Pandora and Local Radio

Harker Research posted an interesting analysis a few weeks ago, comparing the ratings of Pandora to those of the typical terrestrial radio station. Here’s the main point:

If we divide Pandora’s AAS by the US population, we’ll have their rating. The 12+ population of the US is about 256 million, so Pandora’s national rating is something like 0.1. It might be a little higher in some markets, a little lower in others, but on average Pandora has the same average listenership as a typical niche programmed AM station, ranked outside the top 20.

This ties into my earlier point about Pandora’s time spent listening. There’s a lot of tune-in to Pandora, but not a lot of stickiness.

Read the Harker Research piece here.

Pandora Listeners Do Not Seem to Listen Long

Ando Media released their January rankings of streaming audio entities last week. Pandora is on top of the heap in “session starts,” with 100 million plus. Pretty impressive, even when you consider that this is as duplicated number and not an unduplicated measure like cume.

But the really interesting bit is that Pandora listeners spend less than an hour with the service for each session, on average. In comparison, Cox Radio’s streaming listeners spend over two hours for each session, Saga and Citadel’s spend considerably more than 3 hours.

There has been a lot of talk about the stickiness of Pandora, but these numbers expose the fact that the service is cool, but boring. People are finding that the streams from broadcasters like Saga, Cox, Citadel, CBS and Clear Channel are more than twice as engaging.

As more streaming moves from the home and office to the car, it will be interesting to see how these numbers play out. For example, will Pandora listeners bring their listening to the car? The stats will show those connections as additional session starts. For listeners of terrestrial stations who use the stream when in the office and listen to the same station in the car, the behavior might well be to forgo listening on IP devices and revert to the car radio. This will be a complicated behavior to measure; PPM is probably the only tool that can do it.

The complete pdf of the Ando Media January report is available by clicking this link.

Nielsen, Arbitron and the Upcoming Battle – Part II: Engaged

Several months ago, I wrote an article about the potential for battle between Arbitron and Nielsen. It was called “Nielsen, Arbitron and the Upcoming Battle.” You can link to it here.

Since then, Arbitron announced its “ARB-TV” program and just last week the Financial Times broke a story about a new consortium of advertisers and agencies

In today’s “Taylor on Radio-Info,” Tom Taylor had a couple of paragraphs on this topic. He says, in part:

Now, #3, I’m hearing more concrete chatter about Area 51-kinds of research using an electronic detector much smaller than a PPM.

I thought to myself – “huh, top secret development? Didn’t I post a link to the actual photo of the Nielsen Go Meter?”

So, I popped back to the article and – lo and behold – the link to the picture of the “Go-Meter” was broken. A Google search for “Go Meter” or the file name for the original picture turned up nothing. Any reference to the “Go Meter” has been (it seems) removed from the Nielsen website. Interesting.

So – Tom is right. Nielsen HAS gone stealth on its PPM attack project.

Also in his article, Tom mentions the idea of installing encoding software on a cell phone that would perform the function of a device like the PPM. There are many problems with using a mobile device like a cell phone as a measurement device for audience ratings. The biggest one is the instability of the platform. A measurement device for media audience ratings needs to work like a simple appliance. Having other applications running on the platform at the same time raises the possibility of instability exponentially. Thus, the whole platform could crash and detection of media exposure would stop. The way to avoid this is to use an imbedded operating system on a single-function device, like the PPM. These devices can be rock-solid stable with very high levels of reliability. There are also behavioral reasons why the cell-phone solution doesn’t make sense, but this single technical reason is enough.

So – Arbitron and Nielsen are rattling sabers, but in a “muffled” way. It’s clear that the muffles will be taken off the sabers pretty soon. Nielsen will attack Arbitron in the major markets with portable electronic measurement. Arbitron’s move is to encroach upon Nielsen with its ARB-TV project and alignments with companies like TRA and Tivo.

Let the battles begin!

Ratings, Radio, Sales – The Debate Goes On

Chuck Francis, VP New Media at RemergeMedia.com, wrote a well-received article in Radio Business Report recently called “A New Day for Radio.”

Chuck makes the valid point that, in truth, ratings don’t matter, except to “…those that provide ratings.” However, this common-sense statement misses the mark. Yes, it’s not the ratings that matter to the listener or the advertiser. It’s the thing that the ratings represent.

For example, Chuck says:

In the years I spent as a Program Director I cannot recall one single instance where a listener came up to me at an event or a remote and said: “the reason I listen to your station is because your station the number one station.” In fact, even if your station is 13th in the market – I’d argue you’re number one in the minds of the people that listen to you regularly.

One of the reasons that a listener is loyal to a station is because of its reputation among her peers. The listener doesn’t look at the ratings and make a rational decision about whether or not to listen to a station because it’s #1 in one category or another, but she does decide whether to listen partially based on her peer group’s relationship to the station. The ratings are a surrogate for this – they tell us roughly how many and what kind of people are listening to each station. They tell us about the listeners’ peer groups.

On the advertiser side, the ratings are a surrogate for a true measure of results. In the absence of a way to directly measure the ability of a station to deliver results for a specific product, the ratings system was developed in an attempt to predict what will happen if you run a spot on a station.

Those stations who achieve success without ratings have found other ways to measure their reputation with their audience and to predict the results their advertisers will have when they run a spot on the station. However, I suspect that these substitutes for ratings remain surrogates of a direct measurement of reputation or results.

TRA, a research firm in New York, is developing a very interesting way to measure results in TV advertising… a methodology which could well be transferred to radio and other media. Instead of measuring how many people are viewing a particular program, TRA measures how many people actually were exposed to a specific ad and how many of those people actually purchased the product being advertised. This is similar in many ways to the work that was done on Arbitron and Nielsen’s suspended Project Apollo, but TRA is processing millions of households’ viewing data and matching it up against their purchasing data. As this platform matures and is able to capture not only TV viewing but also other forms of media, we will move away from the surrogacy of ratings and towards the direct measurement of the results of radio campaigns.

Of course, it is not likely that this direct measurement of results will be applicable for all advertisers, particularly local retailers. As data is developed inferences will be possible that are much more precise than today’s rating system. TRA could syndicate their data in such a way as to provide local decision-makers tools to determine which media selections are best for their businesses.

Direct response radio marketers are measuring this today, for their clients. One particularly sophisticated firm, Strategic Media, has literally written the book on the subject. Through extensive testing and results measurement, they have built very detailed databases of the stations and creative execution that works best for their clients. Perhaps local radio can learn from this and develop similar data for their advertisers to use.

Please share your views on this subject by commenting on this article.

Comparing PPM Streaming Data to Actual Server Stream Metrics

Mark Ramsey, in a recent article on Hear 2.0, posed an interesting exercise for broadcasters – calculate the difference between your stream’s cume in PPM to the records that your CDN provides for total unique connections for the same period.

He asked for responses to these case studies. So – in an effort to make this easy to do, here’s a link to a spreadsheet where you can enter the info for your stream and compare the results with others: