loss leader or retention tool for a larger organization that makes its money some other way? I don’t think it is possible under the current rights laws. But even if those change I’m not sure there’s ever a high margin business here. Pandora pays a ton of money to rightsholders but even their top line revenue numbers are not mind blowing considering their market cap.

    1. CJ

      Yes, but not as currently envisioned. The money is in discovering, breaking and cashing in on the talent. Basically a music service has to become the label. Existing big names often don’t own enough of their own rights/masters/etc to allow this model to work but if you had sufficient capital you could break a ton of artists and see what sticks and create the next big music catalog that’s already enabled for the internet.Self-Publishing is what you want to emulate. Watch out if Amazon ever figures this out, they have the scale pull it out. So does Apple and especially with the Beats brand and Dre and Jimmy, they could totally create a Music Label 2.0 that could change the game for the music industry.

      1. LE

        Basically a music service has to become the label. One of my accounts owns the domain name and was thinking of this as a market for that (they were in a different type of “label” business.)

        1. JimHirshfield

          ^^ Nice Sponsored Comment LE

          1. JimHirshfield

            How much does it cost @domainregistry:disqus ?

          2. LE

            The idea is to suck as much money out of an interested party as possible.! Consequently it wouldn’t make sense to state a price publicly. But by way of comparison a similar name that was owned (by the same entity) we were able to get mid 5 figures for. Domain pricing relate more to who is buying and who is selling than anything else in general.

          3. LE

            I will no longer sayy you are funny and good looking since you didn’t send me my weekly tribute for doing so. You are in the 90 day past due category.

          4. JimHirshfield

            You’re looking younger and more handsome by the day. You been working out, bro?

        2. CJ

          Think about it, the only way you can make money in online music is if you don’t have to pay the extortion to the labels. How do you avoid that? Produce your own acts, build Motown part 2. Apple has enough capital to stomach the investment and when these acts take off, they have exclusives to the music, or even better, can now charge extortion to other services.

          1. ErikSchwartz

            Music sales are a loss leader for Apple.

          2. CJ

            Not sure I agree on that. I don’t think Apple knows what a loss leader is.

          3. ErikSchwartz

            Copied from my comment below…Last year Apple’s revenue was ~$170B and ~40% marginWithin that iTunes did about $6BWithin that music in iTunes did about $1.5B.The margins on that $1.5B were nearly basically zero.

          4. LE

            Why is that extortion though? (even though you use it in jest and excitement why do you call it extortion?).

          5. CJ

            It’s extortion because you have no choice but to pay it if you want a product in that market. Yes, technically it’s licensing fees, but the labels have a monopoly on licensing so they have no incentive to charge a fair market price. (One could even say that the fair market price is an extortionist price because there is no competition so no reason to charge fair-ish prices)But yeah, mostly just in jest. 🙂

          6. LE

            It’s extortion because you have no choice but to pay it if you want a product in that market.It’s a barrier to entry and a filter. I don’t see it as extortion at all.By your thinking aren’t supermarket slotting fees extortion?

          7. CJ

            Poetic license if you will. The fees are borderline insane so I said to extortion to make my point. It wasn’t meant to be literal.

      2. ErikSchwartz

        A&R is a cost center. Where will the revenue come from? What does “cashing in on the talent” mean anymore?

        1. CJ

          The Next Big Thing(tm) is often found by a label. That talent signs a contract granting the label $X, $Y, & $Z, in exchange the label provides money, marketing, etc. It’s widely known that the artist gets the short end of the stick. The label owns the masters, recoups all of their expenses and are now even gunning for a piece of merchandising and touring income.If you have a music service that owns its own talent and happens to have one of the best music execs and music producers on the roster that know how to make hits and find talent then you can deliver a label that once again OWNS ITS OWN DISTRIBUTION in a world where that model is dead. Cha-ching!

          1. ErikSchwartz

            Traditional radio was a parasite industry (ad supported) that existed to promote traditional recorded music sales.Traditional recorded music sales were predicated in bundling together (and charging for) 10-12 songs to allow the customer to get the 1 or 2 they wanted. The profit margin was in the bundle, they got you to pay $15 for a single song. — The bundle is dead and so are the profits.I still do not know what “cashing in on the talent” means these days?

          2. CJ

            What I mean is if you own the talent – the artists – then you don’t have to pay to lease them. Further, you can lease your artists to other music services. You simultaneously save money and earn it This isn’t practical for a traditional label because they don’t have the tech savvy to build a online music store, if they did, it would have happened by now.However, Apple already has a great online music store AND they just bought the talent to create a label. They’re building a vertical in the music industry, the only vertical left. They’ll own the talent and the distribution and profit from each in different ways. It wouldn’t surprised me if they bought American Idol too. That’s the only thing they need to complete the vision.

          3. ErikSchwartz

            You are missing the point.The profit margin in music sales came from selling the other 9 songs on an album that the buyer didn’t care about and would not otherwise buy. It was a trick to get you to pay $15 for a single. In an a la carte world where customers buy only the songs they want for .99 each where is the profit margin?

          4. CJ

            I think you’re missing mine too. :-)I’m not talking about recapturing that $14 in profit. I’m talking about reclaiming .70 per $1 for every hit song that you no longer have to license from someone else. The model that you refer to is gone but there is a new model that depends on paying an outside business a large percentage of every sale that you make.What I’m saying is that Apple could be putting themselves in a position to not have to pay that percentage for the songs that are in their catalog. It’s a longer term strategy but one that saves them a lot of money in the long term if they can make it work.The record labels take a not insignificant portion of every track sale, reclaiming that money for the tracks that they would own would be huge, over and beyond being able to license those tracks to others and serving as a draw to their music service over the competition. If you see them buying up masters or catalogs then watch out.

          5. LE

            The record take a not insignificant portion of every track sale, reclaiming that money for the tracks that they would own would be hugeYou say that as if record labels don’t do, and haven’t done, anything at all for the money they take it. That it all goes to the Queen of England or something like that. This is the same as the “movie studios control Hollywood” that gets bantered about. Or people who are upset because someone else has the domain that they want and they feel it’s some kind of public trust that is being violated. Or liquor licenses or taxi medallions.

          6. CJ

            Nope. Not saying that at all, nothing emotional about the argument. The labels charge a lot of money because they can and if you want to make a successful online music service you have to charge your customers more than the labels charge you – which can make you too expensive or not profitable. Or you have to supplement your income, often with ads, which makes your service less desirable.So, if you own the songs you put on your service you save yourself a huge chunk of change and change the economics of the game in your favor immediately.

          7. Marty Gottesman

            You’re speaking about merging two different business models. One is a music service. The other is investing in an artist, building their career, and making a return to cover your investment (and the other 9 that recently failed). I don’t see music services taking on that much artist development risk in today’s music market.The music services have done a version of this with established artists making exclusive recordings (e.g., Spotify Sessions). What we may see more realistically is a hybrid where some services could start releasing entirely exclusive artists, but they’ll also have to have the label catalog to supplement (like Netflix).

          8. ErikSchwartz

            Keep things in perspective.Last year Apple’s revenue was ~$170B at ~40% marginWithin that iTunes did about $6BWithin that music in iTunes did about $1.5B.The margins on that $1.5B were nearly zero (because of licensing and content costs). In your scenario the margins on that $1.5B would increase substantially. For the sake of argument let’s say to 100% (but more likely much less than that).Is going after that $1.5B incremental revenue a wise use Apple’s corporate attention? Or like I started with is this whole thing a loss leader for larger businesses?

          9. ShanaC

            sounds like old hollywood – cant last forever

        2. Morgan Warstler

          Erik, i’ll give you a specific example I heard last year from the guy running of the three largest labels.Justin Timberlake was created by N’Sync, his fame is not his alone from talent, and his future movie earnings – the label helped do that. Also any ads he is in, any clothing lines – true 360 deals.It actually reminds me a lot of the music biz in China.

          1. ErikSchwartz

            1 in 50 artists a label supports becomes meaningfully successful.1 in 50 of those manages to successfully cross over to other media.1 in 50 of those leverages those cross media deals into non-media brand deals.50 X 50 X 50 = 125,000Even if I am off by an order of magnitude, it’s still a crappy business.

      3. ShanaC

        Being your own label when you start out is very high risk

    2. Salt Shaker

      Good question. Yes, the licensing fees are screwy, but the biz model for these freemium services isn’t right either. There’s not a large enough diff between the free and pay components, hence free-to-pay conversion rates are still south of what they need/should be. The margins are in premium not the ad supported part of the service.

    3. george

      The key point here is “stand alone.” Perhaps it’s not enough to build and operate on a mainly a single product/service. Spotify and Pandora have both been around for awhile, they have revenue growth but profits lag or at best are comparatively meaningless.Apple iTunes is a multi-billion dollar super profitable business but it’s reach today is now much more than simply music, it’s the glue that exponentially keeps the ecosystem locked in. Google + Apple, they really are an exhibition on platform leverage.My observation and lesson learned, partnerships are essential for mono-product businesses, at least for now…

      1. ErikSchwartz

        I am dubious that iTunes music business is cash flow positive. If it is, it is very thin margin, and I suspect that those thin margins are not negative margins because the marketing budget is co-oped with the iOS marketing budget.

  14. Ana Milicevic

    I’ve also been a Songza fan dating back to the Amie St. days. It’s interesting how a finite set of curated playlists has proven remarkably effective for passive listening: for me more so than Pandora or Spotify where the selection of playlists, songs, and artists is seemingly endless. The one challenge I ran into was repetitiveness — on some playlists there would really only be about 6-8 songs in rotation so if you’re a heavy listener you need to alternate playlists (or moods) pretty frequently. Was hoping they would implement a way to subscribe/pay for a specific curated playlist so that I can finally have that ‘4th of July BBQ music picked out by Nick Cave’ one I’ve always pined for.

  15. sigmaalgebra

    Yes,there’s a lot in digital music and Internet and likely mobile music. And there has been some money to be made.But to me the main issues in music for the masses are discovery, recommendation, and curation. Can also add some value to notification and subscription.A fundamental challenge is the fact that for music very much need in some effective way to ‘get at’, well, the nominees are, artist, title, genre, composer, popularity, and artistic taste. May I have the envelope, please (drum roll please). And the winner is,artistic taste.And accepting the award is, what? Nobody. Why? Because at least for so far we see in the market nothing, nichts, nil, nada, zip, zilch, zero, for a good software based solution.There’s a question: How the heck to do that, that is, how the heck do we figure out what we tell the programmers the program is supposed to DO, with, of course, how the heck to do it? That’s not a question in programming, and the answer is not in computer science and, really, not on the shelves of the research libraries.So, sure, each rock, hard rock, country, hip hop, classical, easy listening, show tunes, gospel, etc. radio station has a ‘curation’, but this approach is old, manual, limited, and does poorly at most of discovery, …, etc.Got the techniques worked out and the software written, timed, tested, and documented. Doing ‘system test’, and so far nothing at all serious is wrong.

    1. Salt Shaker

      Not every biz decision can be solved with an algorithm. Some biz are too reliant on feel, touch, etc., or emotional components that can’t be captured or effectively measured. Music is one of them. All of these music biz are commodity oriented w/ minimal pts-of-differentiation. The competitive landscape is becoming increasingly more complex. Ultimately they’ll be consolidation. The market, and more importantly, the margins aren’t large enough for all of today’s players to effectively compete or achieve sig profitability. Certainly no prob for Google, Apple or Amazon, even if their biz are, at best, break even.

      1. sigmaalgebra

        Thanks for your thoughts.Your statements are fine and, at first glance, would appear to block my goals, but actually I don’t believe your statements really do block my goals. Instead, in a sense, my techniques are ‘another way through the woods’ and do not really encounter or attempt directly to overcome your points. Seeing the other way through the woods is a bit tricky.> Not every biz decision can be solved with an algorithm.Certainly true.> Some biz are too reliant on feel, touch, etc., or emotional components that can’t be captured or effectively measured.Also true, but I do believe that in some cases the challenges you mentioned, especially the “feel, touch, etc., or emotional components” can be ‘handled’ effectively.Also your “can’t be captured or effectively measured” are appropriate and essentially fully correct but, possibly surprisingly, in my view from my work, can be circumvented in that if we look carefully we don’t really need either to “capture” or “measure” in any very obvious, literal, or definite sense.I believe that my techniques will work much better than anything else that is in software and believe I have some quite solid reasons for my belief, but mostly to agree people will have to wait until my Web site goes live.Working on it. Back to it.

      2. awaldstein

        Got an example of a real business decision that was saved solely by an algorithm?No SEO or PPC, no computational errors corrected.I can’t think of one.

        1. Salt Shaker

          Didn’t mean to suggest that any “real biz decision was saved solely by an algorithm,” however a company’s critical pt of differentiation often is rooted in its alleged “proprietary” algorithm. Can’t tell you how many companies, part in the ad space, are positioned that way. Hard to question on the buy side, part when the folks doing the selling can’t articulate exactly what that means themselves. My comments were really specific to music software. I don’t find discovery is done part well. I think music today is quite ephemeral w/ appeal driven by a range of emotional and qualitative factors. My musical tastes are fairly eclectic and not very linear. “This for that” doesn’t really work for me. I actually like Fred’s Soundcloud feed a lot and listen to it pretty often (sans some of the hip hop stuff.)

        2. sigmaalgebra

          Airline fleet and crew scheduling done with 0-1 integer linear programming.Linear and nonlinear optimization for how to run an oil refinery. E.g., here are your candidate inputs and their prices and quantities, and here are the prices of your candidate outputs. Now how to operate the refinery? The main algorithm involved, certainly for the linear case and possibly also for the nonlinear case (with iterative approximation) is the Dantzig simplex algorithm, once rated as one of the best results in engineering in the 20th century.A simple, old textbook problem was, for positive integers m and n, we have m factories and n warehouses and need to ship some one good, e.g., widgets, from the factories to the warehouses. We have the mn shipping costs and the maximum each factory can produce and what each warehouse needs. Now, what to ship where for minimum cost? For this problem as stated, there is a very clean, astoundingly fast solution via the ‘transportation algorithm’ but, now, seen as a special case of the more general linear programming on networks. And apparently Bertsekas at MIT has an algorithm with running time guaranteed to be a polynomial in the size of the problem.I have to believe that Amazon and Wal-Mart have long since attacked versions or generalizations of this problem. If not, then they should give me a call!Here we need to upgrade a major portion of the Internet. We want progress in capacity, reliability, and latency. Here are your inputs and outputs. Now, your mission, and you have to accept it, is to specify what links and nodes will meet the need now, grow for the need for the next few years, at minimum cost now and expected minimum cost for the only probabilistic future. The problem is in combinatorial optimization and was the subject of the Goldman Lecture at Johns Hopkins by T. Magnanti, at the time Dean of Engineering at MIT. There has been at least one startup formed and funded to attack this problem. At least some business decisions have been made from the associated algorithms. Bell Labs was trying to do something similar way back; a lot of money was on the line; the figure tossed around is that their efforts saved ballpark 15% of capex, not too shabby.As in…the ordinary differentialy'(t) = k y(t) ( b – y(t) )caused two guys from General Dynamics to ‘decide’ to stay and, thus, not kill FedEx.There is the Black-Scholes option pricing model; E. Thorpe made money with something similar for a while.

          1. ShanaC

            basically hard industrial goods, some degree shipping of perishables.Stuff that don’t really touch consumers directly at all

    2. Salt Shaker

      Gee, my apologies if my comments were taken a bit too personally. I really was responding to music software that’s avail today, and not something you have in the pipeline. I hope you kill it, honestly I do, cause I’m disappointed in the products and services avail today! I find music software to be a commodity and I wish it wasn’t. The U/I is cumbersome and discovery is extremely time consuming and weak. Would love to learn/hear more about what you’re doing at the appropriate time.

      1. sigmaalgebra

        > I hope you kill it, honestly I do, cause I’m disappointed in the products and services avail today!Thanks for your “hope”. I’m trying.My first clue to get started was an effort to get music recommendations on the Internet. Like you, I was “disappointed” and guessed that there HAD to be a better way, put my feet up, thought about some of my best material in grad school, had some ideas, wrote them out, and typed in code.> discovery is extremely time consuming and weak.Very glad you noticed! I agree. So, there’s a ‘need’! Good! One of the standard pieces of advice for an entrepreneur is to find a need that needs to be met. Yup! Working on it.Main bottlenecks: The need for some 5000+ Web pages of documentation to cover what I needed about Microsoft’s .NET Framework, IIS, ASP.NET, Visual Basic .NET, SQL Server, and more. Some severe computer security and computer system management mud wrestling. Some exogenous events independent of the work. The work itself? Fast, fun, easy.

  16. Ian C Rogers

    Agree 1000%. I first met Elias in Amie Street days and became a fan of Elias instantly even though I was puzzled by the Amie Street model. I’ve been a vocal supporter of Songza and Elias ever since and couldn’t be happier to see him happy with the outcome.

  17. Twain Twain

    Founders and investors alike “win some, lose some, always moving forward”.

  18. shi11

    I would just add to that story, (we hosted Eric Davich last month here:… that after Amie Street when they started to focus on Songza, they iterated through at least four different product ideas before landing on the concept that stuck: curated, experience based listening. Tenacity is taking the necessary time to realize a great product.

  19. Jevon

    Such a great product, I’ve always wished there was a better reason to pay for it as I felt it was worth at least $10 a month, but they were asking for $4 a month for a few features that didn’t really impact my habits.Sonos + Songza really has to be one of the most luxurious things you could ask for. Great music, great sound, and with a minimal physical and app footprint.

  20. Kunal Tandon

    Some people say a good product is one that you happily use, and a great product is one that you tell your friends about…for me, Songza definitely fell under the great product category. Whenever my friends asked what apps I was using, or what was on my home screen, the conversation quickly turned into me gushing praise for the product that the Songza team built. I’m glad that the product will continue to exist with the support of Google.

  21. ShanaC

    Although I never fell in love with the product (I did try it a few times) it was extremely thoughtful.It makes me feel better that it takes time to iterate through products until you exit – that’s the best part of this storyThough you should expand upon your thoughts on passive/active.

  22. Chris Phenner

    Two footnotes (or perhaps, Endnotes):1. The Amie St. guys visited Thumbplay in NYC as I was there while they were still at Brown, and we have them affiliate links to sell ringtones — XML feeds before we even had APIs available.2. The Songza guys started in Chicago on Ravenswood Ave., and were as copyright dirty as imeem, but nowhere near as large, and were not being sued. They spoke about design as I visited them, but felt they were not doing anything wrong with copyrights.I recall them being such good listeners, saying very little until they can back to confirm that they had integrated.Good for them, they cleaned up nice.

  23. George Tziralis

    Fred, thanks for sharing this background story, online music startups seem to be really tough and every single success story is a great stimulus to those who are now starting. On a side note, I’d appreciate your take on, built to be a proper canvas for experiencing music on the web (you may start from here –… ; also, disclaimer: we are investors, and it’s still early days for the service). Thank you.

  24. John Revay

    Re: Soundcloud….I love this mashup track that Fred posted last May…Especially like this song called – Hang with me @ the 33:00 mark,

  25. John Revay

    HUMMMMI was scanning twitter last night and saw that @cdixon post two tweets about MusicSerious question: has decline of music business negatively affected the quality/quantity of music being created?— Chris Dixon (@cdixon) July 4, 2014And then saw this link come across digg today…I subsequently saw that Chris referenced this.