|
Welcome!
Today's
News
Yesterday
Kurt's
essay
Guest
essay
Site
of the Week
Feedback
form
News
archives
Vendor
guide
Readers'
forum
Internet
101
Internet
201
Who's
Who
Interesting
sites
Fave
bookmarks
Coherent
Design
Chat
room
|
 |
 |
|
 |
| Friday's
column analyzing the first Arbitron InfoStream results
(click screenshot at right to read
the original article) drew lots of
reader comments. Some of the best of them are presented below.
To contribute your thoughts, please click here. |
 |
RealPlayer
presets explain some of the webcast cume numbers
From:
William Goldsmith
(wildbill@kpig.com) |
December 10, 1999
11:54:26
|
As the KPIG "web guy" since the beginning (and as a present consultant
for both KPIG/CD93 & Magnitude Network) I read your comments with
great interest. I have access to the raw data that Arbitron used to
base their report on, and I can draw a couple of conclusions.
1. Arbitron does not seem to have accurately processed
this data.
2. Their inaccuracies don't material affect the conclusions
you drew, nor the overall picture painted by the report, which I agree
is less rosy than it might appear at first glance.
The most glaring inaccuracy is the fact that KFAN is a solid #3 (not
#1) in total accesses. Both CD93 and KPIG ranked significantly above
KFAN in terms of avg. daily accesses. Since, as you pointed out, these
numbers are not estimates like Arbitron's radio surveys, I'm
at a loss as to explain how they came up with the ranking that they
did.
I agree that some equivalent of AQH listening is the most critical
measure of the penetration of webcasts. The estimate that you came
up with isn't so far out of whack -- so perhaps Arbitron elected to
not include such a figure because it would be so low.
As nearly as I can determine, AQH listening for October among Magnitude
stations had KPIG, CD93, WEQX, KNSX, and WKPO at the top (with numbers
not too different from what you estimated). KFAN didn't even make
the top ten -- their AQH was significatly lower than your estimate,
due to their extremely low average listening times ("listening times
per access").
Why are the "listening times per access" numbers (which I quoted you
but regretfully can't let you publish at this point) for KFAN (and
KPIG too, IMHO) so low? It's because of their placement on the RealPlayer
presets. A great number of the listeners for certain Magnitude stations
come from the RealPlayer presets. WEQX, for example, owes about 85%
of its listening to the preset. But they're an alternative station
that's listed under "Alternative" -- so they satisfy the expectations
of those who click on them pretty well. So does CD93, listed under
"Pop."
KFAN & KHYI, however, are basically country stations that appear under
the "Rock" preset. So a good number of the people who click on them
are not at all satisfied with what they hear & quickly click away.
Therefore the long-term listeners they pick up are balanced out by
a lot of listeners who are gone within a minute or two. That's
why those stations can look so good in the (basically meaningless)
"total listeners" numbers, while in reality their listening levels
are much less impressive. This phenomenon affects KPIG's TSL numbers
as well, to a lesser extent -- since KPIG also fails many listeners'
definition of a "Rock" station.
It'll be interesting to watch this report evolve over the next few
months. I'm looking forward to the inclusion of data from broadcast.com
(the exclusion of which renders the study severely flawed from the
start).
Webcasting is still in its infancy, and though I'm confident that
it will grow tremendously, I think a "reality check" is in
order at this point in time. Enjoyed the article a lot, & I'll be
checking back for more of your thoughts.
Thanks very much, Bill,
for an insightful analysis. I think you've helped us all understand
the numbers a lot better.
I believe the reason for the apparent inaccuracy in the rankings is
this Arbitron was quoting UNDUPLICATED MONTHLY accesses in their press
release, which is a somewhat different number than AVERAGE DAILY accesses.
Based on the explanation that you provided, the following scenario
would explain the difference in the two rankings:
If CD93 has 5,000 of the same fans listening every day, whereas KFAN
has 4,000 new people every day sampling the station, finding that
it doesn't meet their expectations, and not coming back, then CD93
would have more "average daily accesses" but KFAN would
have the bigger unduplicated monthly cume.
And of course it's a judgment call as to which is the more relevant
number.
But if Arbitron was trying to quote the biggest numbers possible,
then the latter would be the appropriate number to use.
Bandwidth
will never allow it
From:
Dave Van Dyke (dvdla@earthlink.net) |
December
10, 1999
07:57:25
|
Enjoyed your analysis of the webcast thing and I have one thing to
add. I read recently in The Standard (and have seen this figure repeated
before) that audio streaming on the Internet will not pose
any real threat to radio for one simple reason: Bandwidth
will never allow it.
If you took all the bandwidth on the Internet today and used it only
for streaming audio, only 3.2 million people worldwide could actually
listen at the same time. Sounds like an impressive number? Not really,
when you consider that's all the bandwidth the net could offer. This
would account for zero visual content being capable of being fed down
the 'nets tubes.
The Internet bandwidth would have to grow exponentially in order for
audio streaming to be a legitimate contender for our listeners' ears
en masse.
This is not even taking into account the cost of streams and the fact
that we recently did a calculation to determine what the cost would
be for KCBS-FM to stream audio for an audience large enough to reach
25% of our radio cume. (What's my math? According to qualitative research
we know that 50% of Arrow's audience is online. Take half of that
to be conservative. 25% of our weekly cume is 250,000). The best deal
I can get on just the cost of the streams (not the equipment) is $1,800
per month for 1,000 streams. Mulitply $1,800 X 250 and you get $450,000
per month!
As well as KCBS-FM is doing, I don't think Uncle Mel would approve
such an expense, do you? Have a great day.
Ask
someone who has a firewall
From: Lou Josephs (loujo@ix.netcom.com) |
December
10, 1999
10:41:32
|
Mark Cuban has it correct when he says it's all BS. Each report you
read says...bigger, larger, more scaleable.
Ask someone who has a firewall what their users in the office do with
streaming media. WJZW is the no. 1 streaming station daily, with the
stream open for about 3 hours a day. Out of a 1500 person company,
4 users. Walk by the desk...the station is background. It doesn't
change the habits. What it does is allow you to listen to radio...from
anywhere.
ABC's stations have the best Internet streaming capablities. Listen
to WLS...chances are good that the stream is coming from a data center
in the midwest. Listen to WJZW...the datacenters are in No. VA...
The closer the stream is to the end user the better it sounds...no
rebuffering..no breakup..no digital junk.
Media Metrix and AC Nielsen are also entering the streaming media
measurement field...
Arbitron uses server side measurement. When you use data from the
server all sorts of things can happen.
I have a website that is hosted on 3 servers..it gets normally 90,000
page views a month. When a cache server got stuck on the site..my
total use of the site jumped to 6 million. What I'm trying to say
here is...server side measurement is not perfect.
You
are basically on target with many of your assumptions...
From: Geoff Rich (geoff.b.rich@abc.com) |
December
10, 1999
16:45:02
|
I read
with great interest your analysis of the Arbitron.
You are basically on target with many of your assumptions, though
your 9.7 screaming headline is a tad dramatic.
Our entire ABC group is averaging approximately 1,425 streaming "listeners"
(tuners?) on a 24/7 basis, although we peak at about 5,000 listeners
in midday and average over 3,500 between 9a-6p ETZ. Now that's 23
radio stations and 4 network programs combined. The AQH leaders of
the pack are Joyner, WPLJ and WJZW averaging 150 listeners per
station, although they peak at 500-700 concurrent listeners.
Remember, however, Internet listening is still unlike Radio listening
-- it is limited to only 1 or 2 places at max per consumer, whereas
radio can be taken everywhere. The challenge will be when "Bluetooth"
1Mbps wireless internet connectivity will enable internet radio to
be as ubiquitous (and easy to tune) as terrestrial radio. Then...
Katy bar the door.
Sincerely,
Geoff Rich
Executive Vice President, Programming
ABC Radio Networks
Don't forget to
add your comments on this issue here.
Thanks.
...
.
|
|
|
|
