Case Study Of Analytics, Trends And Insights

For a while, I’ve been wanting to write an analytics post to share observations on traffic streams to provide insights to both marketers and bloggers.  While I would love to share metrics and campaign successes from client sites, obviously permission is an issue there.

But since The Future Buzz is my own ongoing project for helping you create buzz, I thought I would take you through a few insights on some web analytics from this site for you to learn from.  Plus it’s always interesting to look at someone else’s site metrics.

Before going into the analytics, I want to point out I started this blog with goals and direction from day 1, which was a key factor for any successes I’ve had at getting traffic.  Having been blogging since 2005 or so on WordPress (here’s my old blog) and previously on some hosted services didn’t hurt either.  I also had been doing internet marketing and PR prior to starting this blog, have been on the web daily for over a decade, and done freelance writing since college.  All of those things were definitely enablers.

Some items to keep in mind:

  • My first post was on November 23, 2007 (Marketing To Internet Communities)
  • There are a total of 223 published posts on this site
  • I have had this blog for around 440 days, which means I post about once every 2 days

I turned on Google Analytics in March of 2008, so we’ve got just shy of a year’s worth of data to go through.  I’m not going to share direct numbers because I don’t want you to focus on those which is what inevitably happens – instead I want to show you trends because they are applicable at scale for any new blog or website.

Let’s start with an overview of all traffic sources:


  • The first spike was hitting page 1 on Digg for my post on BrandTags.
  • The second was hitting page 1 on Digg for my post on Newspapers Have Much To Learn About The Web
  • The spike at the end of August was my 50 viral images post being featured – interestingly enough a Norwegian language site, proving that images transcend language.
  • Large spikes in traffic throw off the scale and make it difficult to point out daily traffic trends – this site does get traffic on days without spikes even though the graph makes that hard to see.  If you removed the spikes, the graph would show a slow, but linear slope up.
  • Successful viral content does not necessarily equal huge amounts of steady return traffic, but it does create lots of links/awareness and should at least be an element of your overall strategy.
  • The spike at the start of 2009 was for my 49 Amazing Social Media, Web 2.0 And Internet Stats – which didn’t go popular on Digg but was linked to by more than 100 other pages, shared highly on Twitter (more than 208 people Tweeted it reaching 400,000+ users) made PopURLS, shared on Delicious, etc. and achieved near the same spike in traffic on that day as a popular Digg story.  Interestingly enough, for a period of time that post had so much juice it ranked on page 1 of Google for Social Media which is a phrase with strong competition (it was  there for week or so for that phrase in particular before the algorithm decided to knock it to page 2):

click image for large

Traffic sources:

  • Want a reason to use social media?  Look at my top 10 traffic sources – 6/10 are social sites.
  • Just 10 sources make up 81.4% of my web traffic.
  • Only 18.6% of traffic is from around 2,000 other traffic sources.
  • Aside from direct traffic, Digg has been my #1 source of referral traffic by a clear majority even beating out Google (thanks mostly to John Boitnott, a Digg power user who found my site on his own and decided to start sharing it)


Digg traffic is fun and gives a website visibility to lots of new eyes, but it mostly is not high converting or sustainable traffic.  Checkout the interaction time from Digg:

Other sources deliver far higher rates – Digg interactions are quick hits, just due to the nature of their users looking for the next hot thing.  I write longer content here so each time I’m dugg it’s a bit of a surprise.

The FriendFeed audience looks to have a bit more patience for longer content.  People across niches share my content in that network so this is a good sample of the difference between traffic from two different social sites:

StumbleUpon and Reddit traffic falls somewhere in between Digg and FF.

Industry-relevant blogs, while not sending nearly as much volume do send the best traffic of all – traffic from Dosh Dosh for example is extremely interested in the topics here as we cover similar subjects:

Now let’s take a look at search traffic:

Insights on search traffic:

  • Search traffic grows steadily over time, as you add content.  I’ve written my blog SEO strategy publicly in case you’re curious to see it.
  • Even in an overcrowded niche (people blogging on marketing is one of the most crowded due to the nature of the niche) there is still room for anyone to carve out search traffic if they are persistent and write link-worthy content.
  • For a brief period due to a canonicalization update from Google, in January my PR was dropped from 5 to 0.  Matt Cutts, Google’s head of webspam, was kind enough to take the time to make sure everything was okay personally (he really goes above and beyond to help people, definitely the white knight of Google).  The learning from this is if you’re PageRank drops to 0 and you did nothing wrong – don’t freak out, it might not effect your rankings too much if it’s a temporary error and Google may correct it soon.  It does appear coincidental my steady trend up had a blip right around the time Google dropped me to zero.  Alternatively it may have been from hitting page 1 for some bigger terms quickly and then Google moving them down.  Regardless, my trend seems to be continuing how it was.
  • Patience is the key to search traffic for a new site.  Don’t ever risk trying to game the system, the punishment will be worse than any short term gains.
  • Also note my search traffic is nearly all long tail, save a few words I’m ranking well for that people Google frequently.  This is due to the fact that I usually write longer-format content. There are just so many potential phrases for Google to discover. As an aside, long tail traffic isn’t even listed as a reason why I write detailed, in-depth posts – but its a nice ancillary benefit.

Other observations

Browser usage:

Most of the visitors here are web savvy, and use FireFox – no surprises.  If you’re part of the 24.61% using IE, go get FireFox.

Operating system

Windows users win, but there are a contingent of Mac and Linux users too.  Also interesting is iPhone comes in at the #4 OS.  I should get a mobile version of the site out – but if you’re going to read blogs mobile, why not just use an RSS reader?

Screen resolution

Screen resolution is actually an interesting metric because it gives you an indication of how tech savvy your audience is.  This is a subjective observation, but generally the more computer literate your audience, you’ll notice higher screen resolutions as they spend more time on the computer so they probably have better monitors.


Interestingly enough, only about half of my traffic is US-based. The next 9 most popular countries make up for 30.93% of visits, and the remaining 141 make up 16.53% of my site traffic.

Other case studies at The Future Buzz

Related posts from around the web

Why You Don’t Need to Tweet to Get Traffic from Twitter (Web Strategist)

How to Start an Internet Company That Will Be Noticed: The Proposal and Outline (SEOmoz)

Build Useful Media (Chris Brogan)

Top image credit:  vgm8383 modified under cc 2.0