Engadget Analysis Part II - News Sources, Press Releases & Engadget20
June 12, 2007 | 4:17 pm
Update: You can digg this here.
This is Part II of my Engadget Analysis. Part I is here. In Part II, I’ll analyze their linking and what we can learn from it. Also, I’ve made a handy OPML file for you so that you can keep up with the top 20 sites that the Engadget folks have to keep up with to pick the best for us. Head down for the download link.
But before we start…
<blatantAdvertisement>I have a list of books I’d like to read in this Amazon Wishlist. If you like what you read, you can either buy me one of those books or donate to me (PayPal to yuvipanda@gmail.com). Also, custom analysis is available for quite a reasonable price, so if you want anything, ping me!</blatantAdvertisement>
Links
The same dataset that was used in Part I was used. So, we have the same 27,953 posts in 1,169 days at our disposal. Those 27,953 posts had a grand total of 117,739 links, at an average of 4 links a post, which is double the linkage of Scoble (at least last time I checked).
Most Linked To Domains
Here’re the top ten
|
Rank |
Domain |
Links |
Most links are back to themselves (as with Matt Cutts). In fact, it has about 25 times the number of links of it’s closest competitor (which incidentally happens to be EngadgetMobile, an offshoot of Engadget). Heck, 42% of all links were links to themselves! This makes a lot of sense, as they’re just linking back to previously run gadgets/stories that might be of interest. Here’s a graph showing the difference visually:
Links to stories and site
No.3 is also Engadget.com, but there’s a difference between finding a link to www.engadget.com and just plain ol’ Engadget.com. The difference is that the www.engadget.com link is copy pasted from somewhere and most probably links to a previous story, while just Engadget.com is hand-typed and most probably link to the site itself, rather than a particular story. The people who write the stories are savvy enough that whenever they want to visit Engadget, they’d type Engadget and press Ctrl + Enter rather than type the .com themselves. The mere idea of putting the totally unnecessary www prefix hasn’t even crossed their mind. So, when they are inserting a link to Engadget the site in their blog posts, their instincts would make completely sure that they would just type Engadget.com and forget about the www. So, if the link is to www.Engadget.com, it is to a story, while if it is missing the www, it is to the site itself J
A-List?
Also, in the Top Ten, 5 are owned by AOL (Engadget, EngadgetMobile, EngadgetHD & Joystiq), two by Google (64.233.179.104 the Google Cache and translate.google.com, the translation service), one is owned by Yahoo (Yahoo News) and one is owned by CNET (news.com). That leaves only one site in the top ten which is not owned by any major corporation, and it is akihabaranews.com. They’re a firsthand news source, located near Tokyo (or at least so they describe themselves in their about us page). This first hand source kinda surprised me, as I’ve actually never heard of them! (But maybe, that’s because I ain’t no Gadget Geek (Except when drooling over the Creative Vision M, that is))
Via Links
Via links identify where they found the material for the post. There won’t be a Via link if the post is found by a Press Release or is original content. Out of the 27,953 posts, 8,869 contain a Via Link identifying where they were found. The rest probably point to press releases & featured articles.
And here are the top ten sources:
|
Rank |
Domain |
Posts |
AkihabaraNews tops the list with 387 posts, or around 1.3% of all Engadget Posts. Google Cache comes second, followed by Textually (Cell Phone News), Joystiq (Gaming News), I4U (Cell Phone News), DAPreview (Digital Audio), and TheRawFeed, RedFerret, WeMakeMoneyNotArt and Boing Boing (Trivia). The presence of four trivia sites in the top ten is because Trivia don’t come out in Press Releases. Also, Joystiq is pretty high up here because instead of having to dig through Gaming news themselves, the people at Engadget just let their friends at Joystiq do it for them.
Pervasiveness of Cell Phones
Also, the presence of 2 cellphone specific sites in the top ten, along with the fact that it was the category with the most number of posts, indicate that cell phone related news has crept into other news channels as well. Joystiq reports on Cell Phone Gaming news, DAPreview reviews MP3 players in Cell Phones and Boing Boing reports on Cell Phones being flushed down toilets. Cell Phones are just getting more pervasive, going into every area of technology. They are no longer a specialty. They are churned out by the million, bought by the million, posted by the thousands, but not exactly commented on by a lot. They’ve silently become the first computer many people would encounter. I would like one. Any PR people out there looking for someone in the young middle-class demographic that is targeted by so many Cell Phones? 
Colorful Sources
When your top source is contributing just around 1% of your total posts, you know your sources are very varied. In fact, 969 unique domains have been listed as sources, with each domain contributing an average of 9 posts. However, of those 969, 489 proved to be the source of only 1 post and an additional 304 contributed less than 10 posts. Then there were 99 which contributed less than 25 posts, leaving us with just 79 sites which contributed more than 25 posts. Here’s a chart showing this:
Only a very small slice of that pie acts as the source for most of the posts on Engadget. That small blue colored slice with 8%. Let’s break it down further and have a look:
Note that this pie is a breakup of that small 8% slice in the pie above. So, an even smaller percentage acts as the source of most of the stuff. Which brings us to the next section about…
Dare to drink from the Fire hose? The Engadget20
Unrelated to TechCrunch20, Engadget20 is a list of 20 sites which Engadget cited as sources a considerable number of times (the yellow slice above). If you really want your daily fix of gadgetry undiluted by the kind folks at Engadget, then download the OPML file here and Go!
Read Links
Read Links are the stories themselves, while Via Links are a hat tip to the site that bought the site to Engadget’s notice. The Press Releases, the links to the product pages, the link to the cached copy of the blog post that was removed after 20 minutes are all put up in the Read Links.
Of the 27, 953 posts, 22,444 posts contain a Read Link, while the rest were probably Live Reports, Hands-on Galleries and Features.
Here’s the list of top ten most linked to domains in the Read Link:
|
Rank |
Domain |
Links |
Google Translate makes this list as well, showing that a lot of Gadget related sites are non-English. The usual big names are all here: Yahoo News, News.com, BBC, PC Magazine, NYTimes, and of course AkihabaraNews as well. BusinessWire and PRNewsWire are Press Release agencies. No surprises here except Google Translate.
A Total of 6,552 domains were in the Read link at least once, but 4,969 of them got only a single link, a further 1316 sites had less than 5 links and 420 sites had less than 25 links. That leaves us with just 120 sites with more than 25 links! Here’s the pie:
The number of sites which have at least 25 links is so small relatively that it doesn’t find any place in the chart. There is a 74% chance that Engadget will not link to a site more than once. J So, just because a new product by a new company shows up in Engadget, it doesn’t mean the company is not goanna kick the bucket. Also, the comparatively small number of sites which contribute most of the Reading Material shows the upperhand of the Press, Big Companies and the Press Releases. Press Releases ain’t yet dead people!
Takeaways
Here are the takeaways from today’s post:
· Each post has about 4 links on average
· Engadget links the most to themselves, with 42% of all links back to their own articles.
· AkihabaraNews is one of their major sources of news.
· Cell Phones are no longer a specialty. They’re just another form of a computer.
· Foreign Language sites are quite rampant, as evident by the generous use of Google Translate.
· There is a 74% chance that Engadget won’t link to the same site more than once.
· If you want your dose of gadgetry fix undiluted by the kind editorial staff at Engadget, you can grab this OPML file and read 20 blogs from which Engadget gets tipped off quite a bit. Be warned if you start feeling guilty that you can’t keep on top of the information inflow though.
What’s Next
This doesn’t quite end the series though. I tried doing a Gizmodo vs Engadget, but hit a brick wall when the Gizmodo server turned me up a blank page here. Hope to do that comparison post soon (Any ideas on getting a proper page turn up there?).
And, there’s bigger game afoot, atleast in terms of size. The processed raw data for Engadget was around 49 megabytes. Right now, my friggin freaky awesome friend Rob La Gesse (I’m just stacking up adjectives because I can’t think of anything exact to say) is using his servers to run an analysis for me that has unprocessed data run into around 7 gigabytes! I expect the processed data to be a file of around 4 gigabytes, which is the largest I’ve handled yet. My resources are constrained enough that even Engadget would be stretching them to their limits, and this new analysis would have just been impossible to think of. And, he’s even given me more access to more processing power, and that too with full administrative access to a smoking Server 2008 box. You rock Rob!
So, wait for that big analysis to hit next. I hope to learn quite a bit about System Administration and SQL Server 2005 from this adventure! Expect some technical details as well, along with some rants on School. (On an entirely off topic, mind-venting, tangential note: They’ve shifted to a new schedule that gives us 8 hours at school. 8 tiring hours L I can’t even think straight for a few hours after the ordeal is over. Made more daunting by the undue (and unneeded) pressure they’re putting on us. I might seem to be ranting out on this, but really, I would rather be learning to tune SQL Server than sit at School and memorize pieces of poetry or do manual mathematical calculations by hand. Who wouldn’t?)
But before you head over…
<blatantAdvertisement>I have a list of books I’d like to read in this Amazon Wishlist. If you like what you read, you can either buy me one of those books or donate to me (PayPal to yuvipandda@gmail.com). Also, custom analysis is available for quite a reasonable price, so if you want anything, ping me!</blatantAdvertisement>
[...] YuviSense: Codin Kid » Engadget Analysis Part II -
STUFFLEUFAGUS » Blog Archive » Engadget Analysis Part II | June 12, 2007 | 4:36 pm[…] YuviSense: Codin Kid » Engadget Analysis Part II - News Sources, Press Releases & Engadget20 (No Ratings Yet) Loading … *** Random Post *** 1 Views […]
Yuvi, It's been quite some time since I opened by Bloglines
Sudar | June 13, 2007 | 7:06 amYuvi,
It’s been quite some time since I opened by Bloglines feeds and when I opened it today, I was very surprised and happy. You have become an overnight star and congrats for that. All the best with your statbot and I am sure you are upto something very big.
Let me know if you need any help and I will be happy to do it for you.
Cheers,
Sudar
Yuvi, you have a very interesting analysis here.
Sam Abuelsamid | June 16, 2007 | 8:14 pmYuvi, you have a very interesting analysis here. If you want to continue doing this sort of thing, I would suggest a look at autoblog.com (another weblogs inc site and closing in on engadget for readers) and autobloggreen.com. The comparison between autoblog and engadget would be interesting.