This article is part of my ongoing personal research about things going around the Music Scenery. Using SAS social listening/web crawling/analysis tools, I aim to detect patterns in social conversation around music, find innovations in Music technology scene and draw some implications on new music tech product launching. As far as my weekly schedule allows, this project is still under development.
I personally have been a heavy Twitter user for a year. This is an outlet to share my knowledge about Music (@phuong_vu_3011) (music metadata, artificial intelligence in music content curation, music technology, marketing in music, music mentorship, etc.) and promote my own music content. Also, based on my long-term engagement with music, I was able to articulate business decisions from my findings. Fascinated by some music hi-tech product/software, I chose #musictech to explore with NodeXL. I limited the search to 4000 tweets, and the results were 489 edges and 307 vertices.
Taking a close look at “Description”, I can tell that most of the vertices were either founder of technology companies, venture capitals, technology magazines writers/reporters, growth hackers, marketing strategist, DJ, sound designers, music producers, music agency. It drop the minimal hint that a mere music lover will not use such specific hashtag. Location wise, it is no surprise to me that the locations are tech incubators or tech hubs (Paris, Stockholm, LA, London, Florida, Colorado, SF, etc.). If I have a music tech product (an automate songwriting software for example), I can shadow a trendy specific hashtag to attract major technology websites & magazines organically.
From the overall graph shape, the centralized connections of the people tweeting about #musictech, including myself exist. In my case, the directed-undirected graph doesn’t make much difference, since the Music tech was a close-knit community, mostly the prominent touch point of interaction was B2B rather than B2C.
Overall metrics
The top 10 vertices:
In-degree | Out-degree | Betweeness centrality | Followers | Eigenvector centrality |
Soundtrap (10) | Visitswedenit (7) | Soundtrap (421.638095) | Beyonce | Soundtrap (0.103) |
Isachintiwari (9) | Peema (7) | Visitswedenit (169) | Techcrunch | Larsbergstrm (0.092) |
element14 (8) | Sthlmstartuphub (7) | Peema (169) | Newyorker | Bbcclick (0.083) |
Moodelizer (7) | Startupguidesth (7) | Sthlmstartuphub (134.067) | Billboard | Sthlmstartuphub (0.075) |
Aimporg (7) | iam_bombshell (6) | Startupguidesth (134.067) | BBCClick | Startupguidesth (0.065) |
Bbcclick (7) | rockstardreams1 (5) | Movetostockholm (102.286) | Soundcloud | Musicedexpo (0.064) |
smashdig_com (6) | Tectoniclabs (5) | Moodelizer (92.286) | Spotify | Fredposse (0.063) |
Larsbergstrm (6) | 7digitaltech (5) | navY (70) | Venturebeat | Alicekeeler (0.054) |
Siliconvikings (5) | Fredposse (5) | Isachintiwari (56) | Ableton | Amandafoxstem (0.047) |
Artistrelations (5) | Heyimjulia (5) | Siliconvikings (44.952) | SlackHQ | Siliconvikings (0.0542) |
You might notice that my top vertices are mostly from Sweden: Soundtrap is a startup music recording/tech in Sweeden, Peema is Soundtrap’s founder, Sthlmstartuphub is the startup hub of Stockholm, same fashion for Startupguidesth which is the information platform for the startups based in Sweden. It is surprising that #musictech isn’t used by incumbent technology magazines/music platforms (Techcrunch, Soundcloud, Spotify) with a huge number of followers, but the hashtag got spread out through a niche twitter account of startup companies in Sweden. Soundtrap is almost the top of every criterion, with only 2770 followers. As a result, if I look to spread my message about my product, without having to ‘follow’ too many people, I can start connecting with key people that Soundtrap follows. This number is much more manageable and of higher quality. In this case, closeness centrality doesn’t matter much, since the top 10 lowest closeness centrality is all 0. The cutoff point is at vertex 47, of 0.010 of ‘auddlyofficial’- a music copyright collecting in Stockholm (again!). Overall, my top vertices are quite consistent.
Let’s take a look at the subgraph. In a nutshell, it has 67 groups, with vertex shape: Disk, Solid Square, Solid Diamond, Solid Triangle, Sphere, Circle.
Taking a close look, I notice that all top 10 players are strongly linked to each other. These players deemed to be the influencers, so the people I would want to mention or tag in my note to spread the words about my product. Playing around with the autofill, I learn that the top 10 stay consistently dominant. When I set vertex size to Followers and set the size from 1.5 to 4.0, the shape of the graph doesn’t vary much. In the middle were:
- Guide_humming: 59 followers
- cantibene0709: 50 followers
- It reinforced my previous assumption: To get my message across, I didn’t need to reach to BBC Click or Spotify.
With Dynamic Filter, I recognize that it is not a heavily centralized network. Some node may have huge followers (Huawei, Beyonce, Venture beat, etc.) (strong network value), but would not be good prospect for outreach effort(not so high customer) for #musictech
Relationship wise, out of 487 edges, I observe 325 mentions, 11 replies to, 151 tweets. A reasonable approach is to mention other influencers or retweet them to expand the spread of #musictech. I experimented when mentioning @soundtrap in one of my tweet. It immediately initiated engagement reaction from their side. Examining 362 Edges which have the domain in tweet, I find some patterns:
Some pointers:
- Only one domain is Facebook.com. Therefore, I wouldn’t link any of my music content to Facebook, because Twitter users would not click on it.
- Apart from some popular domains like Twitter, Youtube, I also find that Hypebot.com (a well-established and worldwide music magazine) appears lots of time, so did musictech.net, aimp.org (Association of Independent Music Publisher), smashdig.com (again, from Sweden!), musicbusinessworldwide.com. Surprisingly, widely respected tech voices such as techinasia.com, venturebeat.com didn’t generate lots of traction, and neither did wired.com (almost one of the most popular music website), musicrow.com (more on country music).
- Famous music tech startup like soundcloud.com or deezer.com doesn’t make it to the list.
- Apple music is referred at super low frequency. My assumption is that Apple Music was already launched in June 2015. Apple may have won more inbound links rather than self promotion. Indeed, Apple generated more outside advocates in pages like thenextweb.com, theverge.com, etc. Some related hashtag could be: #Musichistory, #musicgear, #musicnews, #startup, #techjobs, #techtrends.
- www.artiphon.com (an all-in-one instruments: violin, guitar, keyboard, drum) – one of the most successful crowdfunding campaign on Kickstarter and the most talked about innovation in media) isn’t on my list. The fact that Artiphon kept delaying their product launch date may be a possible explanation. Plus, I only limited the number of tweet search to 4000.
Although there was a tendency in the vertices, #musictech wasn’t wholly centralized to detect the Acquisition cost of customers.
Credit: Featured image: subfold.com