Interview with Arkadiy Kukarkin, developer at the Hype Machine
How about an introduction: *Scroll down for the interview!
The Hype Machine is a website that keeps track of what music bloggers post on their blogs. A selection of blogs are selected by the makers of the website. The careful selection of the blogs is what makes the quality of the website: official labels, Dj blogs with their own mixes and blog run by promoters of parties are not included in the Hype Machine’s database. The Hype Machine follows bloggers that have a pronounced passion for music, and are sharing (good) music no matter what. Each song can be listened to from the Hype Machine, and there is a possibility to follow the link to learn more about the song and read its original post.
The website recently reached one million users, and it has grown to be one of the most popular online music websites since it was created in 2005 by Anthony Volodkin. In 2008, the Hype Machine was chosen in lostintechnology’s Top 10 Ways to Discover New Music. It was then featured among the Guardian’s top 100 essential websites in 2009. In 2010, Lifehacker included it in its list of best five music discovery services. You get the idea: Hype Machine is among the most respected websites to discover music since its launch. Interestingly, it is also one of the few ones that is still up and running as opposed to many others from the preceding ‘top’ lists.
How it works:
The Hype Machine tracks different mp3 blogs. If the post contains an mp3 link, the Hype Machine links it to its database and adds it to its latest list. Registered users of the Hype Machine can show that they like the song by clicking on the heart next to the song. The songs with the most hearts then climb the rankings in the popular list. The popular list is far from your local radio’s top40: the Hype Machine’s default popular list displays the most favorited tracks from the last 3 days. Once the song is older than 3 days, the song will disappear from the popular list.
“For a track to be eligible for being on the Popular page, it must have been blogged in the past 3 days, and received some amount of new favorites. The tracks are placed on the Popular page in order of how many new favorites they are getting. To prevent a track from staying on top of the chart by being constantly reblogged and favorited (because it’s already number one), there is a time limit of 3 days for its presence. Once the 3 days pass, the track can’t enter the Popular page for at least two weeks, so that the charts remain fresh.”
They have also decided to remove the number of times a song was favorited in the popular section because of its distracting nature. They have however added the rate of new favorites, by adding a small bar display that expands to a larger graph of favoriting activity during the past week. Typically, the spikes in this graph help explain a track’s position in the chart.
How to discover music:
Hype Machine allows users to discover music by allowing them to listen to music from different rankings.
- Latest: What’s going on right now in the blogosphere. You can also browse the recent posting.
- Popular: The most popular artists, searches and blogs on the internet right now. Twitter: Interactive music chart of songs being posted on Twitter.
- Spy: Snoop on what other Hype Machine users are listening to (Live).
- Radio Show: A monthly radio show summarizing what’s up on music blogs. Music Blog Directory: The complete list of blogs they track, organized by genre. Zeitgeist: The best 50 artists, albums and songs of the past year.
- Following friends/ Hype Machine users. There is a possibility to follow users of Hype Machine to keep up to date with which songs they favorite.
The Hype Machine is really effective in launching new artists and songs. A recent example is the popularity that Foster the People gained thanks to the website.
Interview of developer Arkadiy Kukarkin by Clement Adam
C.Dear Arkadiy Kukarkin, thank you for sparing a little bit of your time for this interview.
C.First of all, could you tell me a bit about yourself?
A.I’m 24, living in Williamsburg, Brooklyn.
C.You are now working as a developer of the Hype Machine. Have you teamed up with Anthony Volodkin since the beginning or did you join the team later?
A.I was the first serious developer hired, about 3 years after Anthony built the first prototype.
C.Are you working fulltime at Hype Machine? What’s your daily routine?
A.Since the team is very small, the schedule is somewhat relaxed, but it’s generally an average of standard 8 hours, with some days running late into the night and others being very light.
C.Could you tell me –on average- how many unique page views the site receives per day? Does it put pressure on your work?
A.We don’t specifically use pageviews as a metric, since they don’t really represent user engagement (and also because the site layout de-emphasizes individual “pages” with features like infinite scroll, ajax page loading, etc). We generally track the number of registered/active users, time on site, amount of use of different features, etc. However, we are not as metrics-driven as some other startups (we don’t run A/B split testing, etc).
C.Do you still consider Hype Machine as a startup?
A.The definition of “startup” is very fluid — we often joke about different “you are a startup if…” hypotheticals, like whether you still all eat lunch together at the same table (we do, when we’re working in the same space). I like to think of it as a new type of organization — very flat structure, no obsession with growth just for the sake of growth, fulfilling our own (and our users’) goals only, minimal restriction to creativity. You see this type of model popping up in other fields, like design, light manufacturing, the food industry, etc — though with a lot of these things have been historically run in a similar way.
C.Bloggers must delete the mp3 links from their blog sometimes, how does the Hype Machine deal with dead links?
A.We generally stream the songs from the original source via our player whenever possible, but there is also a fallback mechanism where we serve a copy from our cache server if the track was recently posted, but is no longer available. We also de-duplicate tracks (i.e. if the same song was posted by multiple blogs, they are merged together), so we often have several sources we can try. In general, with higher usage of reliable hosts like Soundcloud, this has become less of a problem. We comply with all DMCA takedown requests (treating them much more seriously than some other services), and tracks older than 3 months with no viable remote sources are no longer playable.
C.Hype Machine allows users to follow the people one is following on twitter. Are you considering allowing the possibility to follow Facebook friends?
A.This is something likely to be implemented soon, once we figure out certain intricacies — facebook relationships can be quite complex, and the mapping of following/followers is not quite as neat as with twitter.
C.The Hype Machine states that the website is far from being complete. What are the latest changes made to the website? Are there any exciting new features currently under development?
A.FF & Freshest page are the latest rollouts, though we’re in a state of continuous incremental improvement all across the board. The most recent set of changes concerns the interface, making it less obtrusive and freeing up more screen real estate, as well as simplifying the feature hierarchy. In terms of features, we have a set of tools for exploring remixes that’s pretty close to release.
C.It seems that Hype Machine is developing slowly but surely. Do you think that the Hype Machine has a chance to compete against all in one solutions such as Spotify or LastFM which are available across diffrent platforms?
A.We don’t really see these services as direct competition, because they tend to be aimed at either listening to the familiar or very passive, decontextualized discovery. I think the viable audience for Hype Machine is quite a bit larger than the current user base, though there are limits — ultimately, not everybody cares about music. Regarding platforms, we do have an iPhone app (offering a more radio-like/guided discovery experience), and several unofficial Android apps.
C.Spotify and Hype Machine are both services used to discover music. What are the main differences between the two services?
A.The discovery aspect of Spotify is largely limited to the social features and the “similar artists” section tucked away in the corner of the artist pages. The process of encountering new music is enriched (or made meaningful at all, as a stronger version of this argument) by the context — from who, when, where, with who, why. We try to capture this value and give you a framework for each little discovery — the obsessive passion of the blogger, the way the community reacts to it in the context of contemporary (and not-so-contemprary) artists, the way other users relate to it. Spotify doesn’t really focus on this type of experience, in part because their catalog is so large and atemporal. It is a great tool to sample the work of artists you discover through other channels, though.
C.Apart from your twitter account, I haven’t been able to find your personal website, how come?
A.I have pretty weakly attributed content scattered all over the place, mndhl.org is my personal tumblr, at the moment. Specific types of media live on their own services, like code on https://github.com/parkan
C.How do you think that the online (and free) distribution of music affects the traditional structure of the music industry? Are people more open to different music now that they don’t watch mtv anymore?
A.Much has been said about the evolving nature of genre and it’s hard to deny that greater connectivity (or the parallel societal change) has erased some lines. Cross-pollination and boundary interfaces between different cultural spheres have always been key factors in creativity, but it’s really hard to say whether this process is qualitatively different now without being a cultural theorist.
As for the industry, I would have said it was totally screwed, but the success of the US spotify deal has given me some hope. Though frankly, nobody knows anything.
C.Where do you see most potential for development in the (online) music industry and how do you think music will be distributed in 20 years time?
A.I’m hopeful that the next wave of startups can find a space to innovate without the constraints of the tarpit of present copyright law and labyrinthine licensing agreements (and associated profit extraction schemes by catalog owners that contribute nothing whatsover to the industry), while finally getting the artists paid. However, it would be extremely dishonest of me (or anyone else involved) to make predictions for 10 years in the future, let alone 20.