Skip to main content

Most Read: Final Version

Back in 2003-2004, ComixTALK ran a series of articles under the banner of "Most Read" trying to work through how to measure the respective audience shares of various webcomics.  More recently, T Campbell borrowed the idea to generate a list of such webcomics for the former version of webcomics.com.

I've pulled together one more Most Read list, this time relying fairly heavily on Project Wonderful data.  Mostly though I went to the trouble of compiling this to point out how someone else could do a better job of it in the future.

I've always thought there was a better way to do this - some sort of clearing house of server data (or verifiable third party data).  I don't think at this point I'm going to create such a service (or pay someone to code it).  It would be pretty straightforward to do this though -- essentially you'd want to create a database where each comic would be given an account and associated with each comic you'd include some basic information about it and then identify the software package/third party service you're obtaining the data from.  Then for each statistics source you would need to write the appropriate "scraper" code to go out and collect information on a regular basis.  After that you'd just need to write a little PHP (or your language of choice) to pull data per time periods and present it in a nice format onscreen.  If someone did this they would probably be able to continue to tweak how the data is presented to balance out any idiosyncracies between the different data sources.

One great source of such data would have to Project Wonderful which does present all of the data it accumulates from serving ads.  Although not as elegant as the automated system I describe above, one can easily do a quick and dirty search of sites on Project Wonderful to get a rough estimate of not only the order of comics' readership but some actual numbers of readers (one of the flaws, among others, of using sites like Alexa and Compete). This weekend, I pulled the average uniques for the last 15 days from Project Wonderful and compiled the following list of webcomics:

  1. Questionable Content
  2. Saturday Morning Breakfast Cereal
  3. Sinfest
  4. Adventures of Dr. McNinja
  5. Girl Genius
  6. PvP
  7. Something Positive
  8. Dinosaur Comics
  9. Misfile
  10. Sluggy Freelance
  11. Softer World
  12. Phoenix Requiem
  13. Menage a 3
  14. Shortpacked!
  15. Fanboys
  16. El Goonish Shive
  17. Diesel Sweeties
  18. Devil's Panties
  19. Day By Day
  20. Scary Go Round
  21. Three Panel Soul
  22. Garfield Minus Garfield
  23. The Noob
  24. Wapsi Square
  25. Girls With Slingshots

I think tracking uniques (as opposed to visits or pageviews) is probably the closest data point to "readers", however imperfect it is in reality.  Project Wonderful data is complicated by the fact that some sites show the ad on all the pages on a single site, some on only some of the pages and others show the ad on multiple sites.  Still it's a verifiable source based on actual displays of the ad on the site (as opposed to the browser-bar approach of Alexa and Compete)

Looking only at Project Wonderful data, of course, leaves out several webcomics that are probably going to rank highly on any complete "most read" list.  That unfortunately does leave us back in the land of Alexa as the only publicly available source to sort out the audience sizes for these webcomics. (I used to look at ranking.com, but it in its current incarnation I think it's usefulness is diminished.  I note that T Campbell also used Compete.com but it doesn't seem to have a very large data sample to work with.  It also uses a very similar approach to Alexa, with again a smaller data set.  Whatever flaws come with the Alexa approach it seems are only magnified by Compete.)

Just taking the last list compiled by T Campbell, Alexa seems to sort out a list (using the 7 day measure of "rank") as follows:

  1. Cyanide and Happiness
  2. Penny Arcade
  3. xkcd
  4. Ctrl+Alt+Del
  5. Looking For Group
  6. VG Cats

VG Cats has a comparable Alexa rank to Questionable Content (the top of our Project Wonderful-generated list) so we can have some confidence in putting these 6 comics at the start of our more complete (if probably less accurate) list. 

The attempt to sort out webcomics by website breakdowns very quickly though.  The URL http://www.giantitp.com/ hosts two popular comics: The Order of the Stick and Erfworld.  The Alexa rank of the domain is higher than VG Cats and Questionable Content but what does that mean in terms of readership of the two comics?  Without more data that's unanswerable. 

UPDATE:  I should have been clearer here that almost certainly both The Order of the Stick and Erfworld are on the list somewhere.  I've emailed back and forth with Rob Balder of Erfworld and Rich Burlew of The Order of the Stick and they state that based on their server stats that Erfworld has approximately 65% of the audience that The Order of the Stick has.  Another way to put it is that about 65% of The Order of the Stick readers also read Erfworld.  That would put The Order of the Stick between the #5 and #6 spots and Erfworld between the # 8 and #9 spots.  I've updated the list to reflect this (I haven't renumbered it though). 

But since this is the last one* let's do our best to slot in webcomics without Project Wonderful data into the list using comparative Alexa ranks.  I'm not bothering with The Perry Bible Fellowship because it's not updating (add it in yourself if you feel compelled to include it) and I'm not adding Garfield, Dilbert or any other primarily newspaper comic.  One, they are hard to sort because they appear on various URLs and two, it seems odd to measure their "web" numbers while ignoring their daily audience in print.

  1. Cyanide and Happiness
  2. Penny Arcade
  3. xkcd
  4. Ctrl+Alt+Del
  5. Looking For Group
    The Order of the Stick
  6. VG Cats
  7. Questionable Content
  8. Least I Could Do
    Erfworld
  9. Megatokyo
  10. Saturday Morning Breakfast Cereal
  11. Nuklear Power
  12. Sinfest
  13. Dominic Deegan
  14. Adventures of Dr. McNinja
  15. Girl Genius
  16. PvP
  17. Something Positive
  18. Dinosaur Comics
  19. Misfile
  20. Sluggy Freelance
  21. Softer World
  22. Phoenix Requiem
  23. Menage a 3
  24. Shortpacked!
  25. Fanboys

Is it accurate?  No more or less than any other time we've run this list.  While the order may be imprecise, these 25 webcomics certainly have large readerships.  It is possible that there are other comics (especially ones that just haven't had a lot of interaction with other webcomics and therefore might be unknown in a narrow sense) that would bump any number of these down the list, but I think it would be highly unlikely to find as many as 25 new comics to bump all of these down the list.  On the other hand I would also be surprised if there weren't around five or more comics (both known and unknown) that would claim spots on the list if we had better data on them.

Re: Most Read: Final Version

 Always nice to see others who appreciate data -- and how hard it is to gather.

Thing is, Project Wonderful only counts hits on pages carrying Project Wonderful ads. I am confident that Ryan North is not claiming the superiority of his count because he has made it quite clear that his count measures something different.

Think of Google Analytics. You have to put the code on every page, or that page doesn't get counted. With PW, the ad is the code, so pages with no ads are not counted. And don't worry, they don't double count pages with two ads.

I've been using a bucket of analytics sources in my circulation verification research, and all the Brand X versions are prone to producing incorrect data with moderate confidence. It requires a lot of double and triple checking.

Even Google Analytics is cumbersome when it comes to reporting readership, which is so much more important to content based sites like comics. By coincidence, I'll have a blog post on how to calculate readership going up the evening of Friday Oct. 24th (tomorrow for me now). As soon as the HalfPixel wing ding has run its course, it will be up at (forgive the indulgence) floatinglightbulb.blogspot.com

 

Bengo

lilnyet.com

scratchinpostcomics.com

psychedelictreehouse.com

Re: Most Read: Final Version

Xaviar, chat with me on GTalk sometime and I'll share everything I worked out to track Alexa and Compete data. It required a small investment on my part, but the amount of time saved was totally worth it.

Re: Most Read: Final Version

I am really surprised by this list - I thought Penny Arcade was the unbeatable king of webcomics. Also none of my favorite comics made the list - which is surprising because I thought everyone loved them. I'd love to see a list of 25 most influential/groundbreaking webcomics, 25 most successful financially (good luck getting that data), or 25 best, highest quality, webcomics. Really interesting, thanks for putting in the work!

Maybe Penny Arcade IS the king of webcomics.

Xaviar Xerexes's picture

Well it may be the king of webcomics but it's hard to decipher.  Consider some variables of what's on the URLs of the 3 top webcomics on the Alexa-flavored list:

  1. Cyanide and Happiness - there's a lot of animation on this site; there's a forum that's big and other non-comic stuff.
  2. Penny Arcade - a lot of traffic goes to read Jerry's blog posts - maybe as many page views are for that as for the comic itself.  Data on uniques would sort that out but I don't think Alexa helps with that.  Plus PA has a big forum too.
  3. xkcd - actually this is the cleanest of the three but still has a forum too.

What we really want are readers of the webcomic -- not hits on the URL.  Unfortunately we don't really have that in terms of publicly available data for these 3.

As for the other ideas for lists -- all interesting.  A bit of a challenge to figure out how to do (and probably no way not to include loads of subjectivity on some of them) but might be worth thinking more about.

I run this place! Tip the piano player on the way out.

Re: Most Read: Final Version

 I don't understand what happened to The Order of the Stick on the final list...was it just thrown out? That seems sort of arbitrary.

Re: Most Read: Final Version

Xaviar Xerexes's picture

It is a little arbitrary, but without a breakdown of the traffic between OOTS and Erfworld I have no way of determing where to put either on the list.

I run this place! Tip the piano player on the way out.

Re: Most Read: Final Version

 We don't have a breakdown between Something*Positive and Super Stupor, either, yet S*P got on the list. Or between Cyanide and Happiness (the comic) and Cyanide and Happiness (the web animation). Or between Phoenix Requiem and the Inverloch archives, at least as far as Alexa is concerned. At what point does having diverse content eliminate an entire domain from being listed?

I just checked on Alexa, the giantitp.com domain has a 3-month average reach of 0.01526%. SInfest, the #12 comic on the list, has a 3-month average of 0.00572%, far less than half as much. So even if you just sliced giantitp.com's numbers down the middle and gave half to OOTS and half to Erfworld, each would still beat more than half the comics on the list. It's no more or less scientific than using the rankings of a commercial ad service that not everyone participates in.

 

Re: Most Read: Final Version

Xaviar Xerexes's picture

Hey interesting comments! Just to clarify -- if your point is to convince me that OOTS and Erfworld are very popular -- then reread the article, the website would be in that top 25 list of PW + Alexa.  I just didn't have a way to breakdown the two comic on the website.  It's a unique issue - there aren't a lot of domains I can think of with two popular webcomics - when I worked on this list I honestly wasn't even sure which of the two comics were more popular.  (I have never had any indication or evidence that Super Stupor is anywhere nearly as popular as Something*Positive but to the extent it is then it's another example of the same problem that's going to bedevil any effort to list comics by readership).

So sure if you want to guess, than split the Alexa number in half and put each on the list where that would lead.  (I was using a 7 day figure from Alexa btw not the 3 month figure -- as messy as Alexa is you don't make things better by mixing and matching different data sets from them.)  You raise a broader version of my two webcomics on a URL point though -- there are a lot of reasons why URLs are a bad approximation of a webcomics readership and your point about Cyanide and Happiness is well taken.

Ultimately, I don't think the second list (with Alexa data) is terribly important here -- although I think a list that was the result of an automated process could be -- why? This list is a one-shot list that I nor probably anyone else is going to try and do on a regular basis (i.e., weekly)  which you would need to do to take a list seriously as an indicator of anything.  Billboard uses automated sales data from stores and puts out lists weekly.  If Billboard had a couple of editors every year or so sit down to explain to their readers why they couldn't really do a reliable list but explain what was wrong with what they could think of it wouldn't be the Billboard lists that we do have. 

Lastly I do disagree with your characterization of PW data as inferior to Alexa data.  Alexa data is flawed and is never going to be the basis for a serious industry list.  PW data has the benefit of being very clear as to what it is and for those sites that use that ad service, it could be a reliable data point for a serious industry list.

I run this place! Tip the piano player on the way out.

Re: Most Read: Final Version

 Some problems with PW data:

Sites like mine use one PW ad for multiple domains, but only as a default; therefore, our PW score is much lower than it was when we used it as our sole advertising network.

Sites like QC use one ad format for their main page and another for their archives.

Some sites put PW on their forums, greatly increasing the total number of ad views, and some do not.

To my knowledge, no one has independently tested the data's accuracy against something like Google Analytics, even allowing for the problems cited above. I never used PW in a straightforward way with one PW ad on every single page, so I couldn't test for a pageview-to-pageview match. Ryan North is a great guy, and I have no doubt he's proud of his system, but if you're going to state that the PW data is superior, you really want to be able to back that up.

Re: Most Read: Final Version

El Santo's picture

Thanks for doing this, Xaviar.  Are you planning to do this on a semi-regular schedule?  I sorta miss the days when I could pop over to webcomics.com and see what the highest-ranked webcomics were.  Despite the iffyness of the methodology, it was probably the best reflection of what were the most read webcomics at the time.

Also, if I recall correctly, T Campbell also used compete.com, which measured the sites in terms of number of unique visitors.

Re: Most Read: Final Version

Xaviar Xerexes's picture

No plans to do regularly although if there was a lot of interest I could spec out a proposal to see how much it would cost to get a coder to write up the scrape scripts and database/php to do a more automatic collection of data from sources like PW or third party statistics packages (or even in theory stats on the server too).  That would be interesting -- to see a regularly updated pull of uniques from sites that wanted to participate or that public data was available for.  If you've got the right set of code-fu email me at xerexes AT gmail DOT com.

I run this place! Tip the piano player on the way out.