At first glance, these numbers for Twitter look a bit low, but after looking at how people use Twitter, these numbers do make a lot of sense. According to another recent study from the Conference Board, the top reasons for people to tweet are "connect with friends (42%), update their status (29%) and look for news (26%)." The study also found that two out of three Twitter users use the service to interact with friends.

It is also interesting to note that another recent study from the NPD Group found that Twitter users are more likely to buy music than non-Twitter users. Chances are that this is also related to the demographic makeup of Twitters user base which tends to skew a bit older, but it also clearly shows why Twitter could be such a valuable source of traffic for retail sites.

In many ways, it is probably a good thing that brands are still trying to figure out how to best utilize Twitter. If brands want to make good use of Twitter - which, for many would mean driving traffic to their sites - they have to become part of the community. We would love for Twitter to find a viable business model so that the service can stay afloat even as it grows, but in the end, most of us use it as a personal communications medium and unless brands can find a way to become part of that in an authentic, non-creepy way, they won't be able to profit from Twitter's rapid growth - and maybe that's a good thing, too.

We've posted below a series of charts showing how many people clicked through that article hour by hour and from where they came. It's just one blog post, but this example sheds some light on a few interesting questions people are asking about the social web.

Here's how it went down. I was scanning RSS feeds Monday morning and discovered Rankin's video in the feed for Delicious Popular. I assumed that one of our competitors had written about it already, but in looking around I couldn't find anyone at all who had. I sat on it for hours, unfortunately writing another post about Twitter and one from a news tip I got via Twitter (is there anything else to write about these days?).

At 3 PM PST I wrote up the post about Rankin's class. I embedded her video and added more examples of Twitter in classrooms and related resources discovered on Delicious. I visited delicious.com/tag/twitter+classroom, opened up several pages worth of recent bookmarks using the Autopagerize Greasemonkey Script, then sorted all the bookmarks visible by number of saves using the Delicious Sort Visible Links script. That made it easy to find high quality related links to include in the post.

The post could have been much better written than it was, but it did much, much better than I expected it to. It was also an exception to the rule gleaned from a much larger aggregate study we wrote about last year finding that 4 PM PST on a Monday is one of the worst times to put up a blog post if you want it to get social media traction.

At 4 PM I put up the post and Tweeted out the link. 500 people clicked through and read it in the first hour, mostly from Twitter but some from Google Reader.

Note that Twitter desktop clients, a thousand random URL shorteners and other factors lead to a major under-counting of actual traffic from Twitter. Imagine the Twitter line below bumped up but trends remaining the same.

Traffic slowed a bit, then the post made it to the bottom of Delicious Popular after three hours. From there it hit aggregator site PopURLs, the secret winner in all these numbers. Traffic stayed strong throughought the night (West Coast night, that is) except for a short dip at 3 AM. Then the East Coast of the US started waking up and people started sharing and bookmarking the post again.

The link moved from Delicious Popular to the much more popular outright front page of Delicious.com at 9 AM PST, leaving PopURLs in the dust for a few hours. A steady trickle of visitors kept coming in that morning from Facebook.

Suddenly, perhaps because it was lunch time (noon) on the West Coast and 3 PM near or at the end of school on the East Coast, a flood of people started coming in from Twitter and elsewhere. It was 20 hours after we wrote, published and Tweeted about the article ourselves. I sat down to lunch with an iPhone developer thinking the post had lived out its short-term life at several thousand views, only to come out of lunch to discover a huge spike in readers.

Twitter search and archives are such a mess that it's probably impossible to reconstruct exactly what happened at noon, but it may have been that people got some free time, found the link on the front page of Delicious and then Twittered it out to friends. It's also possible that Delicious and Twitter users are different but overlapping groups and there's some other explanation for the Twitter spike. Google Reader click-throughs were at a day-long high at the same time.

As fast as the influx of traffic came, it was gone again. Within two hours all the sources of traffic had fallen back to their half-day lows, except for the slowly growing Facebook referrals.

Note the "other" category above includes small contributors like FriendFeed, JimmyR.com and OurSignal - but it also includes a whole lot of Twitter desktop clients and other hard-to-track sources.

A Few Take Aways

I've spent some time throughout the day looking at this data and have thought of the following:

Obviously PopURLs is not to be forgotten. It's old school, not exclusively focused on tech but clearly remains a popular place to find content.

Twitter is clearly becoming a major traffic driver, though this was a post about Twitter. It's very interesting that the spike in Twitter traffic came almost a full day after the first wave of Tweets about this article. Something else kicked it off, it's hard to say what.

ReadWriteWeb should probably be developing a bigger presence on Facebook. It's much bigger than all of these other sites but just barely made the chart for traffic. It stayed steady throughout the 24 hour period, though.

This wasn't the kind of post that was going to do well on Digg or Hacker News but those aren't the only games in town. I thought it might do well on StumbleUpon but it hasn't yet.

There's an interesting rhythm in that chart of traffic by sources, isn't there? What other patterns of interest do you see?

You can find ReadWriteWeb on Twitter, as well as the entire RWW Team: Marshall Kirkpatrick, Bernard Lunn, Alex Iskold, Sarah Perez, Frederic Lardinois, Sean Ammirati, Doug Coleman Dana Oshiro, Steven Walling and Lidija Davis.

As a website or blog publisher, you've likely often wondered, "Who is visiting my website?" Traffic stats are readily available, so you already know that your unique visitors are up this week, and average time on-site is holding steady, and you're slowly beating down that ugly bounce rate, but what you don't know is who these people are. Are they social media geeks? Are they stay-at-home moms? Are they 20-something males with an interest in extreme sports?

We're often asked why site-centric demographic data is better than direct-traffic measurement (like Omniture, WebTrends, and Google's) or syndicated panel measurement (like comScore and Nielsen's). The truth is that all three have a place in the audience-measurement landscape. But the site-centric approach has some advantages that deserve special attention.

Site-centric research has two major advantages:

1. Access to the entire site audience,
2. Customized measurement of audience attributes.

Site-centric research is the demographic equivalent of direct-traffic measurement. Publishers simply tag each page of their site with the survey instrument. This means that the sample is selected based on the entire audience, from which a small percentage is invited to participate. This is unlike a panel survey, which does not have access to the entire audience and which can sample among only the small subset of the audience that happens to belong to the panel. With site-centric research, you get a more representative sample of your audience, the ability to conduct research on very small, targeted audiences, and, because of the potential for a greater volume of collected data, the ability to dive deep into particular parts of your site and audience segments.

Complementing greater access to audiences across different websites is site-centric research that executes customized questionnaires. Instead of a "one survey fits all" approach, each survey can be driven based on the category of the site and the context of particular visitors. So, your audience is asked relevant questions, and the data collected is far more valuable to publishers and advertisers alike. Compared with the sophisticated mathematical models that behavioral tracking companies use to infer the demographics and psychographics of Internet users, site-centric research can seem very simplistic. And in some sense, it is, though there is a tremendous amount of value in "just asking," and a lot of theory behind why it works.

Based on this, there are seven reasons why everyone should be thinking about site-centric demographics.

1. Know your audience. These are the folks actually on your site: you need to know them inside and out. Beyond basic demographics are visitor psychographics.
2. Profile who is visiting your site. Traffic stats are only part of the picture. Detailed and accurate portraits of online visitors are worth their weight in gold.
3. Tailor your content and offers. Detailed profiles allow you to customize offers based on who is visiting your site at any given time. Or, you could allocate ad space based on categories, interests, and behaviors.
4. React faster to emerging trends. Because you're getting real-time info, you can spot key indicators early and adjust your messaging and offers accordingly.
5. Ensure your advertising spend is hitting the right target. Whether you're a publisher or advertiser, profiles ensure that you are delivering the right message to the right audience.
6. Increase the impact of your research dollars. Site-centric research is very cost-effective. More importantly, the data is richer and allows you to do more.
7. Develop relevant partnerships. Partners are a great way to broaden your reach, as long as you can show them what you've got.

If any of these reasons strike a chord, maybe it's time to consider what site-centric demographics can do for you: all you need to do is ask.