I attended a Facebook Developer Garage event in SF a month ago. Andreas Weigend, former Chief Scientist of Amazon.com, articulates a PHAME methodology for optimizing results from social data:
Problem -> Hypothesis -> Action -> Metrics -> Experiments
This would be an excellent model to apply to:
- Social Media Marketing of bands/artists
- Application Development (Music player apps for example)
- Social Media Marketing in general.
So for example, for promoting an artist or a band, one might utilize the PHAME model to engage in a number of marketing activities. For example, one “P” problem might be ‘how to bring new fans to our myspace site and get more plays’. An “H” Hypothesis might be ‘Using a Myspace friend add automation tool could be an effective and not too costly way to get new fans’. “A” Action might be “Google search on ‘Myspace friend automation’ and select one of these tools to purchase and use (sparingly … so as to not overdo it with ‘friend adds per day’). “M” Metrics – track friends requests resulting in acceptance vs total requests, track number of song plays before and after. Track increase in song purchases (if applicable). “E” Experiment: What are the results like? Any new ideas surface from results so far. Can we add more fan requests per day without getting banned? Should we drop back on friend requests per day? Checking out some of the new fans playlists, is there a similarity in their musical taste to our songs? Are there features or adjustments in the automation tool we have not fully explored that might make a difference?
The point is that there is no ‘set in stone rule’ that is going to magically work for everyone. In fact, some musicians might be turned off by the idea of using an automation tool as just ‘being another spammer’. Or maybe your band gigs and tours enough that you are doing fine with building a fan base by more traditional means. You may have your own PHAME experiments to do (YouTube Video? Last.FM ? Use crowd source funding for next recording session using http://www.sellaband.com/ ?)
What should you try next? There may be many ideas. Start with one. Use the PHAME model. Measure results and tweak the experiment if needed. Have some fun doing it.






