Chris Grant, CEO of NetDynamix, has had to deal with a massive media storm over the past couple of weeks. A blog by Interweb’s Shaun Dewberry called into question the listenership numbers of the burgeoning internet radio sector. NetDynamix is the company that provides “services and solutions to the radio industry”, including, in this case, audio streaming. The story went viral, and was taken up by not just the IT press, but the media in general.
Dewberry said in his “ professional (and personal) opinion, I state that there is only one explanation for these proclaimed levels of listenership – they are complete fabrications. Utter nonsense. Lies, even.” And Dewberry went on to analyse the numbers to back up his claims.
There was massive fallout for NetDynamix and its young CEO as story unfolded. In an interview with TheMediaOnline, he explains how it all happened, and the lessons he has had to learn.
Your first response seemed to be to issue a lawyer’s letter to Shaun Dewberry. In retrospect, was that the right move?
It was the view of our management team and our legal represenatives that the words that Mr Dewberry had chosen were particularly slanderous and we believed that the lawyer’s letter was a good step to take. In hindsight however, I believe that the technical community as a whole saw it as a ‘scare tactic’ more than what it was, that of being a response to his slanderous words.
Once you take a step back from the entire situation, however, and you view all of the cards on the table, then I think that there was more than simply our company’s reputation to take into account and thus for public interest, the sake of our clients and their brands, a lawsuit would simply have been a waste of time. We stick by the merits of the legal advice that we got, and we do not believe that his article was written with enough consideration, but after looking at the entire process unfold, we do not believe that the lawyers’ letter was a good move which is why we haven’t pursued it further.
Was his story accurate?
Without trying to sound too complicated in my response, I believe that Mr Dewberry’s article had certain points that were correct and some that were only a partially correct, which I say on the basis that he simply did not have enough information at his disposal when writing the article.
For example, the ‘Shoutcast Status Page’ that he was collecting his information from does show an exact number of users who are connected to the service at that particular time. He then took it a step further and monitored the status page every five minutes (I believe) in order to collect data and compile it into graphs and reports.
What he did not have access to using this very simplistic method, is information such as: IP Addresses of the users; listening time per session; and sessions tally (sessions used very loosely).
Why is it important to have this type of information available? The IP Address of the user can tell you things such as geographical location of the user. ISP/Network information of that IP and the route in which it takes to receive the audio from our Shoutcast Cluster.
Using IP Address information you can also, as best as possible, try and determine the ‘uniqueness’ of a user and his ‘return rate’ to the service. Listening time is also very important. One of the terms that we have come to learn through this entire process is that with internet radio audiences there is a vital need to ‘discount’ certain sessions from the reports as ‘unqualified’ – these, as we understand now, are sessions that last less than 60 seconds.
Session tally, as after discounting unqualifiying sessions from the mixture, you get a better understanding, again, of the ‘uniqueness’ of the users over a full reporting period (7 days or 30 days) rather than a ‘concurrent listener view’ only.
An example of sessions is as such: USER 1 and USER 2 each use the same internet connection to get out onto the internet. This single internet connection has a single unique IP address. USER 1 connects to the radio stream at 09:30 and listens all the way through to 13:30 USER 2 connects to the radio stream at 11:00 and listens all the way through until 14:00.
Now typically, and in ideal conditions, USER 1 and USER 2 have created two unique sessions or connection to the radio service through a single IP address. On the certain interrogations of the log file for that particular radio station you would see one single IP address. Others will show you two sessions created. You need full and unhindered access to the log files from our systems in order to get clearer and more reportable information before you state something as fact.
As I had stated above, in ideal network conditions we hope that a single user creates a single session throughout his listening time on the radio service. In practice however, it is quite possible on certain platforms and devices that multiple sessions could be created for a single listening time. This produces the need to have full access to the RAW log files of all of the servers in order to produce correct numbers.
I do not wish to rubbish what Mr Dewberry had said, because I think that at the time that he went about collecting his ‘statistical data’ – he had done a fine job. He used his skills along with some solid industry accepted software to obtain a certain view of the traffic. He achieved what he had set out to achieve. I do however believe that should he have had access to the RAW log files, it might have changed the view he was presenting and would have enabled him to give a seven day report or 30 day report on a usable period rather than just a 5 minutely concurrent sessions report.
Lastly on this question, without access to the previously mentioned RAW data. There is absolutely no way possible that Mr Dewberry could provide the public, our stations or ourselves with the total pool of listeners that had tuned into those radio services for the month. In his view there were at best a total of 300 people listenening to one of our clients in a given month. This is wrong. I think that his method of collecting the data and presenting it, at best, comes close to what the radio industry refers to as ‘Average Quarter Hour’ or AQH. Not exactly, but close enough!
Mr Dewberry cannot tell us how many unique people had tuned into those stations in a seven day period, or a 30 day period or a 60 day period. All he could see and collect was “partial” information and I do not blame him for that.
Can you unpack how he came to the conclusions he did?
Mr Dewberry, from what we can see, has made use of a piece of “monitoring” software that would draw certain data out of the Shoutcast Status Page. He would then allow this software to run at a pre-determined interval and to collect and record said data into a database. From there he produced a graph showing “at that time” number of connections to that particular server. He then reported exactly what he saw and what his monitoring software was telling him. He made no error in his collection of the data but rather a mistake in reporting something as “absolute” when he did not have full access to the RAW log files.
Has the legal threat passed or is it still in play?
After much pondering on this particular point we decided that it would not be beneficial to either party to continue. From a NetDynamix point of view it would do nothing to restore reputation. Our client portfolio has been more than accepting through the entire process as we have and continue to be transparent to them. Many of them have also been with us for more than a few years and they know how we operate and our integrity.
For Mr Dewberry, it would have been unnecessary for us to go after him legally as, at the end of the day, he has highlighted a very important topic. One that we’ve voiced our opinion on to our clients since October 2011. There is the lack of two things in the online radio industry. Firstly there is no clear and defined standard that the bulk of the radio stations are using to report online listenership and secondly, perhaps most importantly, there is no independent body to take hold of the verification and publication of these figures.
For the technical community I believe, that because of the legal threat, had in their minds painted a very poor picture of NetDynamix and should we have continued to give them the paint brushes and the paint to go on, the portrait would have looked very nasty.
Is it in play? No. We see no point in continuing with the legal action at this point as we are only interested in presenting the facts and putting new systems in place, with the consultation of our clients, in order to avoid such events in the future.
How are you managing the fallout with your clients such as Ballz and 2Oceansvibe?
Ballz and 2oceansVibe Radio have already publicly stated that they are looking for alternative service providers. Since we enjoy a very close relationship with most of our remaining clients we have been on the phone to them and on email to them constantly over the past three weeks not only to manage fallout, but mostly to get their input as we implement new statistics systems. We believe that it is more their choice than ours and we need to cater for their choices.
How have they reacted subsequent to realising there WAS an issue with the numbers?
At the end of the day we were called upon to report session number which is what we had done. There was no fabrication of these numbers, these are clearly visible on various reports that we have. At the time of generating those reports we did raise a number of concerns around the obscurity of the figures to our clients that we were seeing and although we had incorrectly labeled some columns in our reports as LISTENERS rather than SESSIONS, something which we have apologised for and do not hide from anybody, we now understand that because there are so many factors that need to be applied to discount “unqualified sessions” to get to a more realistic number, our sessions-only reports should have accompanied other data as well. As for their specific reactions, I believe that their reactions are quite well documented on their websites.
What, in your view, is the solution? (Matt Buckland’s piece on Memeburn suggested similar incidents took place in the early days of online publishing.)
An independent organisation, body or research firm needs to be setup and made responsible for the development of a methodology/standard and for the reporting of online radio listenership figures to the public. This organisation should not be controlled or run by the radio station, it’s employees, service providers or anybody else linked to the station.
We believe that in the same way as SAARF reports RAMS, AMPS and TAMS… this organisation should be run.
Matt Buckland’s piece on Memeburn highlighted “what it all boils down to” in my opinion – as a technology service provider, who are providing services to our clients to get their content to their users, we do not see how we could extend our area of expertise and training to now also include statistical research. Also based on the simple fact that we are not independent in the reporting of the findings either. Our Terms and Conditions that cover our audio streaming service clearly state that the data collected in log files and such mechanisms remain the ultimate property of our clients. This means that we do not publish their RAW log files or generated reports of their listenership or online activity. That is up to them, or as we are proposing, an independent and qualified third party.
What, exactly, does NetDynamix do?
Our core focus is providing services and solutions to the radio industry. These are typically audio streaming and outside broadcast solutions but we also provide services such as SLA Contracts for maintenance of equipment; studio builds; studio upgrades and custom software development.
Looking at audio streaming alone, we make it possible for our client’s content to get from their studios to the ears of their users (listeners).
What is the difference between audio streaming and traditional FM radio?
FM Radio receives it’s signal (or feed) via terrestrial transmitters and is governed by various organisations. ICASA is responsible for the granting of the FM frequency and license, broadcasters are, mostly, part of the NAB and are governed by a code of conduct enforced by the Broadcasting Complaints Commission. Their footprints are also “geographically” limited due to the nature of FM. National stations overcome this with the use of multiple frequencies and a network of FM transmitters, but ultimately they are limited in the reach that they can achieve. FM stations rely on SAARF to produce their listenership figures.
Audio streaming is received via an internet connection. There are certain additional “usage” costs involved, which are not charged by the stations or by us, but rather by the ISPs or methods of access of the internet user who tunes into the stream. There are very few regulations surrounding internet radio in South Africa besides that of music and content rights and to date there is no single body or association that all of the internet radio stations are a part of. In the same breath there is no independent body producing their listenership figures.
Who, if anyone, actually publishes listernship figures in this sector?
For FM Broadcasters, an organisation known as SAARF (The South African Advertising Research Foundation) produces a quarterly study known as RAMS (Radio Audience Measurement)
For online broadcasters…nobody. The station’s streaming provider can produce reports from their RAW log files, but there is no body that is applying any methodologies or research methods to these figures and ultimately nobody is publishing these figures.
Does NetDynamix actually track sessions yourself or do you use other technology?
At the core of each of the technologies that we use to deliver our services to our clients is a logging function. This function records in RAW unedited format, all of the information for those sessions.
We now make use of two pieces of industry accepted of software that go through the log files and generate usable data. We then provide the RAW log file (unedited) along with two generated reports to our clients for them to interpret.
A new step that we’re busy implementing is to make those two pieces of software available to our clients along with direct access to our servers in order to draw those log files into the software and to generate their own reports. This adds a new layer of transparency into the entire process.
Is there a system by which all streaming radio can be reported, some kind of ‘baseline’?
If we look overseas, there are some standards that we are busy looking into. These standards, or baselines, are being shown to our clients so that they can make the choice on which one we should use and make available to them.
If not, are there moves to standardise reporting to prevent this from happening again?
We still believe that an independent body needs to take on the responsibility of drawing, generating and reporting listenership figures to the public and to potential advertisers. If an independent body is setup to create a methodology around the collection, reporting and publishing of these figures and they create the “BASELINE” or “STANDARD” then us as a streaming service provider will provide them with the RAW data that we have in the formats that they require in order to aid them in what they are tasked to do.
This crisis must have presented a few lessons for you, not least how social media can impact HUGELY on a story. What are the lessons you’ve learnt?
Broadly we’ve learned that in our own company we have many obstacles to overcome. These are mostly in the field of communication. As a company we do not have a mouthpiece from which to speak through or from, so our responses were not always well published. We also realise that we need to engage more in the media, in public and on social media.
We’ve also learnt that in a time of crisis, as you’ve put it, your integrity and ethics to date will dictate the outcome of whether you are here or not tomorrow. 90%+ of our clients know us. They know how we operate and they know that we would never skew or fabricate data or figures which is why we are still here today! Which is why we’ll be here tomorrow, because the vast majority of our clients are still with us. Yes we need to overhaul our entire reporting system, but this is an easy step to take in order to rectify all of the problems that have been caused.
What would you do differently?
Firstly, we made a mistake in some of our reports to our clients where we had labeled “sessions” as “listeners” which was a design flaw in our collection and production of that report. If it is something that I could fix that it not have been made, I would certainly take that back and do it correctly because then there would be no misunderstanding at the end of the day.
We would have taken our suggestion of our web only radio stations forming an association to a wider client base rather than the five or six that we had consulted with since October last year and perhaps been more of a driving force for them to get this organisation setup.
Put simply however, we have made a few mistakes and we’ve apologised for them. I do not believe that there is any point in wishing that you had not done X or not done Y. I believe that in life and in business, if we did not make mistakes then there would be no basis for the company or our employees to grow from. We would be stagnant and perhaps be too focused on the wrong things. Mistakes are there to be made and no human being (or the company that they represent) is perfect…it is how you bounce back from your mistakes that I think is important.
We will continue to provide our clients with a top notch content delivery service and we will strive at all times to better that service where possible by putting additional methods and systems in place that can better serve them. Right now… that service is better statistical guideline generating software.