5 ms·

What do you think the employees do? I imagine there's a big team dedicated to getting inventory on the service through negotiations with music producers around

What do you think the employees do? I imagine there's a big team dedicated to getting inventory on the service through negotiations with music producers around the world. You'd need to keep them to continue expanding, adding newly produced music, and re-negotiating expiring deals. There will be a large editorial team producing playlists, curation, tuning the algorithm with expertise, etc. That all needs to be localised and to cover lots of niche music tastes. That's also not evergreen content or work, it needs constant updates to remain relevant. Then there's international expansion and enabling the tech to work for more countries, more languages, more payment methods, more types of music publishing. Honestly, where do you get "80%"?

maxglute 3y ago
>Honestly, where do you get "80%"? Lazy hyperbole and picked pareto 80%. Not remotely well calculated remark. Should have constrained comment to engineer/design team.
- danpalmer 3y ago
  Well thanks for owning up to it! I agree that it feels like the product from an engineering perspective was "finished" a while ago, however when I think about it for a while I can come up with many things that will need ongoing engineering work. For example, bug fixes (no one ever finishes this), support for new devices/phones/smart speakers/TVs/cars, support for payments and authentication in new regions, scaling work to scale to users in more regions, reliability and operational work. And that's just the stuff that isn't about growing more. I think it's reasonable to expect that businesses want to grow, and this isn't always anti-consumer either, as it can result in services that appeal to more users and cost less. This means engineering work on the marketing funnel, onboarding, new markets like podcasts, and lots more.
madeofpalk 3y ago
There's at least one team devoted to screwing up the homepage, pushing content that I don't want like podcasts. Will not miss them.
- maxglute 3y ago
  200m on Rogan has caused more than few of my friends to drop spotify permanently because they couldn't get him off their home screen recommendation for months. If only one engineer added a dismiss button.
  - danpalmer 3y ago
    I want to dismiss Rogan as much as anyone, but I think this underestimates the complexity involved. It's not just a button, it's what the button does, it's respecting opt-outs like that in a scalable way for many millions of customers, it's having a way to un-do those for users and for support agents when customers inevitably accidentally hide a podcast they want to see, it's GDPR data export and deletion of those flags to hide things, it's UX testing around whether users even understand what the button does. I agree that a good product would do all these things and have a way to not show recommendations you don't want, but I get why these things might not exist. When I worked in a <100 person startup we'd "just build it" and I'd probably spend a few hours on this, but we didn't have all these concerns. We didn't do user testing, we didn't care about scalability at this level. Now I work on Google Play, and if you want to add a button like this (that will need a database query) to a frontend it's a ton more work because the scale is so different.
    - madeofpalk 3y ago
      Yet somehow they managed to do the impossible and add a "hide from Discover Weekly" button.
    - ta1243 3y ago
      update users set "show_podcasts=0" where userid = 12345
      - danpalmer 3y ago
        If you think it through there is so much more to do here. This is the 3 person startup solution, but really doesn't get you much further than that.
apwell23 3y ago
Yea I worked at groupon when it was taking off and we had around 200 sales ppl calling up and negotiating deals with local business. Everyone back then used to ask me the same question: Why does groupon have 500 ppl, isnt' it just a wordpress site.
- dmd 3y ago
  People who make CRUD sites think everything is just a CRUD site.
  - Gooblebrai 3y ago
    Engineering-bias at its best. Thinking about the product complexity but not about the business
- makestuff 3y ago
  It is a shame for Chicago that groupon didn't continue to find success. I really think that if they had continued to grow the Chicago tech scene would be significantly larger than it is today.
  - boiler_up800 3y ago
    Absolutely agree. I’m not sure I can name another name brand Chicago tech company, which is devastating for their startup and venture ecosystem.
    - tempusalaria 3y ago
      Citadel. Chicago has been a big centre for developers for finance companies for a long time. Hedge Funds, Market Makers, Asset Managers etc.
      - local_crmdgeon 3y ago
        Citadel is moving to Miami because Chicago government sucks - the move was due to "riots, an employee stabbed, and the attempted hijacking of his Ken Griffin's car" https://www.businessinsider.com/ken-griffin-citadel-moving-headquarters-to-miami-from-chicago-2022-6?op=1 https://www.businessinsider.com/ken-griffin-citadel-moving-h...
aerhardt 3y ago
I find their algorithms for recommendation (song and artist “radios”) far inferior to 2000’s Pandora, so at least in that area I agree with GP that throwing 1000 times the resources at the problem doesn’t seem to be doing anything…
mrkeen 3y ago
> What do you think the employees do? Product stand-up, agile stand-up, team fika, cross-team fika, town hall, team activity, tribe activity, one-on-one, backlog grooming, team retro, incident post-mortem, ping pong, live music. If you do scrape together a few minutes to get part of the codebase into your head (despite the open-floor-plan chatter), you've got no way to trigger that bug you think you saw. But persevere anyway and diagnose it. Oh it looks like it's happening in someone else's microservice. File a JIRA? Bring it up at the next meeting? Try a different fix anyway. Happy with the fix? Wait in pull-request limbo until it's sorted. Release it straight into prod? You could bring down prod... Maybe you didn't really see that bug after all. Coffee time!
- smus 3y ago
  I think you're missing the point that coding isn't the only type of job
  - mrkeen 3y ago
    Do you think that the non-coders have more time to do their actual work than the coders?
    - danpalmer 3y ago
      I think your representation of a software engineer's work is made in bad faith. There may be some like this, but from anywhere I've worked and anyone I've spoken to in the industry, it's a small minority. I haven't worked at Spotify but I have worked with an ex-Spotify engineer who was very productive and one of the best engineers I've worked with. The representation is in bad faith because it overstates the balance of non-coding tasks, but also because it assumes these are "not work" or worth less than coding.
      - mrkeen 3y ago
        > but also because it assumes these are "not work" or worth less than coding. I once walked out of an improv class (on company time)
NickC25 3y ago
Spotify had a large headcount. Let's look at some of your examples and see where things can logically be cut. >I imagine there's a big team dedicated to getting inventory on the service through negotiations with music producers around the world. This can be streamlined considerably through standardized contracts. I'd wager 90% of the inventory that said team gets is from niche artists, independent artists or small labels. Very few artists actually have the clout or resources to negotiate their own deals with Spotify. That's sort of the point of having executives - to handle the Taylor Swifts of the world. Your local band, rapper, singer-songwriter or producer doesn't matter to Spotify. Have a clear and straightforward tier system for compensation. You get a billion streams a year? You get $X. You get 10,000 streams a year? You get $Y. >There will be a large editorial team producing playlists, curation, tuning the algorithm with expertise, etc. That all needs to be localised and to cover lots of niche music tastes. That's also not evergreen content or work, it needs constant updates to remain relevant. You don't need a large team to curate playlists when you can have users curate playlists. In fact, they already do. You don't need a huge localization team either. Users already take care of the legwork for you. Same goes for algorithm tuning. >Then there's international expansion and enabling the tech to work for more countries, more languages, more payment methods, more types of music publishing. Spotify is kinda already there in terms of international expansion. Anywhere they aren't already, they probably won't get there for political reasons (such as China). More languages? Sure, but that doesn't require a massive team. Fortunately, there are a few dominant languages around the world, and Spotify has those covered. More payment methods? Let the creators of those methods handle that - that's their job. Credit Card companies or banks or whatever don't create their own payment platform tech and just say "now it's yours to try to implement", they have entire teams dedicated to that. More types of music publishing? Like what? There's only so many methods of publishing music, and I'd wager there's going to be very little (if any) innovation there for a while. Spotify doesn't really distribute any other forms of published music besides digital streams. They aren't selling records or CDs, are they? EDIT: you could also say Spotify has vastly overspent on the likes of Joe Rogan, no matter how popular he is. $200 million dollars to a company that isn't very profitable, no matter how large, is just bad business. He's worth nowhere close to $200 million.

maxglute 3y ago

>Honestly, where do you get "80%"? Lazy hyperbole and picked pareto 80%. Not remotely well calculated remark. Should have constrained comment to engineer/design team.

danpalmer 3y ago
Well thanks for owning up to it! I agree that it feels like the product from an engineering perspective was "finished" a while ago, however when I think about it for a while I can come up with many things that will need ongoing engineering work. For example, bug fixes (no one ever finishes this), support for new devices/phones/smart speakers/TVs/cars, support for payments and authentication in new regions, scaling work to scale to users in more regions, reliability and operational work. And that's just the stuff that isn't about growing more. I think it's reasonable to expect that businesses want to grow, and this isn't always anti-consumer either, as it can result in services that appeal to more users and cost less. This means engineering work on the marketing funnel, onboarding, new markets like podcasts, and lots more.

madeofpalk 3y ago

There's at least one team devoted to screwing up the homepage, pushing content that I don't want like podcasts. Will not miss them.

maxglute 3y ago
200m on Rogan has caused more than few of my friends to drop spotify permanently because they couldn't get him off their home screen recommendation for months. If only one engineer added a dismiss button.
- danpalmer 3y ago
  I want to dismiss Rogan as much as anyone, but I think this underestimates the complexity involved. It's not just a button, it's what the button does, it's respecting opt-outs like that in a scalable way for many millions of customers, it's having a way to un-do those for users and for support agents when customers inevitably accidentally hide a podcast they want to see, it's GDPR data export and deletion of those flags to hide things, it's UX testing around whether users even understand what the button does. I agree that a good product would do all these things and have a way to not show recommendations you don't want, but I get why these things might not exist. When I worked in a <100 person startup we'd "just build it" and I'd probably spend a few hours on this, but we didn't have all these concerns. We didn't do user testing, we didn't care about scalability at this level. Now I work on Google Play, and if you want to add a button like this (that will need a database query) to a frontend it's a ton more work because the scale is so different.
  - madeofpalk 3y ago
    Yet somehow they managed to do the impossible and add a "hide from Discover Weekly" button.
  - ta1243 3y ago
    update users set "show_podcasts=0" where userid = 12345
    - danpalmer 3y ago
      If you think it through there is so much more to do here. This is the 3 person startup solution, but really doesn't get you much further than that.

apwell23 3y ago

Yea I worked at groupon when it was taking off and we had around 200 sales ppl calling up and negotiating deals with local business. Everyone back then used to ask me the same question: Why does groupon have 500 ppl, isnt' it just a wordpress site.

dmd 3y ago
People who make CRUD sites think everything is just a CRUD site.
- Gooblebrai 3y ago
  Engineering-bias at its best. Thinking about the product complexity but not about the business
makestuff 3y ago
It is a shame for Chicago that groupon didn't continue to find success. I really think that if they had continued to grow the Chicago tech scene would be significantly larger than it is today.
- boiler_up800 3y ago
  Absolutely agree. I’m not sure I can name another name brand Chicago tech company, which is devastating for their startup and venture ecosystem.
  - tempusalaria 3y ago
    Citadel. Chicago has been a big centre for developers for finance companies for a long time. Hedge Funds, Market Makers, Asset Managers etc.
    - local_crmdgeon 3y ago
      Citadel is moving to Miami because Chicago government sucks - the move was due to "riots, an employee stabbed, and the attempted hijacking of his Ken Griffin's car" https://www.businessinsider.com/ken-griffin-citadel-moving-headquarters-to-miami-from-chicago-2022-6?op=1 https://www.businessinsider.com/ken-griffin-citadel-moving-h...

aerhardt 3y ago

I find their algorithms for recommendation (song and artist “radios”) far inferior to 2000’s Pandora, so at least in that area I agree with GP that throwing 1000 times the resources at the problem doesn’t seem to be doing anything…

mrkeen 3y ago

> What do you think the employees do? Product stand-up, agile stand-up, team fika, cross-team fika, town hall, team activity, tribe activity, one-on-one, backlog grooming, team retro, incident post-mortem, ping pong, live music. If you do scrape together a few minutes to get part of the codebase into your head (despite the open-floor-plan chatter), you've got no way to trigger that bug you think you saw. But persevere anyway and diagnose it. Oh it looks like it's happening in someone else's microservice. File a JIRA? Bring it up at the next meeting? Try a different fix anyway. Happy with the fix? Wait in pull-request limbo until it's sorted. Release it straight into prod? You could bring down prod... Maybe you didn't really see that bug after all. Coffee time!

smus 3y ago
I think you're missing the point that coding isn't the only type of job
- mrkeen 3y ago
  Do you think that the non-coders have more time to do their actual work than the coders?
  - danpalmer 3y ago
    I think your representation of a software engineer's work is made in bad faith. There may be some like this, but from anywhere I've worked and anyone I've spoken to in the industry, it's a small minority. I haven't worked at Spotify but I have worked with an ex-Spotify engineer who was very productive and one of the best engineers I've worked with. The representation is in bad faith because it overstates the balance of non-coding tasks, but also because it assumes these are "not work" or worth less than coding.
    - mrkeen 3y ago
      > but also because it assumes these are "not work" or worth less than coding. I once walked out of an improv class (on company time)

NickC25 3y ago

Spotify had a large headcount. Let's look at some of your examples and see where things can logically be cut. >I imagine there's a big team dedicated to getting inventory on the service through negotiations with music producers around the world. This can be streamlined considerably through standardized contracts. I'd wager 90% of the inventory that said team gets is from niche artists, independent artists or small labels. Very few artists actually have the clout or resources to negotiate their own deals with Spotify. That's sort of the point of having executives - to handle the Taylor Swifts of the world. Your local band, rapper, singer-songwriter or producer doesn't matter to Spotify. Have a clear and straightforward tier system for compensation. You get a billion streams a year? You get $X. You get 10,000 streams a year? You get $Y. >There will be a large editorial team producing playlists, curation, tuning the algorithm with expertise, etc. That all needs to be localised and to cover lots of niche music tastes. That's also not evergreen content or work, it needs constant updates to remain relevant. You don't need a large team to curate playlists when you can have users curate playlists. In fact, they already do. You don't need a huge localization team either. Users already take care of the legwork for you. Same goes for algorithm tuning. >Then there's international expansion and enabling the tech to work for more countries, more languages, more payment methods, more types of music publishing. Spotify is kinda already there in terms of international expansion. Anywhere they aren't already, they probably won't get there for political reasons (such as China). More languages? Sure, but that doesn't require a massive team. Fortunately, there are a few dominant languages around the world, and Spotify has those covered. More payment methods? Let the creators of those methods handle that - that's their job. Credit Card companies or banks or whatever don't create their own payment platform tech and just say "now it's yours to try to implement", they have entire teams dedicated to that. More types of music publishing? Like what? There's only so many methods of publishing music, and I'd wager there's going to be very little (if any) innovation there for a while. Spotify doesn't really distribute any other forms of published music besides digital streams. They aren't selling records or CDs, are they? EDIT: you could also say Spotify has vastly overspent on the likes of Joe Rogan, no matter how popular he is. $200 million dollars to a company that isn't very profitable, no matter how large, is just bad business. He's worth nowhere close to $200 million.