Kelly: Pandora and Radio Are Not Equals

5

In part two of our three-part preview of Brad Kelly’s Radio Ink cover story interview, we cover the topic of online ratings. The radio industry has been anxiously waiting for when it can show advertisers a combined listening total of listeners that includes both over-the-air listening and online listening. Nielsen has been working on the service for years now and has hit the “pause” button several times. There has also been a battle over whether pure-play streaming companies would be included in the service. In our cover story interview, Nielsen Audio Managing Director Brad Kelly gets us completely up to date, including whether your advertisers will see the word Pandora next to your radio station in the ratings.

Radio Ink: Give us an update on online ratings, and tell us why it isn’t being used yet.
Kelly: Let me first offer up some background. SDK, which stands for “software development kit,” is a piece of code that rides on top of apps and players. When a user fires up an app to stream digital audio, what’s happening is that SDK pings back to Nielsen — it fires and says a session has started. We capture that session and the station being streamed, and the duration and amount of time they’re spending with that station. We then use a variety of third parties to put demographics against that individual. We then calibrate the data to PPM data, and then we have census-level information. This is not a sample or panel, this is census-level information for everyone who is streaming for all players and apps that have the SDK enabled. We have been working very hard for a couple of years on this. We have 7,500 SDKs installed, which represents about 2,500 radio stations and radio streams. We hit the pause button on this one back in late 2016. Part of the reason is our customers said, “We’re laying data up against the other service that’s out there, and there is a differential delta.” They were saying that the data from Nielsen from the SDK is, in some cases, lower than the other service, and want explanations as to why. So we took a step back and started looking at alternate sources to measure streaming, other methodologies, other companies to partner with. What we’ve come back with is that the SDK is the right solution. We are further emboldened by the fact that this same technology is in use right now and has been MRC-accredited on the TV side of the business.

There are some reasons for the delta. We’ve employed restrictive crediting rules, and we started asking ourselves if we are being too restrictive with these crediting rules. If the other service out there is MRC-accredited and their crediting rules are looser, then maybe these are areas we should be looking at. That’s what we’re doing now. We have hit the un-pause button on SDK, so we’re committed to that as the data-collection mechanism for streaming. The vision is a bit different from what we had originally envisioned. This is not what we had originally suggested would be a syndicated service, meaning full-market coverage, just like we do in our standard local markets. This is not something we can force on the industry. It has become an opt-in service. There is engineering work that needs to be done to get the SDK embedded to allow the data to flow through. So for broadcasters that wish to participate and opt in, we can and will produce digital audio numbers for them.

Radio Ink: How will the numbers look on a sheet of paper so salespeople can present them to an advertiser?
Kelly: With regard to an integrated number, for folks that opt in and take the SDK, there are two different ad load models. If you simulcast 100 percent of your commercials, your stream is a mirror image of your over-the-air. That situation qualifies for what we call total line reporting. In that case we will take the stream and combine it with your over-the-air numbers, so you can present it as a single number. For stations that elect the ad-insertion model, we’ll produce two numbers, a number for over the air and, below that, on a TapScan report, a number for your stream. Then we will further take the step of showing the total combined audience for the two.

That’s slightly different. Total line reporting is a single-line-reported number that flows through the buying system. Total audience comes with a footnote and caveat that these numbers, although we’re showing them independently, are additive. The only way to reach this combined total is to have a mirrored ad load on both of these. For folks that opt out, they simply would not have a streaming number associated with the station.

Radio Ink: So the reason there are two options is that you really can’t present an advertiser a total line number if a station has different ads playing, since their ad might not be played on one of the listening options?
Kelly: Yes, that is it. It’s truth in labeling. They break for commercials and there’s the potential for having different loads on these different platforms, so they should be reported independently. We wanted to give the broadcasters the ability to show the combined number: Here’s the potential delivery between the two if we run your spots on both of these platforms, but that is not a guarantee in a non-simulcast situation.

Radio Ink: When will these numbers finally be in use?
Kelly: We’re generating some preview data for stations jumping back into the pool with us again. We want to make sure everybody who wants to participate has got the engineering tightened up and is feeling good about it. We will produce preview data this summer. I don’t want to commit to the definite “go live” launch button, but we’re looking at an October time frame, end of the year. It is subject to the marketplace and the adoption rate; we will move at the speed at which the industry wants to move.

Radio Ink: Has there been a decision on whether companies like Pandora and Spotify will be included?
Kelly: We would be happy to include them, and I think they would tell you that they would be thrilled to be included. The tricky part is their desire to show up in the New York or Los Angeles local market report with an average quarter-hour stacked up against Power or Hot 97. We believe at Nielsen that it’s fundamentally different; pureplay digital streams are fundamentally different and require different metrics associated with them. Average quarter-hour is uniquely broadcast. Digital is about impressions, and because you can mathematically back into an “average quarter-hour” number for these digital pureplays does not mean it’s the right way to evaluate them. We would be happy to produce the number using the SDK on the digital side with impressions. I think it’s their desire to be measured more akin to broadcast, which is not a conversation we are entertaining currently.

More from our interview with Brad Kelly in our Friday headlines. Read Part One (What is the Perfect PPM Size HERE).

To subscribe to Radio Ink magazine and have this interview and the entire May 22 issue downloaded to your digital device today GO HERE

5 COMMENTS

  1. As the former employee of the “other service” that conceived, developed and accredited that offering, I find a bit of Mr Kelly’s descriptions and explanations a little hard to swallow. Let’s call the SDK what it really is, and that’s a sample or a panel. Not that having another sample or panel is a bad thing, but in my opinion, trying to disguise it as “census-level” information is misleading. When “7,500 SDKs” represents “about 2,500” stations and streams, and needs to be calibrated to PPM, it’s a sample or a panel.

    Mr Kelly then goes on to talk about crediting rules as the possible reason for the “differential delta” of the two services. AQH Persons is defined as the average number of persons that listen to a particular station for five minutes within a quarter hour. Seems pretty straightforward, and I’m positive the MRC makes sure that there is no loose-ness in the crediting rule.

    Maybe what’s really being described is the reluctance to simply measure the audience and present it as a whole (broadcast and digital combined), unless you adhere to Mr Kelly’s simulcast rule?

    In the simplest terms, measurement services should be trying to provide the most accurate sizing of an audience. Seems sensible…

    My two cents.

  2. From this article, I’m still not real clear on why pureplay digital stations are or should be viewed differently from broadcast streams when it comes to audience measurement. I run an online indie rock station in the Dayton, Ohio market and we fight for online listenership along with our broadcast friends in the market. If “Bob” is sitting in his cubicle and chooses to listen to MyGreeneRadio online or on an app over some other Dayton station online, shouldn’t the ratings reflect that? Perhaps I’m missing the rationale but this seems like an attempt to deny the veracity of a potential competitor.

  3. The REASON we bought into the PPM was the PROMISE that all audio would be measured live and put into a single report ; TV, Radio, digital, loud speakers! All of it. That’s why I lobbied hard for it on these pages and in many industry forums. At the moment I feel scammed.

  4. Radio -and online services should be a little upset (or a lot upset) at the hair-splitting going on. If WXXX has 14000 listeners through broadcast at a given moment…and 2000 online listeners, their cumulative audience should be listed as 16000 right? (That is given the other criteria…5 minutes etc.). A pair of ears is a pair of ears whether online or through AM/FM. My sales people should be able to show a prospective advertiser how he/she can reach all 16000 people (by a commercial “simulcast”) or splitting their buy by choosing either broadcast or online. If my station stops for a 7 minute commercial break, whether online or on AM/FM, it’s going to have an impact on my total audience. The only time the content of that break is relevant is when discussing which platform delivers the most efficient audience…or if the buyer should combine the two. I think the big issue is the AM/FM “estimates” are going to be much much higher than the online data. No one seems to want to address that. WLTW has a cume estimate of over 6 million. Does Pandora have that kind of number in the New York Metro? Anyone?

    • 7 minutes of commercials???
      Seems like you have a bigger problem than your streaming numbers?
      Who gets the last spot in that break??
      Ugh.

LEAVE A REPLY

Please enter your comment!
Please enter your name here