Written by Alex Iskold and edited by Richard MacManus
Understanding your audience
is the key to success in any business - including blogging. Lately the Read/WriteWeb
authors have been discussing what it is that keeps readers coming back here. Our recent poll
indicated that most of you come back to this blog for Analysis and
Reviews. We are thrilled to hear this, because we focus a lot on those
two things.
But the poll results got us wondering about which posts in particular are the most popular? And we're not talking about simple page views - we want to know what content you actually liked. In the web 1.0 world, understanding what people liked was a voodoo science. Luckily, in these days of blogs and social software, there are fairly definitive ways of measuring what people like. Comments on posts, del.icio.us bookmarks, Technorati links and of course Diggs, are all entries into the fascinating world of social popularity. So we decided to put our investigative hat on and do a deep dive on Read/WriteWeb popular posts.
The obvious place to look for popular posts is in the Read/WriteWeb archives. We looked for the most commented-on posts. Below are the posts that had at least 50 comments:
Note: We close off comments on posts after about a month, in order to prevent spam.
Not surprisingly, readers of Read/WriteWeb are most passionate about Search. This is a topic close to the blog's core focus - because of Google vs. Yahoo!, Google vs. Microsoft, Google vs. the World and most importantly because this is where Web business is centered nowadays. The other posts on the list also reflect this blog's character. Comprehensive product surveys, profiles of top applications in different countries, and predictive analysis are definitely topics we spend a lot of time on. The Netscape post btw just outright hit a nerve!
Yet just like the poll, how many comments a post gets is just part of the picture. To find out more about what people like about Read/WriteWeb, we need to analyze external links to posts.
del.icio.us has become a social phenomenon, but it is now turning into a gem of hidden information. We have written before about the possibility of using del.icio.us as a recommendation engine. Today, we will look how to use it to distill the popular posts from your blog - and to understand how people perceive those posts.
You would think this would be an easy thing to do, but unfortunately it is not - because del.icio.us does not yet allow search by URL prefix. So you cannot just search for posts that start with http://www.readwriteweb.com. Instead, we had to use a trick. We searched for readwriteweb and then sifted through the posts to determine the ones that belong to this blog. As it turns out, 34 posts from R/WW were saved by at least 100 people (note: given that it was a manual process to get that data, it's possible we missed a couple). We saved these popular posts for you under a new rwwpopular account.
Here are the top R/WW posts in del.icio.us, bookmarked by at least 500 people:
The pattern on del.icio.us is less obvious, but things become more clear once we realize that del.icio.us and comments on a blog reflect different kinds of actions. Comments reflect passions, bookmarks serve as references - so there is little overlap between them. More importantly, comments (like posts) are short lived. Unfortunately in our day and age, news and even analysis has a life span of a few hours. Once a post is off the front page of a blog, it is less discoverable and typically is not commented on anymore.
The bookmarks of del.icio.us, however, have a longer lifespan. After the first person bookmarks a post, it starts traveling through the del.icio.us network, acquiring more and more links, and growing stronger. What popular bookmarks indicate is a combination of time and usefulness. All of these posts are roughly 6 months old. It is likely that in another 6 months a new batch of R/WW posts will cross the 500 threshold on del.icio.us. This is just how references and networks evolve.
Of course no popularity contest would be complete these days without checking out Digg. This social news site has become a huge source of endorsement and traffic for bloggers. Many R/WW posts have made it to the digg front page, since Digg users have an appreciation and passion for technology. So naturally, we went looking for what stories were especially popular.
Unlike with del.icio.us, it is really easy to find this information on Digg. Here is the query, using advanced search. Here then are the R/WW posts with at least 1000 diggs (which is a lot on digg):
We noted that the posts that did well on Digg are somewhat different from the ones that got a lot of comments and picked up more links on del.icio.us. The full query results told us that while Digg users love posts about search, they also love the posts about browsers. In particular the Firefox vs. IE battle is dear to their hearts. And of course, digg users love posts about Digg - especially when it's about Digg kicking competitor Netscape's butt!
Using social information to measure user information is an effective way for bloggers to understand what their readers like. It is also possible to use the methods we've outlined here to measure the popularity and effectiveness of pages on a corporate web site.
Another useful thing to do is to dive deeper into del.icio.us and digg tags and comments. These pages contain a wealth of insightful information about how your audience perceives your content.
While doing the research for this post, we compiled a list of over 50 of the most popular posts on Read/WriteWeb. We are thinking about making this available to you as a permanent tab. Please let us know what you think about this idea, as well as the techniques that we've discussed.
Image credit: www.eyeassociates.com
TrackBack URL for this entry: http://www.readwriteweb.com/cgi-bin/mt/mt-tb.cgi/1970
Comments
Subscribe to comments for this post OR Subscribe to comments for all Read/WriteWeb posts
This is fantastic Richard - I'd not considered this type of analysis before - great stuff.
Posted by: Darren | February 15, 2007 12:26 AM
Great roundup.
check the link for "Social Bookmarking Faceoff (74 comments)" there seems to be an error :) (i think)
Posted by: Diana | February 15, 2007 2:15 AM
I'm glad you didn't just look at the number of comments. That doesn't necessarily indicate which are the most popular among your readers, but perhaps just indicate which were linked the most from outside sources or were dugg. In other words, they might be posts that matter most to people who aren't your regular readers. ;)
Posted by: Josh | February 15, 2007 2:45 AM
Wouldn't it be great if these kind of statistics were pulled from social sites into existing web-analytics software? Imagine opening Google Analytics and seeing a correlation between number of del.icio.us bookmarks and sales, or ad hits.
Problem is - surely this only works for large sites with a mostly tech audience? If half of your readers don't use digg, how relevant is the information?
Posted by: James Brown | February 15, 2007 3:12 AM
@Diana - thanks for spotting it, I just fixed it.
Alex
Posted by: Alex Iskold | February 15, 2007 5:27 AM
@Josh, we are thinking about having most popular posts page with all this information. In this post you see the rankings of some articles from del.icio.us and digg and we think it would be useful to have this info on one page for all.
Alex
Posted by: Alex Iskold | February 15, 2007 5:29 AM
@James,
Richard and I were just talking about this last night - we need a tool to do that. The fact is that since there apis for del.icio.us and digg, building this should not be difficult.
Alex
Posted by: Alex Iskold | February 15, 2007 5:30 AM
I did blog this but not sure if the trackback worked :) Please feel free to check out my response at the blogged linked on my name!
Great article btw, social media is packed-full of powerful demographic information.
Posted by: Matt Harwood | February 15, 2007 6:20 AM
Nice analysis report. Now, I know which article is interesting. Keep it up :)
Posted by: PohEe.com | February 15, 2007 7:25 AM
I would keep this kind of analysis in proper perspective. Bear in mind that in a given time frame a certain topic may be the flavor...does not mean that this maybe so in the medium to long run.
In the case of RWW I understand its vertical positioning. The challenge in my opinion is to keep the topics as varied as possible and the analysis and reviews relevant.
Posted by: Adrian keys | February 15, 2007 8:34 AM
To be blunt, I think this analysis of traffic from social sites is backwards.
The statistics coming from getting on Digg, Del.icio.us, Reddit or spikes from popular blogs like Slashdot, LifeHacker and Problogger are anomalies (all of which happened to me this week -- it's been a busy week).
They don't tell you what your readers like, they tell you which of your articles intersected the interest of another community -- which aren't necessarily the people who read you day in and day out.
Also, none of these are closed systems. How an article does on one of these social sites is largely a function of when it was submitted, who submitted it, and whether or not there is another high volume of readers coming from another site.
Get linked on LifeHacker *guarantees* hitting the front page of Del.icio.us. Get Dugg increases your chances of getting Lifehacked... but none of it really has anything to do with what your readers liked.
Taking outside linking out of the equation, it's like saying that the articles your readers liked are the ones that have the most traffic, when realistically those are the ones that are ranked best for a common search term.
I think going by the # of comments is spot on, as is going by the number of people who clicked from the RSS feed to the article (which is a stat I think feedburner gives).
Posted by: engtech | February 15, 2007 11:17 AM
FYI, some people were talking about an API for search for your most popular posts on the various site. To my knowledge all of the social sites except for StumbleUpon give you a mechanism for search on URL. On my "most popular posts" page (http://engtech.wordpress.com/most-popular-posts/ ) I give links to searches for my site.
I think another valid metric would be "most linked to posts" -- but I haven't seen any mechanism for getting that info... not that Technorati catches even 50% of the times you're linked to.
http://digg.com/search?s=engtech.wordpress.com&submit=Search§ion=news&type=url&area=all&age=all&sort=most&search-buried=1
http://reddit.com/search?q=engtech.wordpress.com&s=ups
http://del.icio.us/search/?setcount=100&all=engtech
Posted by: engtech | February 15, 2007 11:25 AM
@engtech,
You raise good points. Let me ask you a question. Do you think people would bookmark article on del.icio.us without reading it, just because they heard about it or were recommended by a friend?
Alex
Posted by: Alex Iskold | February 15, 2007 11:26 AM
engtech, those are good points and we will check out those Feedburner stats. All of the 3 current sources (comments, delicious, digg) have their weaknesses, but mashed up together with something like Google Analytics (as James suggested in comment 4), they can give you clues as to what your audience likes.
But in the end there is also a bit of 'magic' involved. For example we're still trying to find the magic formula for getting more community participation on R/WW - i.e. comments on each post. If anyone knows that, then please do tell ;-)
Posted by: Richard MacManus | February 15, 2007 11:55 AM
You raise good points. Let me ask you a question. Do you think people would bookmark article on del.icio.us without reading it, just because they heard about it or were recommended by a friend?
I would bet dollars to donuts that del.icio.us saves fall almost exclusively into three categories:
- to read later
- I've skimmed this, and I want to keep this for a reference
- I like it and I want to save it (maybe for reposting on your blog...)
I've started to get in the habit of doing the later with:
http://del.icio.us/engtech/linkblog
---
Del.icio.us would be a great metric if only you could differentiate the saves of your regular readers from the saves of one time visitors... that's really the problem with using any social site as metrics, once they take off it's all the outsiders / community regulars voting on it -- not your readers. Social sites are a source of traffic, not a reflection of existing traffic.
One way to correct the metrics might be to divide the votes by the number of views. So del.icio.us saves / times that page was viewed would be a percentage like "9% of the people who viewed this article saved it".
---
http://www.majikwidget.com/
I've seen Guy Kawasaki use MagicWidgets to let readers rate articles. It has the same fault where if an article is dugg/whatever that you get so many votes that aren't regular readers... but the average (discounting the number of votes) might still be a good indicator. At the very least it is a metric of an engaged visitor (someone who takes the time to rate the post). Theoretically, you could also track return visitors with a cookie and only count the votes of regular readers (not sure if Majic widget supports that -- they should).
---
re: How to encourage comments.
This is counter-intuitive but it really works. Don't tell the full story / be wrong / don't cover all your bases / leave open areas for debate. People will leave a comment if they want to correct/disagree. If they agree with your point of view then you won't get that many "me too" posts.
Insert Kathy Sierra graph of an upside down bell curve showing how users are most engaged when you're "mostly right".
---
The more I think about it, the more I want a tool where I can type in my URL, and it'll tell me the most linked to posts on my blog. That would be so much more useful than Technorati WTF.
Posted by: engtech | February 15, 2007 1:24 PM
We were trying to find an easy way to perform popularity analysis of web pages when I found socialmeter.com. Very useful. Great tool!
Daniel
duzzio.com
Posted by: Daniel | February 15, 2007 3:58 PM
I found this really interesting reading. I don't think I'd show up in your count because I subscribe to your blog through NewsGator and save specific posts to my Clippings file (not to delicious). Typically I use delicious for saving URLs to static sites (not blogs); however, if I see a link to a site I think I want to check out, blog or not, then yes I DO save it to delicious without having read it.
Sorry if that stuffs up an earlier theory about delicious, and your counting!
Posted by: catherine | February 15, 2007 5:10 PM
@catherine
Why not save the actual links to articles?
Alex
Posted by: Alex Iskold | February 15, 2007 5:32 PM
@Richard
Regarding more comments I have no magic for you, but two ideas.
One of the problems with comments are that the gold is often among a lot of small-talk (RWW fortunately has a mature audience and does not suffer as much as i.e. TC), if the first few comments are not great then people are likely to get discouraged and move on. A system that would allow users to rate comments, and have brilliant comments float to the top would make reading the comment more interesting.
A threaded or other wise more debate friendly system would then encourage more users to reply to the comments and not just the post.
Since a lot of your reader probably never visit the site but just reads the RSS-feed putting the top 5 rated comments at the end of the post inside the feed might be a way to encourage further commenting (but also might be disruptive to some RSS readers if the article content is continuously updated).
Posted by: Emil | February 16, 2007 12:16 AM
You start off by saying "And we're not talking about simple page views - we want to know what content you actually liked!"
Wouldn't it be nice, to actually prove your point and supply the actual pageviews from your own log?
Posted by: Niels Jakob Buch | February 16, 2007 8:58 AM
I think this analysis is spot on. The trick to having access to so much information is the trick of being able to sift through the haystack and find the needle, or needles, that you love.
Digg and the other sites are awesome b/c they allow really targeted searches on the content we love. Hopefully the process of writing, providing, tagging and sharing will get even easier!
Posted by: Lawton Chiles | February 16, 2007 9:08 AM
The process will definitely get easier. I envision a one touch key-in system where the process will be uniformed and efficient.
Posted by: sewdough | February 16, 2007 11:28 AM
and one to add about what digg users love posts - Ubuntu!
Posted by: jetpeach | February 16, 2007 5:26 PM
Another way to discover what posts your readers find most interesting would be to add a little extension to your blog so that each of your blog posts can be rated by your readers.
We just launched this feature at BlinkList. You can get an early look here:
http://www.blinklist.com/mreining/get_blog_extension.php
In the future, we will also offer you a feed of the "top rated" posts of your blog so that you can automatically see which posts your readers enjoyed the most.
Would love to hear your thoughts on this extension and how to further improve it.
Mike
Posted by: Mike | February 17, 2007 7:48 AM
Mike - looks interesting, I will check it out.
Alex
Posted by: Alex Iskold | February 17, 2007 10:45 AM
socialmeter.com is pretty cool, except I'm not entirely sure how they calculate their numbers. When I manually compare the links I get different numbers.
The joy of statistics. No wonder they're all made up on the spot :)
Posted by: engtech | February 17, 2007 3:44 PM
Great overall analysis of audiences vs social realm.
Posted by: Joost | February 17, 2007 4:05 PM
@Alex re #17 and #18:
Sorry for delayed reply - I was on holiday at the beach nr Auckland reading the Herald article on RWW :-) I may save the actual link or I'll save an article on a topic that may contain several links potentially of interest. Why? because I'm lazy, because I come across stuff when I'm supposed to be doing something else and don't want to completely lose sight of my real job, etc. And often it's because I want to check out more of the blog where I got the ref from as well as the ref itself. I use Newsgator's clippings because it's easy and it's there.
Posted by: catherine | February 22, 2007 5:00 PM