Earlier today, techy blogger Shaun Dewberry posted a paper detailing “The Truth Behind Streaming Radio in South Africa”. In his research, Dewberry gives a number of reasons why he believes that stats provided by various online SA radio stations are significantly overstated.
His argument can be summarised into the following points:
- Social media numbers don’t add up.
- Not enough internet users in SA to justify numbers.
- Mysterious and sudden growth.
- NetDynamix’s published stats
- Bandwidth stats don’t add up.
- Servers aren’t configured to handle the reported load.
It’s a bit of a pity that Dewberry led with points 1 & 2 as they are difficult to prove. Points 3 – 6, however, are unequivocal facts. All these stations experienced very rapid growth, NetDynamix does publish those stats (checkout http://126.96.36.199/ for Ballz stats), the bandwidth stats don’t add up (3.2 Gigabits per second or 11.75 terabits an hour at 51 000 users), and the NetDynamix servers show a limit of 3 000 users per server.
On point 5 (bandwidth stats), the Johannesburg Internet Exchange (JINX) currently shows a peak of 3.8 gigabits per second over the last 30 days. One could fairly safely assume that much of the streaming traffic would go via either JINX or CINX – yet it’s unlikely that over 80% of local traffic was streaming radio.
Futhermore, NetDynamix is not a large-scale operation. All commercial stations in SA use Antfarm for their streaming, given the superior infrastructure and ability to handle the load. As Shaun points out, there is no such thing as a NetDynamix CDN. It simply couldn’t handle the load they claim.
Anyway, later today Darren Scott – long time SA radio personality and co-founder of Ballz radio – phoned Shaun Dewberry live on air and proceeded to launch an ad hominem attack on Dewberry, calling in to question his technical skills and then focusing on the social media points in Dewberry’s arguments. At no point does Scott attempt to address the latter points in Dewberry’s argument.
I feel that Scott ambushed Dewberry and tried to discount his research, not by addressing the research itself but rather by attacking Dewberry – pretty bad form in my opinion and often a tactic used by someone who has no credible way to disprove the argument.
So, how do we decide whether the listenership numbers are Just Plain bullshit or whether we’re in the middle of explosive online radio growth? Well to start with, I think an independent audit would help to set the record straight. From there, online radio stations should look to create an Effective Measures or DMMA type body which establishes how sessions are logged and ensures that listenership stats aren’t provided by a single party who has something to gain from inflated numbers.* As an aside – I’m a big supporter of online radio and listen to 2OV every now and then.