Sonibyte Podcasting Blog

Blackberries can now get Podcasts easily

One of our customers asked me if Blackberry devices can get Podcasts.  I told him "of course," and then realized I should check.

It seemed to me that any device that is:

• Connected to the internet.
• Has a Web browser.
• Has local storage.
• Can run simple programs.

Should be able to find and download podcasts.  I knew that Apple iPhones were good at grabbing podcasts--I've seen big pick ups in the traffic for all of our feeds, partly because of the iPhone's popularity.  I also had heard that the new Android phones (based on a Google operating system) were good for podcast listening.  So, why would Blackberries be different?

Well, it seems that until earlier this year, there were limits on how much bandwidth Blackberries could use in a given time period.  This limit seems to have been imposed by RIM and its network providers, to keep from overwhelming their networks.  If you wanted to download a podcast, you had to either go after one that was really small (under 5MB) or use a work around method.

The new Blackberry Storm handsets seem to have come with improved bandwidth and new "apps" that make it easy to find and subscribe to podcasts.  So, my original answer turns out to have been somewhat right--newer Blackberries can get podcasts.

However, there is a cause for caution in this story.  The iPhone has caused a huge run up in network traffic that is already causing delays, lost connections, and dissatisfaction among some iPhone users.  As more smart phones ship from more vendors, this problem is going to get worse.  A point could come in the future where streaming video, video downloads, and podcasts are all "throttled" by carriers in one way or another.

We could try to cope in several ways.  A simple change would be to force podcatchers to be intelligent about their file requests.  As I have mentioned before, some podcatchers will request the same file many times because they do not bother to check to see if they have already downloaded it.  We could also compress our files more or lower our audio sampling rate (although that might bother those who are trying to podcast high quality music!).  And, we could break up our episodes into smaller sub episodes--perhaps with instructions to stitch them back together, once they have all been downloaded.

In any case, something good happened and we don't have to worry yet, about the bad stuff.  We welcome another community of 28.5+ million users to podcasting.  There is plenty for you to listen to and enjoy, and we are glad to have your attention.

Podcasts can be an educational tool

We have mentioned in previous posts that many people use audio podcasts as a way to learn or practice a language.  Sites such as Chinese Pod (a Praxis site) and LearnItalianPod show that canny entrepreneurs are starting to build businesses around this concept.

What about other types of learning?  Should podcasts be considered as an educational tool?

I spotted a study in New Scientist, that looked pretty encouraging.  Dani McKinney, a psychologist at the State University of New York in Fredonia, had some students listen to a live lecture, while others listened to a podcast of the lecture.  Then, she gave the students a test on the material that had been presented.  The students who listened via the podcast got a significantly better score on the test than those who did not.

The SUNY researchers speculate that the students who listened to the podcast benefited from being able to "rewind" to listen again to material they didn't understand.  They may also have benefited from not having to take notes (we humans are not very good at walking and chewing gum--or doing any other two tasks at the same time!).

It struck me that podcasting could be a perfect solution for a number of educational challenges:

• Home schooling.  Children who are studying at home would benefit from being exposed to a broad variety of material and teaching styles.  Parents would not have to worry about exposing their kids to the Internet--they could directly control which podcasts went into their child's curriculum.
• Distance learning.  Students who study at home suffer from endless distractions and interuptions.  Adult students may have to squeeze study time into their commute or during breaks at their regular job.  Podcasts are portable and a student can listen to an episode as many times as he or she requires, without any extra cost or trouble.
• Students with disabilities.  A number of students have trouble handling long hours in a classroom, carrying and reading books, etc.  Podcasts take almost no equipment and are easy to play and control.
• Giving top teachers more exposure.  Most great teachers only get to reach one or two classes of students at a time.  it is a shame not to give them the chance to influence hundreds or thousands of students--around the world.  Some of the material that a top teacher presents may be suitable for only a few students.  We can gather these scattered audiences together and move them all forward at once, using a podcast.
• Covering academic fields that generate a lot of new stuff.  Doctors and lawyers need to stay up to date on new medicines, scientific studies, court cases, appeals rulings, new laws, etc.  Podcasts of abstracts are an easy way to help them.

The folks at Finding Duclinea have done a super job tracking ideas like these, and pulling them together in various posts such as this one.  There are already some good examples of educational podcasting sites.  (E.g, The Education Podcast Network and Education.podcast.com.)  We hope that more Internet entreprenuers will see these opportunities and start pursuing them.  Of course, when they do, we encourage them to use our great tools for creating and managing podcasts!

We are adding another Podcast directory

We help our clients by automatically uploading all of their Podcast episodes to a variety of Podcast directories.  These include the well-known ones like Apple's iTunes and Microsoft's Zune, plus several others we believe add value through their indexing approach or who have interesting and unusual feeds.

Our friends over at Tomorrow Will Be Televised recently found a new directory site that we had not heard of before, Syndic8.com.  As with most of the other sites we list on, Syndic8.com hosts many types of feeds other than audio Podcasts.  In fact, we didn't see many Podcasts yet on their site (only 70 or 80?).  Unlike some other sites, Syndic8.com seems to rely on crowd-sourced/user-generated input.  The two founders of the site have tech backgrounds--and that is reflected in the site's look and feel.

For instance, their tag page practically SHOUTS its tags at you and their stats page is awfully data heavy.  Things work a bit randomly...but the Podcast marketplace is pretty randomly organized, so maybe that's the way it should be?

Anyway, we are glad to know about these guys and happy that they are helping people find interesting Podcasts to listen to.  Our Tommorw Will Be Televised Podcast feed is up there, now.  Got another directory we should consider working with?  Please let us know and we'll consider integrating it into our system.

Tomorrow Will Be Televised Joins Sonibyte

Sonibyte just added a new customer who takes our service in a new direction.  Tomorrow Will Be Televised is an interview program that has been produced for several years by a well-known media journalist--Simon Appelbaum.  The program has been broadcast as a live feed over the Web, and is currently hosted on BlogTalkRadio (http://www.blogtalkradio.com/simonapple04).

We were impressed with the high-level guests that Simon has attracted and enjoyed listening to the give and take his interviews generate.  For instance, his June 1 show featured Kelly Goode, programming chief for GSN.  Other recent shows have featured guests such as FX President John Landgraf and Rescue Me co-creator/executive producer Peter Tolan.  It seemed a shame to tie a good program like this to a one-time, weekly distribution cycle.  A lot of the material in the shows will remain interesting and relevant for many months and perhaps even years.

We discussed our ideas with Simon and he welcomed the opportunity to repurpose his show to produce a parallel podcast feed.  We did not feel that his standard one-hour show length fit the typical podcast format.  (As we have discussed in an earlier post, the average podcast runs around nine minutes.)  However, Simon tends to have two or three guests each week, so we felt we could break his shows into two or three distinct podcasts.

Our trusty sound engineers went to work.  We added a new intro and outro (including a bit of music we created specially for the show).  We both began processing Simon's current shows and also went back and reprocessed eight of his best old shows.  The result is a pretty cool podcast feed that already contains more than 20 episodes.

We have only been working with Tomorrow Will Be Televised for a few weeks, and we can't be sure how quickly this feed will get traction.  However, we feel certain that our podcast format will both broaden the impact of the show and bring in new listeners for the live broadcasts.  We expect soon to add sponsorship and other ad opportunities to the feed.  (Please feel free to contact Simon directly--he announces his email address of simonapple04@yahoo.com during each episode--if you have any interest in participating as an advertiser.)  The show is now listed on iTunes, Microsoft Zune, and several other directories, and we hope many people will subscribe to it and listen "asynchronously."

Our automated podcast creation tool kit and low sound engineer and voiceover support costs make it easy for us to provide this type of content conversion service.  We hope that other content creators will give us the chance to help them soon, in similar ways.

Useful Sources of Podcasting Information--Part 1

We thought it might be helpful to offer a set of links to useful sources of information on Podcasting.  We have found similar lists to be helpful, when we find them on other sites.

 

This type of list always seems “random,” and is necessarily incomplete.  We invite our readers to add their suggestions.  We hope to offer updates and additions to the list, every few months.

 

Podcast Free America was created by Ryan Irelan to provide information on Podcasting for all levels of users.  It includes a guide on Podcast marketing, news and information on companies in the area, and a bunch of “how to” stuff.  All well written and thoughtful.

 

As the head of a successful Podcast production firm, Scott Whitney is well-positioned to give advice to business and corporate Podcasters.  His Podcasting for Business blog has a lot of great tips for the growing ranks of business Podcasters.  The only thing I don’t like is that he allows annoying self-promo pop overs on his site.

 

Awakened Voice from Robert Safuto, covers all types of technology and new media.  However, he has done a lot of posts about Podcasting, and does a great job of fitting audio feeds into the broader context of social and streaming media.

 

Podcast411 carries interesting interviews with the creators of a number of the most popular Podcasts.  The hosts of this site have added a bunch of useful tools and links, to make their site a hub of information for those interested in the area.

 

Rawvoice offers the Blubrry Podcasting hosting service (yes I spelled it right).  They also have a pretty useful blog.  They have some of the best info I’ve seen on Podcast statistics.

 

Feedburner’s blog is pretty much a must-read for anyone interested in feeds (and Podcasts are a type of RSS feed!).  Feeburner is now owned by Google…so it is natural to see this blog also as a way of monitoring Google’s activities in the feed area.

 

NMP (New Media Pro) has a whole set of Podcasts that discuss various aspects of creating Podcasts.  Interviews, commentary, polls and advice.

 

The Hopkinson Report has done two episodes on Podcasting that are pretty handy.  The first talks about setting up to do a Podcast and the second talks about Podcast marketing.  The blog is associated with Wired, one of our Conde Nast clients.

 

Kevin Davis has published three posts in his blog that describe how Podcasting can be used to boost the visibility of smaller businesses.  I was glad to see someone talking about this aspect of Podcasting.

 

The Podpiper blog has a lot of information on using Podcasting for educators.  Ted Lai makes a good case for Podcasting as an educational tool.  The Around Podcasting blog recently offered a related post on using Podcasting in Elearning.  This blog also has a lot of other useful info on Podcasting for businesses.

 

I thought I’d close by sharing two nice examples of people talking about our in-house Podcasts.  The first is from a site on Islam that links to our Daily Koran reading.  The second is from Gerald England, one of the artists who we featured on Poem of the Day.  Unfortunately, we neglected to get his permission before we read his poem (we apologized subsequently—and apologize again, here).  I thought his post on this would be interesting to those of you who either like to track the use of your work or those who are thinking about using the work of others without permission (don’t do it!).

 

Podcast Length--Does a Podcast's length affect its popularity?

Does the length of a Podcast affect its popularity?  It is possible that it could be better either to create shorter episodes (more to the point, snappier) or longer episodes (more complete, serious, and more value per effort to download).  It is also popular that being "at the mean" would be best (reliable length, consistent).

We have shown in an earlier post that Sonibyte's Podcasts have a long "tail" of downloads.  However, it seemed to reasonable to use the number of downloads a Podcast receives over the first three days after it is published as a proxy for its ultimate popularity.

We compared this figure for about 1,600 Podcasts against their length in words and found a mild positive correlation (longer length = more downloads):

As you can see, the chart is pretty messy and somewhat unconvincing.  The 0.25 r-squared figure on the chart indicates that 25% of Podcast popularity could be attributed to how many words are in an episode.  However, there is a LOT of scatter on this diagram.

There is a region of Podcasts with high downloads that have only a middle range of words (the cloud of dots that reaches up in the middle of the chart).  There is also a group of high-word-count Podcasts, that extend across the chart in almost a horizontal, low-download-count line.

We look at this download vs length analysis for a few individual feeds.  Here is the result for the "Weekly column" feed one that we also studied in our last post:

The correlation here is poor, and if there is any slope to the line, it may be negative (i.e., longer episodes do worse).  We looked at other examples, for a variety of Podcast types.  They all showed a similarly poor correlation within the episodes from a particular feed between length in words and the number of downloads.

The data above supports some suggestions regarding optimizing the length of a Podcast:

1.  Longer podcasts may be somewhat more popular than short ones.  One explanation for this is that they give the user a better return on the effor of subscribing and downloading each episode.

2.  Podcasts with between 800 and 2400 words seem to generate the best levels of downloads.  Very few of the feeds that were longer or shorter than this, did as well.

3.  There is no benefit or cost to make any particular episode of an established Podcast feed longer or shorter than the average length for that feed.  There is no reason to pad or edit existing text, before converting it into a Podcast.

We have used the data from our partners to see if publishers should worry about the length of their Podcasts.  Our overall conclusion is that they should probably worry more about the quality of their content and which topics they focus on, than on how long their episodes last.

There is one final business point to consider.  More Podcasts are starting to carry advertising (see our recent announcement regarding our new dynamic ad insertion feature).  A twelve minute Podcast can carry three or four fifteen second ads, while a three minute Podcast could probably only carry one or two.  We predict that this factor may drive many publishers towards longer formats, in the future.

Podcast Length--Does length vary between Podcast types?

Our last post showed a wide range of Podcast lengths.  The peak in the distribution was around 1,200 words, but some Podcasts were much shorter and some much longer.

When we examine Sonibyte's data by feed, we find that our feeds divide into three groups.  A group of shorter posts that are mostly quick tips and comments peaks at around 600 words.  Feeds that are based on weekly columns and stories center around 1,200 words.  Commentaries, reviews, and opinions seem to stretch over longer time periods, and peak at 2,000 words.  Here are charts showing the distribution of lengths for one feed from each of these categories:

 

 

Notice that the Tips and Comments type of material sometimes "runs long."  A subscriber for these feeds will sometimes find that a normal three to five minute Podcast is stretching out to 10 or 20 minutes.  However, most of the feeds in this category are personality-driven--often with the writer doing his or her own voiceover conversion.  Most subscribers to this type of feed should welcome a few more words on a particular topic, from someone they feel they know and like.

The column-driven and commentary-driven Podcasts are more closely tied to the corresponding types of print articles.  Weekly columns tend to have a rigid length limit, and are not allowed by the publisher to vary much from their target length.  Commentaries, interviews, and other longer articles also tend to have a pre-specified length.

To keep subscribers happy, publishers should probably set upper and lower limits on the number of words in each feed.  Most writers will stay within those limits as part of their normal editing cycle.  If a writer has ocassional longer articles, a publisher might consider making those into a separate Podcast--perhaps with a related title (to make it easy for subscribers to find).  This separation will keep listeners from having to adjust their time schedules to suit the length of the publisher's feed.

Podcast Length--How long is a typical Podcast?

How long are most Podcasts?  Are different types of Podcast longer or shorter?  If a Podcast has long episodes, will it be more popular or less popular?

Our next three posts will try to answer these questions.  Of course, the best answer to questions about Podcast length is to say that it shouldn't matter.  The most important thing should be the quality of the content in the Podcast.

Still, I have found that many publishers want guidelines.  They want to know which text articles may be too long or too short to turn into effective Podcasts.  They also would like to know if there is a "sweet spot" for the length of a Podcast.

Let's start by looking at the average length of a Podcast.  We are a little hindered in our research, because all of our data comes from the feeds we help produce.  Sonibyte's customers are high-quality publishers who produce first-class, professionally-written content.  Their Podcasts may have a different length distribution from individually-produced podcasts, podcast-based radio shows, personal journals, etc.

We will start by looking at the distribution of lengths for the Podcasts we produced between November 1 of 2007 and October 31 of 2008.  (All of the statistics for the set of posts are based on the same time period.  We weeded out some outliers and used a set of about 2,000 episodes for our analysis.)

 

The Podcast lengths in our data set range from as few as 91 words to as many as 9,000 words.  The peak of the distribution is around 1,200 words and more than half of the episodes we produced were between 800 and 2400 words in length.  A 1,200 word Podcast would take about nine minutes to read and would result in about a 5.3 MB file.  (In case you wondered, a good rule of thumb is that a professional reader can pronounce about 2.2 words per second.  One minute of professional-quality Podcast audio occupies about 0.7 megabytes of file space.)  For most broadband Internet users, it should take less than a minute to download this file.

Podcast time-scales are in between contemporary music or a radio program segment (one to three minutes) and audio books or classical music (thirty minutes to several hours).  We have suggested already that most Podcasts are consumed during activities such as driving or exercising.  The range of times we see above fit the size of these time slots--and Podcasts are probably the only streaming/feed-type material, that suits them well.

Next, we will look at the distribution of time lengths within different types of Podcast.

How Do People Find Podcasts?

When I tell people about all the podcasting services we offer at Sonibyte, one of the questions I often get is, "Where do people go to find Podcasts?"  Most people think of YouTube, when they want a video or Amazon, when they want a book.  What is the equivalent place to browse and shop for Podcasts?

The short answer is Apple's iTunes site.  I mentioned in a previous post that more than 50% of our downloads come from users who found one of the Podcasts we host on iTunes.  I have also described the 30,000 Podcasts listed on the iTunes site as one of the most comprehensive directories of Podcasts that is available.  Many of those who use iTunes to find Podcasts also use iTunes as their podcatcher (podcatcher = software that watches for new Podcast episodes and manages the downloading process).

According to Wikipedia, Apple controls 70% of the downloadable music market.  Our data shows that iTunes holds an equally dominant position in the Podcast download market.  Here are the top five "readers" for the Podcasts we host:

 

                             

Top Five Readers
Reader	% of All Downloads
iTunes	73.0%
Mozilla	18.0%
Zune	1.7%
Windows RSS	1.1%
Yahoo FeedSeeker	0.5%
Total	94.3%

After iTunes, a lot of people like using Mozilla (in other words, the Firefox browser) to manage their Podcasts.  This makes sense, because most of us spend our entire day with either Firefox, IE, or Safari open on our desktop.  I am surprised to see anyone trying to use native Windows RSS technology to manage feeds, but this statistic result from other more sophisticated podcatchers announcing themselves as Windows RSS, even when they are not.  Microsoft (Zune) and Yahoo together control only a bit more than 2% of the market.

There is more to this story, though.  When you look at the details by market category (see our previous posts on the subject of market categories), it becomes apparent that iTunes dominates some categories more than others:

 

                                                                 

% of Downloads Using Various Download Agents
Category	Use iTunes	Use Mozilla	Use Other
Arts	73%	3%	24%
Business	58%	27%	14%
News & Politics	82%	17%	2%
One-Time	95%	3%	1%
Society & Culture	74%	4%	22%
Technology	61%	12%	27%
TV & Film	64%	32%	4%

As you can see, big categories such as Arts, News, and Society and Culture are more than 70% iTunes.  We host a few special one-time event Podcasts--these are more than 90% found via iTunes.  It is not that surprising that the geeks who like Technology blogs tend to use fancy podcatchers and disdain using iTunes.  What is interesting is the high Mozilla/Firefox use in areas such as Business and TV & Film.  This indicates to me that these users looked for Podcasts due interest in a particular subject or content-creating author.  They probably found this content while using an on-line publication site--or via a Google search.  They then did the easiest thing and subscribed using their trusty Firefox browser.

I believe that Podcasting will continue to penetrate mainstream categories such as business and entertainment.  As it does, iTunes may start to lose some of its dominant position in this market.  To counteract this, iTunes would need to make iTunes easier to use and perhaps integrate it into Mozilla or other browsers.  Any Apple-watchers out there want to comment?  Do so anonymously please, so we don't get sued for revealing Apple's product plans, before they announce them!

Describing the Podcast Market: Part 4 = Popularity and Publisher Strategies

We broke down the 40,000 Podcasts we think are currently available into a set of 15 categories. We then stepped out on a limb and estimated the size of the Podcast market last year at 500 million downloads (see our last post). OK, so which categories of Podcast are most popular and get the most downloads, and which lag behind?

Feedburner, iTunes, and Podcast Alley list which of the Podcasts in their directory had the most downloads. We categorized the Podcasts they listed and calculated what percent of them were in each category. If we assume that the popularity of all Podcasts falls in the same general pattern as it does for these top entries, we get an estimate of overall popularity by category.

The table below gives the estimate we get from the above analysis for popularity by category, along with the distribution we gave earlier for the categories in which Podcasts are published. The contrast between popularity by download and popularity by number of Podcasts produced is interesting:

                                                                                                                                       

iTunes   Podcast Subject Categories	Estimated % of   Downloads	Estimated % of   Podcasts Published	Download/Publish   Percentage Ratio
Arts	6.70%	9.90%	0.7
Business	4.10%	6.00%	0.7
Comedy	16.20%	7.40%	2.2
Education	5.50%	4.30%	1.3
Games &   Hobbies	4.50%	4.10%	1.1
Health	4.60%	3.60%	1.3
Kids &   Family	0.80%	2.00%	0.4
Music	6.40%	16.50%	0.4
News &   Politics	9.30%	8.40%	1.1
Religion &   Spirituality	6.80%	7.60%	0.9
Science &   Medicine	3.30%	3.10%	1.1
Society &   Culture	12.10%	11.20%	1.1
Sports &   Recreation	2.90%	6.30%	0.4
Technology	8.10%	6.10%	1.3
TV & Film	8.80%	3.40%	2.6
Total	100%	100%

If you are considering launching a new Podcast, you could focus on an area that has a lot of downloads and relatively fewer competing feeds. With this strategy, our list above suggests the best areas are TV & Film and Comedy. Instead, you could choose to go where the traffic is and to heck with the competition (aren’t you brave!). Then, you’d pick Comedy (again), Society & Culture, or News & Politics. Note that it looks hard to get much traction from launching a new Podcast in Music, Sports & Recreation, or Kids & Family.

These past four posts have presented an overview of the category structure of the Podcast market, its size and growth rate, and the relative popularity of different types of Podcasts. We have tried to be conservative and to share the evidence that supports our estimates. Please argue with and/or discuss these numbers. Our industry needs more information and an open discussion about its size and composition, if we are going to attract the respect and attention these numbers suggest we deserve.

Describing the Podcast Market: Part 3 = Total Downloads and Ad Potential

So far, we’ve provided evidence that there are about 40,000 “stable” English language podcasts (plus some number of “ephemeral” ones) that were downloaded in 2007 by about 20 million American podcast listeners. How many times did this audience download Podcasts each year and which types did they download most often?

Let’s assume the following things are true:

1. The Podcast market is a Long Tail market that follows a power rule distribution.

2. There are about 40,000 total podcasts in the market.

3. Feedburner’s subscriber statistics cover one third of the top one hundred podcasts.

4. For each subscriber to a Podcast, there are another four people who download its episodes. (See previous posts by us, on this issue.)

5. The average Podcast has one episode per month. (This makes monthly subscription info fit well with annual estimates.)

If these things are true, we can use the number of subscribers for top 100 audio podcasts on Feedburner, to calculate the number of downloads made in 2007. We can fit their data to a power law curve, and estimate the coefficients of the curve. We can then integrate across an assumed universe of 40,000 feeds to get the total downloads for all of these feeds. (If you want to see the charts and math behind this, send me an email!)

Our estimate from this process is that there were about 500 million podcast downloads last year. If 20 million people listened to these Podcasts (per the surveys I mentioned in the last post), they would have downloaded an average of about 25 Podcast episodes during the year.

Of course, downloads were probably skewed and concentrated. The Universal McCann survey I mentioned earlier estimated that about 25% of the Podcast audience made a daily download and 35% downloaded at least once per week. If they are correct, the core 60% of our audience (12 million users) consumes about 97% of all Podcast downloads.

Most Podcasts do not currently carry advertising. However, those that do are earning between $5 per thousand downloads and $50 per thousand. (These numbers are based on our own experience with ad placement and on discussions with other companies in the Podcast industry.) If we pick a median value of $20 per thousand downloads, the English language Podcast ad market was worth about $10 million last year.

We have seen a download growth rate over the past two years within our feeds of more than 75% per year.

At that pace, the Podcast advertising market has the potential to reach a value of $17 million in 2008 and $100 million in about four years (by 2011). While that is still a small market, relative to behemoths such as radio advertising or Web advertising, it is big enough to be interesting!

Describing the Podcast Market: Part 2 = Number of Podcasts and Listeners

The last post described a reasonable way to categorize Podcast subject matter, based on fifteen iTunes categories. There are more than 100 active Podcast directories, and it would be impossible to search, inspect, and categories the thousands of Podcasts they list. Instead, we looked at seven directories that had summaries of their listings by category. We then mapped their categories to the iTunes category system, to get eight independent estimates of the distribution of Podcasts by subject. Based on this data, we made an estimate of the distribution of Podcasts by subject.

This was a list only of what feeds have been created—not which subjects are listened to, most often. The categories with the most feeds seemed to be "social" ones—music commentary, news & political commentary, rants about life or love, personal journals, and other soft stuff that comes from the heart.  Second place went to what I would describe as "education."  This included language lessons, help with personal finance, information on games and hobbies, and podcasts about science, medicine or health.  The last group of categories--but not far behind--was entertainment.  TV & film, and comedy are in this group.

As I mentioned earlier, iTunes has about 2,000 entries for each of its categories. If their catalog is comprehensive (i.e., included every published podcast feed), the Podcast market would have about 35,000 feeds. However, when one compares the listings on the different directories, each has some unique items. It is likely that some Podcast publishers don’t get around to listing on iTunes. So, let’s be conservative and add another 5,000 “active” podcasts to those listed already on iTunes, for a total of 40,000 active Podcast feeds. (English language-only!) Of course, there could be 10,000 or more additional Podcasts that have only a single episode, are no longer being produced, etc.  Some of these “ephemeral” entries could develop into stable feeds.  But most are likely to remain undiscovered, unappreciated, and unexploited.

A recent survey by Edison and Arbitron showed that 18% of the households they surveyed had listened to an audio podcast in the past year.  That suggests that at least 20 million US consumers listened to Podcasts last year.  A similar study by eMarketer apparently estimated 18 million Americans listened to Podcasts in 2007. (I say apparently because the details of the study are available only to those who buy their study.) A recent Universal McCann study says about 30 million Americans have downloaded a Podcast. This study lumps audio and video Podcasts together.  Therefore, I think it supports the figures for audio Podcasts only, from the other studies.

Let’s assume we had about 20 million Americans downloading Podcasts in 2007. How often did they download Podcasts? Which categories did they download the most? We’ll put our guess on the answers to these questions in a future post.

Describing the Podcast Market: Part 1 = Categorizing Podcasts

The podcast market suffers from a sorry lack of reliable and objectively-derived statistics.  There are many reasons for this. One is that a podcast can be hosted and served from almost anywhere in the Web--there is no central repository or clearinghouse for them.  A second is that the amount of money spent on podcast ads is still small.  Until it gets bigger, the market will not attract the kind of tracking and validation services that provide audience figures for TV, radio, or large traffic Web sites. Finally, the biggest distributor of podcasts is Apple, via its iTunes store.  And, as most of us know, Apple is pretty secretive about its products and their performance. As far as I know, Apple does not share any download or subscriber statistics on the thousands of Podcasts it lists on iTunes.

The lack of statistics means that we must be creative if we want to answer questions such as “How many podcasts are available?,” “How many podcasts are downloaded each year?,” or "What podcast subjects are popular?"  We can not get any absolute answers to these questions, but we can put together some reasonable estimates by looking at various podcast directories and feed listing services.

Before starting this process, we need a way to categorize the subject matter of podcasts. I decided to use iTunes’ categories for my study, because they seem to represent a sensible and fairly complete schema. Also, most podcasters upload their feeds to iTunes, and therefore have been influenced by the iTunes category structure:

                                                                       

iTunes   Podcast Subject Categories	Number of   Podcasts Listed on iTunes
Arts	2,364
Business	2,277
Comedy	2,115
Education	2,180
Games &   Hobbies	2,161
Government   & Organizations	2,167
Health	2,362
Kids &   Family	2,219
Music	2,172
News &   Politics	2,162
Religion   & Spirituality	2,223
Science &   Medicine	2,338
Society &   Culture	2,175
Sports &   Recreation	2,325
Technology	2,209
TV & Film	2,176
Total	35,625

(BTW, this series of posts is based on data gathered between May and July of 2008. It is of course quite possible, that iTunes will change their category structure without notifying me!)

I've given above the number of podcasts that iTunes carries for each subject.  Do you find it as strange as I do, that each category seems to have almost the same number of entries?  Does this mean that Apple "trims" old podcasts out of its directory after a while, to keep the list for each subject from growing too big?  Or, has Apple constantly juggled and redefined its categories, so that they stay the same size? I'm relatively new to this industry and would love to have help on this question, from someone who has observed iTunes longer than I have.

If you are interested, you can find similar category lists on Podfeed.net, Blast, Odeo, Podcast Pickle, Zune, Podcast Nation, and Digital Podcast. (Note that I am only studying English language Podcasts in these posts. There are many directories in other languages, that I have not addressed.) Each of these directories has its own variations and oddities—partly because each is oriented towards its own sub market within the podcast space.  Still, we can consider each of these lists to be an independent estimate of the distribution of Podcasts by category. By combining category data from these seven directories with the information from iTunes, we were able to build this estimate of the overall distribution of Podcast publishing by category:

 

                                                           

 

 

iTunes   Podcast Subject Categories	Estimated % of   Podcasts Published
Arts	9.9%
Business	6.0%
Comedy	7.4%
Education	4.3%
Games &   Hobbies	4.1%
Health	3.6%
Kids &   Family	2.0%
Music	16.5%
News &   Politics	8.4%
Religion &   Spirituality	7.6%
Science &   Medicine	3.1%
Society &   Culture	11.2%
Sports &   Recreation	6.3%
Technology	6.1%
TV & Film	3.4%
Total	100%

It makes sense to me that the biggest number of published Podcasts is in the Music category. Music is where Podcasting got its start, and many Podcast listeners are well-educated on musical and other audio subjects. (Our company, Sonibyte, was started by musicians!) Society & Culture made sense as number two, since this category contains all types of personal commentaries, rants, and Web-cast radio shows.

As I struggled to map the data from the smaller directories to the iTunes categories, I found three categories that were especially troublesome:

ð Arts: Apple includes literature, performing arts, and visual arts in this category. The latter group overlaps with the TV & Film and Entertainment categories that other directories use. Apple also includes fashion it its Arts category. Other directories put these Podcasts into their Society & Culture types of category. There is a large group of Podcasts that relate to story telling or serial novel creation. These should probably be in the Arts category, but some directories also put these into a Society & Culture category or into a separate literature or storytelling category.

ð Government & Organizations: No one else besides Apple seems to have this category. I didn’t see any entries from this list on Apple’s list of top podcasts. Since I can’t match it up anywhere and the traffic from this area is probably not that high, I dropped this category from our study.

ð Health and Science & Medicine: iTunes splits health-related stuff into self-help, fitness, and sexuality topics (Health) and scientific/factual stuff (Science & Medicine). Spiritual approaches to health go into the Religion & Spirituality category. This is complex enough that I suspect a lot of entries in these groups overlap or are classified differently in each directory.

In our next post, we use this directory-driven data as one source for generating some market size and segmentation numbers.

Is there an 80-20 Rule of Podcasting (a Pareto distribution)?

The 80-20 rule is also called the Pareto principle.  It states that "80% of the effects come from 20% of the causes."  Does this rule apply to Podcasting?

Our clients are major publishers.  Their writers crank out weekly or monthly columns and newsletters of consistently high quality, that focus well on their readers' interests.  When these articles are converted into podcasts, using Sonibyte's voiceover system, should they all garner about the same amount of attention (as measured by total downloads).  Or, will some podcasts turn out to be home runs?

We studied a set of 185 podcasts created by a selection of our clients during the second and third quarters of 2007.  These podcasts came from five different publications and included between six and twenty podcasts from a single "feed."  (Most podcast users subscribe to a feed of podcasts that are produced by a particular author on a particular subject.  They then automatically receive each new podcast article, when it is published.)

We counted the number of downloads each podcast received during the 180 days after it was published.  We then calculated the average number of downloads that each feed generated and examined how each podcast did relative to its average.  For instance, say the average number of downloads for a podcast feed over 180 days was 10,000.  If a particular podcast in this series had 15,000 downloads, we would give it a score of 1.5.  If it had 3,000 downloads, it would get a score of 0.3.

When we broke apart the downloads in this manner, we found they fell into three distinct groups.  About 10% of the time, a podcast in particular feed would produce less than 50% of the normal number of downloads.  About 5% of the time, a podcast would produce more than 50% as many downloads as normal.  (The best performing cast produced more than four times the normal number of downloads.)  However, most of the downloads--about 75% of the total--came from the 85% of podcasts that had ordinary performance.

This analysis suggests that the best strategy for podcasting (to stay with the baseball analogy) is "small ball."  Don't hope that a podcast will be a home run, and swing away.  Instead, develop a core franchise for your articles, and hope to steadily increase your number of subscribers and average number of downloads.  We will look more at how to build subscribers, in a future post.

How long does a Podcast last?

One of the clients who uses our podcasting voiceover services recently asked me, "how long does a Podcast last?"  Since I prefer to be lazy, if I can, I turned to Google, for an answer.  Surprisingly, I did not see much information on this subject.

There is a good (if a bit dry and academic) study on Web news article on this subject from authors Dezso and Almaas.  It shows that (at least for a Hungarian site) 28% of visits to a Web news article occur on the first day, 7% on the second, and that visits decline following a power law formula to a low, but surprisingly steady value.  Is the same true for podcasts?

Our Sonibyte clients don't do music podcasting or radio program podcasting.  Instead, they are doing news, newsletter, and informational casts.  That means that their content is similar to that of the news items that the Hungarians studied.  We have several years of good information on how many downloads each of our client podcasts receives.  The chart below shows the first 180 days of download activity for about 90 podcasts that were published in the third quarter of 2007.  As you can see:

1.  About ten percent of all downloads occurred on the first day of publication.  Another 13% occurred on the second day.
2.  The number of downloads declines pretty smoothly, except for a "blip" at day seven.  This is probably because most of our customers are all on a weekly publication cycle.  The release of a new podcast in a series seems to stimulate demand for the most immediately preceding podcast (almost doubling downlaods, versus the trendline).
3.  About 50% of the first six months worth of downloads occur in the first 15 days.  After that, there is a slow and fairly smooth decline through the end of the period we measured.

What does this say about podcast longevity?

1.  The first two weeks are pretty important.  In fact, if your podcast doesn't get a lot of traffic in the first three days, it is probably never going to be widely downloaded.
2.  More total downloads will happen after the first two weeks than during them.  This means that you should probably continue to monitor, update, market and promote your podcasts, long after they have been originally published.

We will next look at the variation between different podcasts and see how downloads are distributed, among them.