Don’t fall for the podcasting myth of “monthly downloads”

Photo Credit: me and the sysop via Compfight cc

Photo Credit: me and the sysop via Compfight cc

Podcasters love their stats. I’ve previously podcasted about RSS stats, but the even more mythical number is “monthly downloads” and I’ll explain why.

WARNING: this post is long, contains simple math, and includes honest stats from my own podcasting. Do not attempt to read right after lunch.

I frequently see someone touting large monthly downloads for a podcast. In one particular case, someone even boasted going from 1,000 to 20,000 downloads per month.

That sounds impressive, doesn’t it?

Misleading math of monthly downloads

“Monthly downloads” are actually highly inaccurate. They will always be cumulative and exponential with each new episode you publish and each new subscriber you gain.

Let’s assume you start a new podcast and publish four episodes in the first month. You see a total of 1,000 downloads. But this is really 250 people who downloaded all four of your episodes.

250 new subscribers × 4 episodes = 1,000 downloads

Total: 1,000 downloads

The next month, you publish four more episodes and gain 100 new subscribers. Assuming those 100 new subscribers love you enough to download your past archive, your second month will have 1,800 downloads.

250 previous subscribers × 4 new episodes = 1,000 downloads
100 new subscribers × 4 new episodes = 400 downloads
100 new subscribers × 4 previous episodes = 400 downloads

Total: 1,800 downloads

The third month, you publish four more episodes (total of twelve) and gain another 100 new subscribers. These new subscribers also go back and download your entire archive. Now your monthly downloads will be 2,600.

350 previous subscribers × 4 new episodes = 1,400 downloads
100 new subscribers × 4 new episodes = 400 downloads
100 new subscribers × 8 previous episodes = 800 downloads

Total: 2,600 downloads

With each new episode you publish and each subscriber you gain, you exponentially increase your monthly downloads. A single new subscriber could result in 100 new downloads if they download your 100 episodes.

Monthly downloads are easy to boost

Want to take 1,000 downloads per month to 20,000? One way would be to produce more episodes without getting more subscribers.

250 subscribers × 80 new episodes = 20,000 downloads
No additional subscribers
No downloads of previous episodes

Total: 20,000 downloads

Although this seems ridiculous to produce 80 new episodes in a month, it would boost your monthly downloads, all without adding even one new subscriber—you still have the same 250.

But let’s be more reasonable and mix it up. You producing more episodes, get more subscribers, and convince new subscribers to download previous episodes (you already have 50).

250 subscribers × 25 new episodes (one each weekday) = 6,250 downloads
200 new subscribers × 25 new episodes = 5,000 downloads
175 of the new subscribers × 50 previous episodes = 8,750 downloads

Total: 20,000 downloads

The more episodes you have, the higher your monthly downloads will look. But monthly downloads are essentially meaningless.

My own real stats exposed

Let’s use my complete Noodle.mx Network stats. In January, 2013, I had about 124,000 unique downloads (I could inflate this to 151,000 by including non-unique downloads).

Noodlemx Network podcast downloads-January 2013

Stats provided by Blubrry

That sounds impressive! But let’s digest my own honest statistics.

By the end of January, 2013, I had about 450 episodes in my entire network’s archive. 124,000 divided by 450 archived episodes works out to be about 275 downloads per episode for that month.

Ouch. No longer as impressive. But that whole calculation is based on the false premise of “monthly downloads.” And that monthly number goes up every month as we release about 20 episodes per month across Noodle.mx Network.

The truth-telling podcast statistic

Now let’s consider a more accurate measurement. One of our episodes, from a month ago, now has about 7,500 unique downloads. That’s not as impressive as “124,000 monthly downloads,” but this is truthful. It indicates that about 7,500 people downloaded that one episode.

(Yes, some people will download the same episode on multiple devices, which may increase the unique downloads, but this is a minority.)

Do I have 124,000 listeners? No. I had 7,500 for one particular episode. If that stat continues as the average, then I’ll know roughly how big my audience is.

Don’t lie to advertisers

The worst way to use this inflated “monthly downloads” number is when approaching advertisers. If an advertiser sponsored all of my podcasts, they would not get 124,000 impressions unless I placed their sponsorship in all 450 of my previous episodes and had my entire audience download everything all over again.

If my podcast averages 7,500 downloads per episode after a month, then I can give an advertiser a truthful number: 30,000 impressions.

4 episodes × 7,500 listeners (1 month after publication) = 30,000 total impressions

Episode longtail

I recommend one month as the general measurement for episode stats because most of your downloads will be in that first month (mostly weighted on the first week). Depending on how timeless your content is, you may continue receiving many downloads long after the first month.

For example, I looked at one of popular podcast episodes in Noodle.mx Network from 90 days ago, and here’s how the downloads look over the last 90 days.

Stats provided by Blubrry

Stats provided by Blubrry

You can see the rapid falloff about a week after the episode was released. The continued downloads in the time since then are indication of the episode’s longtail. I’ve heard that Grammar Girl‘s episode longtail is extremely impressive since she produces timeless, valuable content that is easy to consume (about 5 minutes in length).

Advanced stats would help us determine the source for those longtail downloads.

Now layer this example longtail with 450 similar episodes published a different times and you’ll see where the 124,000 downloads come from and how inaccurate the big numbers actually are.

How to track downloads

For audio and video podcasters, I recommend only two services for tracking episode downloads:

LibSyn

LibSyn provides media hosting with basic stats starting at $5 per month. They can be your one-stop shop for podcast media hosting (audio or video), podcast website, syndication to iTunes and other podcast directories, and even iOS and Android apps. You can also use LibSyn as just your media host and power everything else from a separate podcast website using WordPress (my recommendation).

Downloads will be tracked no matter how you link to your media files hosted on LibSyn, because the tracking is built into the hosting.

Get your first month free with LibSyn if you use my affiliate code “noodle” when you sign up for any plan.

Blubrry

Blubrry provides free basic stats, premium stats for $5 per month (or $50 per year), and media hosting with premium stats for $12 per month. They are the easiest media host if you don’t want to leave your WordPress interface. I also like their media stats system the best. You can integrate Blubrry‘s free or premium media stats on any media hosted anywhere (including LibSyn).

Blubrry works by adding an extra part to your media URL http://media.blubrry.com/[YOUR UNIQUE ID]/p/[PATH TO YOUR MEDIA FILE]. Downloads will be tracked anywhere you link to your media file with this full URL.

Blubrry‘s PowerPress podcasting plugin for WordPress will appropriately adjust your podcast media URLs on your website and in our RSS feed so they will be tracked.

Start telling the truth

Most of you aren’t intentionally deceiving or inflating your numbers. But as podcasters, especially with other podcasters, we need to be honest and accurate.

I recommend that we drop the “monthly downloads” number in the same trash can as “RSS subscribers.”

  • http://justinjackson.ca/ Justin

    Wow. Great post. Thanks for the transparency!

  • Jeremiah Miller

    Our RSS feed? :P

  • Jeremiah Miller

    hey, thanks for telling me about blubrrys tracking thing! I was using feedburner. What do you think of feedburner?

    • http://DanielJLewis.net/about Daniel J. Lewis

      FeedBurner is unnecessary if you have a WordPress blog with PowerPress. The only way I would recommend FeedBurner is with its MyBrand feature so you’re having people subscribe to an feed URL you own, not FeedBurner’s.

  • http://twitter.com/podcasthero Andrew McGivern

    Monthly downloads are only relevant if you want to communicate the number of total “exposures” to your content within that month. It isn’t very useful to advertisers unless they have been sponsoring the show since episode 1. The number of downloads within one month of release per episode would be a better indicator of subscribers. Or an average of the last 10 episode downloads.

    • http://DanielJLewis.net/about Daniel J. Lewis

      Exactly. Like saying, “This is your potential monthly exposure.”

  • http://epicplayground.com/ msitarzewski

    Great post Daniel. I’d suggest (as you’d expect me to) that all download numbers are a myth. They’re like print distribution numbers – and tell nothing of whether the show/content was actually listened to. So, if “monthly downloads” are a myth, so too are downloads. For the same reason.

    • http://DanielJLewis.net/about Daniel J. Lewis

      Thanks, Michael! I look forward to when all podcatchers can employ the MediaGauge tracking.

      • http://epicplayground.com/ msitarzewski

        We’ve pulled back from that initiative, so the hole remains in the market. It’s certainly no less true today than it ever was.

        • http://epicplayground.com/ msitarzewski

          What’s funny is that everyone knows this – even the people selling ads, and no one is in a hurry to fix it. Wonder why? All it takes is for the media creators (podcasters) to demand real, accurate, and verifiable numbers from every outlet that plays their media. :)

      • ftwpodcast

        If you put any content on YouTube, those stats are a bit more accurate on content. It always sucks to see a video that has 1000s of views, lose a majority of those viewers after the first few minutes.

  • Todd Cochrane

    Just remember though that 98% of podcast have burned all their downloads in the first 14 days the content has been online. “Very few shows” “less than 1%” have evergreen content. So for most shows their monthly uniques is a decent representation of the actual audience size.

    • http://DanielJLewis.net/about Daniel J. Lewis

      At this moment—and I wanted to email Angelo about this—it seems that Blubrry doesn’t show me a “true uniques” number per month. By that, I mean that if John Smith downloads 10 different episodes of mine in a month, it seems that Blubrry reports those all 10 of those downloads in my total.

      So if I released 4 episodes in a month, 1,000 unique downloads could mean 250 people downloaded all 4 episodes. Or are you saying that Blubrry‘s system reports that 1,000 people downloaded up to 4,000 episodes?

  • http://msileanespeaks.com/ Ileane

    Thank you for the providing the breakdowns Daniel. I had no idea that those stats were cumulative. I honestly I don’t think we can really know the true reach of our audience – in terms of the number of actual listeners.

  • Podcastcoach

    Great post Daniel. Not only do you do yourself a disservice by inflating your numbers, but you potentially hurt the podcasting community as the sponsor may think, “Podcast doesn’t work for us” because their expectations were set on false numbers.

    • http://DanielJLewis.net/about Daniel J. Lewis

      Great point! “You said you have 20,000 downloads per month. Why did only 20 people take action?”

  • Rueben Marley

    Excellent article, Daniel! My own little podcast has just started out, and I’m only beginning to discover how the numbers work… but you’ve provided some very useful info here. Many thanks!

  • http://inallairness.com/ Adam

    Top notch post Daniel! Thanks for the insight, it’s clear why you are an award-winner in the podcasting field.

  • ftwpodcast

    Yeah, I constantly hear “the competition” talk about thousands upon thousands of downloads, but they’re always complaining about not getting the support you’d expect from such a huge audience. It’s good to be honest with yourself, not just the community around you.

  • http://www.facebook.com/swthomas55 Spencer Thomas

    I wanted to see whether my audience was growing, and also to make “apples to apples” comparisons between episodes. The stat that I came up with for this is a cumulative age-based download count. A sample plot can be seen at http://therecoveryshow.com/wp-content/uploads/2013/04/statsplot.png . Each curve represents one of the most recent 10 episodes. The start of each curve is the day it was released, and the horizontal axis is “days since release”. I can easily see that after 20 days episode 14 (sky blue) had almost 200 more downloads than episode 11 (olive) did 20 days after its release. I can see that, in general, newer episodes have faster uptake than older ones did, which says to me that my audience is growing. I can see that older episodes continue to have new downloads (the long tail effect). And I can see that certain topics are more (or less) popular than others.

    It wasn’t trivial to create this. It’s based on Blubrry‘s “day to day trend” numbers for each episode. I wrote a script using the curl command to download the 90 day trend numbers for each episode, and then a small R program produces the graph. Because it’s based on the trend numbers, which are not reduced to “unique” downloads, the numbers themselves are inflated by 20%-30%. But the trends and relative popularity numbers are pretty clear.

    Three of the curves on the graph show a sharp uptick near the end. That’s an artifact of an incident where one client downloaded each episode about 1500 times in one day (and about 50-100 times in the two preceding days). The actual curves go out of the top of the graph, and are cut off at the previous data point.

    I find this presentation helpful. What do you think?

    • http://DanielJLewis.net/about Daniel J. Lewis

      That’s fantastic! I love the visualization. Any plans to make this publicly available? Or maybe you could offer the code to Blubrry.

  • http://foodpsychpod.com/ FoodPsych

    This is super helpful! I just started a podcast three weeks ago and was confused about how to determine the size of my audience for potential advertisers. Monthly downloads look so impressive, but the numbers on the individual episodes really are much more meaningful.

    • http://DanielJLewis.net/about Daniel J. Lewis

      A way advertisers may like to hear it is if you can calculate a minimum monthly impressions. This means something more to them.

      So if you release 4 episodes in a month and each episode will average 100 downloads within two weeks, then a one-month sponsorship of your episodes would result in 400 impressions. Release two episodes per week with these same listeners and your impressions double to 800.

      • http://foodpsychpod.com/ FoodPsych

        That’s a great way to frame it. And good to know that impressions don’t necessarily have to be an entirely different group of people every time–I’d been confused about that too. Thanks so much!

        • http://DanielJLewis.net/about Daniel J. Lewis

          A single impression actually isn’t worth much. It takes at least three impressions … to actually leave an impression. :)

          Sent from Mailbox for iPad

  • Andrew Hellmich

    Great post, thanks Daniel.
    Do you know of a way to calculate podcast subscribers or are we only able to measure downloads?

    • http://DanielJLewis.net/about Daniel J. Lewis

      “Subscriber” is an ambiguous term. Would you consider someone a “subscriber” if they faithfully visit your site every week and listen to your episodes through your site?

      I think the better term is “consistent audience.” You can get that by looking at your download numbers per episode after four weeks.

      If you really want to know about “subscribers,” you could look at the user agents/clients to see how many are coming from podcast apps and then assume those are subscribers.

  • http://www.catalystjohn.com/ Catalyst John

    Awesome read! Thanks for the info Daniel.

  • dlouismartin

    Good stuff. Thanks for the tips

  • Pingback: All Alternatives Considered: How Slate thinks a daily podcast can fit into your evening commute

  • http://blogforthelordjesus.wordpress.com Mike Gantt

    Daniel,

    This post is very enlightening. If I were to sum it up, I would say, “Do not pay attention to monthly downloads; do pay attention to recent average unique downloads per episode.” Is this an accurate summary?

    On a related issue, is there any way of learning number of subscribers (in iTunes, Stitcher, etc.)? Your last sentence seems to suggest that knowing subscribers is not helpful, but maybe you were only referring to RSS subscriptions and not directory subscriptions like iTunes, Stitcher, etc.. However, it seems subscriber counts from the directories would be a good metric to view in combination with the one you are suggesting. For example, if your recent average unique downloads per episode were 7,500 (as you use in your example in the post), then whether your total subscriptions from directories (e.g iTunes, Stitcher, etc.) was 1,000 or 5,000 would tell you something valuable, wouldn’t it?

    • http://DanielJLewis.net/about Daniel J. Lewis

      Hi, Mike!

      Yes, that’s a decent summary.

      It seems you’re actually interested in platform numbers instead of “subscriber” numbers. You can use Blubrry‘s premium stats or LibSyn‘s stats to see information about how people are downloading your episodes. This is pulled from “user agents” and you can see the breakdown of Stitcher, iTunes, mobile devices, and more.

  • Derek C. Olsen

    I never thought about having to insert an ad into every single previous episode AND have everyone re-download those episodes in order to actually reach everyone that you claim to be reaching with your monthly download stats.

    Very well thought out post. Thanks Daniel.

    • http://DanielJLewis.net/about Daniel J. Lewis

      Thanks, Derek! There are a few emerging technologies that can do that. I’m reviewing one right now. :)