The saga started on Monday. On that day, a Centurion-based Unix systems administrator called Shaun Dewberry published a report on his blog which contained some shocking claims. “The Truth Behind Streaming Internet Radio In South Africa”, as he titled it, alleged that the audience numbers being claimed by online radio stations in South Africa were wildly inflated. 

Within only two months of launching, he noted, Darren Scott’s station Ballz Visual Radio was attracting an alleged 51,000 listeners in its peak period. The longer-established 2Oceansvibe Radio was claiming as many as 60,000 listeners an hour, and even a community radio station like PE’s Kingfisher FM boasted monthly listeners of 20,000.

“What is the reason behind this amazing growth in the sector?” wrote Dewberry. “The falling cost in broadband? Increase in mobile access to Internet? A general dissatisfaction with terrestrial radio?” None of the above, he concluded. “There is only one explanation for these proclaimed levels of listenership – they are complete fabrications. Utter nonsense. Lies, even.”  

To understand why Dewberry felt confident making such a strong assertion, we need to go back a couple of months. Dewberry is himself involved in internet radio as a sideline, albeit on a tiny scale – he is the “tech guy” for an embryonic station called Interwebs Radio. 

“Wednesday was an awesome day – we got up to 17 listeners,” he tells Daily Maverick. When he became involved, he began to do some research in order to suss out the South African market. “I saw these incredible stats starting to pop up,” he says. “So because I’m a techie, I thought, let’s go listen to these other streams and see how they do it.”

Dewberry discovered that when he visited the IP addresses and port numbers provided for the servers via the software they used called ‘Shoutcast’, it displayed a listener count for each station – one that was dramatically different from the numbers being claimed. He was finding listener peaks for each station within three digits: one midday screenshot he captured of listener numbers showed that Ballz Visual Radio had a listener peak of 831, and 2Oceansvibe Radio had a listener peak of 122.

This was a bizarre discrepancy, but there were also other things that didn’t add up. Dewberry searched the Shoutcast directory, to which every streaming radio station in the world that utilises the software has to post their listenership. He found that the largest had a maximum of 12,000 listeners. “If the biggest stations in the world only have 12,000 listeners, how can two start-ups (2Oceansvibe Radio and Ballz) in bandwidth-starved Africa have tens of thousands more listeners than that?” he wrote in his report.

Initially, Dewberry thought the radio stations themselves might be falsifying figures, but the common denominator among them was that they all shared a streaming hosting provider: a company called NetDynamix. Dewberry cites the PE community station, Kingfisher, as particularly revealing in this regard: between December 2011 and March 2012 the station’s listenership grew from 1,000 listeners to 20,000 – at the same time, Dewberry says, that they began using NetDynamix’s services.

After what Dewberry says amounted to two or three months’ research, he decided to go public with his findings. When he posted his report on his blog on Monday at 13:00, it began to circulate fast on social media. There was, however, no immediate response from NetDynamix. 

On Tuesday, Dewberry received an ambush phone call from Ballz Radio founder Darren Scott, live on air, purporting to question him about the report. As anyone who has seen the video can attest, however, Scott’s method of “questioning” reads more like an attempt at intimidation. 

Scott focused mainly on one aspect of Dewberry’s report – his claim that the stations’ relatively small following on social media does not parallel the large listener numbers – and rubbished this. Dewberry concedes that this is one of the report’s weaker points, but says that it was in any case only one small aspect of it. “I don’t think Darren Scott had read the full report properly,” Dewberry says now. 

At 14:30pm on Tuesday – a full 25 hours after the report had been published – NetDynamix broke their silence to bring out the big guns, in the form of legal threats. Dewberry was told that he had six hours to take down the blog and issue a full apology. “I didn’t think that was a very good way to respond, because you’d assume they would be able to pull out their own stats to rebut (mine),” Dewberry says. “When they came back with a legal threat as opposed to a technical argument, I figured, they’re scrambling.”

On Wednesday, NetDynamix put out a proper statement on the matter. “We do not agree with the basis and method that Mr Shaun Dewberry has used to arrive at his findings simply because he does not have full and unhindered access to our entire network. He does not have an understanding of the methods that we employ to manage our network and its resources,” they wrote. 

What exactly these methods were, however, they were unwilling to divulge. “The technology and means used to distribute our clients’ radio stations is not something that we are willing to make public as this enables us to deliver a cost-effective solution to our clients where others find it difficult.”

Their statement did include the intriguing fact that NetDynamix does not actually measure individual listeners, but rather listening “sessions” – that is, a connection to the server in order to receive the stream. As a result, if one listener’s stream is interrupted three or four times during the period she is listening, each time she re-connects she would be counted as a new listener. This is a possible explanation for some of the discrepancy between numbers, but it’s not the full picture. 

It would not explain why one of the major claims made by the stations (on the basis of figures supplied by NetDynamix) is how long the average listening time is – 2Oceansvibe Radio, for instance, says on its website that they have an average listening time of “a whopping 2 hours 28 minutes”. It’s unclear how this would square with the fact that NetDynamix is counting each re-connection as a new listener.

The main thrust of NetDynamix’s counter-argument, however, was that they use multiple platforms – and not just Shoutcast, which Dewberry monitored – to stream their audio. NetDynamix spokesman Hans Stricker told tech website MyBroadband.co.za that Shoutcast was mainly used to stream audio to mobile devices. 

In an interview on 2Oceansvibe Radio on Wednesday, Stricker repeated this claim: “He’s only looking at a very, very small portion of the network we use to stream our radio stations,” he said. “The Showcast platform…constitutes under 10% of our total audience.”

This is a point that Dewberry refutes. “At no point during my investigation did I encounter any other server,” he says. Other tech-heads appear to have carried out similar tests and support Dewberry on this, judging by comments and forum discussions on MyBroadband.

On Thursday the plot thickened, as MyBroadband reported that it appeared NetDynamix had suddenly changed from using Windows Media players, which used Shoutcast, to Flash players, which use a Flash Media Server. “It is currently not clear why the old Flash media player, which seemed to work well, was replaced with an updated version which seems to point to a new server,” the website noted. 

A Flash Media Server would apparently make it impossible for outsiders to monitor listening figures. NetDynamix gave a technical explanation in response, including the suggestion that “the script that runs and checks user agents is constantly being modified by our developer as and when bugs are found”. 

NetDynamix has now said they will be asking an independent agent to conduct a thorough audit of their listening figures, though exactly how this will happen is not yet clear. Nobody was answering phones or emails at NetDynamix on Thursday afternoon, so we’re unable to illuminate you further in that regard.

Dewberry says he hasn’t heard anything further from their lawyers, despite his refusal to take down his report. “If they had technical evidence that I was wrong, you’d assume they’d be pursuing it by now,” he says. But he’s confident that his findings are correct. “I’ve staked my entire reputation against everything I’ve said,” he says. 

“When the report started gaining traction on MyBroadband I was sweating a bit, because those guys are techies, and I knew that if there was anywhere I would get nailed, it would be there. But when the MyBroadband guys agreed with my conclusions, I said: I’m in the clear here.”

The Daily Maverick turned to Julian Mountain, Senior Digital Strategist at Prezence Digital, for an independent opinion on the matter. Mountain said he found Dewberry’s report “credible, well argued, and alarming, given that I wear a digital planner’s hat.” 

Does Mountain think that advertisers who may have been lured in through inflated listener figures will be reticent about choosing to advertise on streaming radio in the near future? He thinks not. “Internet radio is phenomenal for advertisers,” he says. 

“You can track location a lot closer than terrestrial radio, and you gain a lot more insight into who your listeners actually are. I don’t think advertisers will hesitate to keep using internet radio stations, but perhaps the budget allocation will drop a little.”

2Oceansvibe Radio, who have declared themselves horrified by the idea that NetDynamix might be inflating figures, told the Daily Maverick that they hadn’t experienced any advertisers pulling business as a result of the report. “Initially there were a few justifiable concerns, but once those clients were given a value breakdown of our cross-platform offerings, any jitters were quickly put to rest,” 2Oceansvibe founder Seth Rotherham said. 

Would the station continue to use the services of NetDynamix, in light of what has happened this week? “The week’s events definitely do play a role in our decisions – we are exploring other options,” said Rotherham, which may be a diplomatic way of saying that NetDynamix is in for the boot. The station seemed to be putting a brave face on things, however, keen to focus on the fact that they have just signed an agreement to partner with News24. This will see the radio stream available on the homepage of News24 – a move likely to increase their listener figures in any case.  

One question yet to be satisfactorily answered is: what would the motivations of NetDynamix be in inflating listener numbers? As they pointed out in a press release, they are not paid by radio stations per listener. Dewberry thinks it’s straightforward: “Streaming radio is their industry,” he says. “It’s to their benefit to make it seem like a major growth area”.

Until NetDynamix releases a full set of audited listener statistics, we won’t know exactly what’s going on. Their evasive behaviour to date, however, strongly suggests that there’s something fishy in the state of internet radio. DM

Read more:

• Bizarre twist in online radio stats debacle, on MyBroadband
• NetDynamix hits back in online radio debacle
• Shaun Dewberry’s blog

Photo: 2Oceansvibe Radio website