In response to Bob’s popular post about Lawn Signs, Nemo commented:

“So: are the lawn signs the viral marketing strategy, or is THIS ARTICLE the real viral marketing tool?”

There’s no question about it, a significant part of the rationale for setting up a company blog is to increase brand awareness and convert readers into customers. Needless to say, our company was pleasantly surprised when the Lawn Signs post ended up on the front pages of Digg and Slashdot, among others. Along with a wealth of readers, we also found ourselves with some interesting data on just how much traffic those powerhouse links can drive.

We thought it might benefit other bloggers (both business and personal) to share this data and our main take-away: not all referred traffic is created equal.

As background, Wikipedia describes The Slashdot Effect as “the phenomenon of a popular website linking to a smaller site, causing the smaller site to slow down or even temporarily close due to the increased traffic.” There are similar effects from other high traffic websites such as getting Farked, Dugg, or -my favorite- the Goon Rush from SomethingAwful.com.

On November 6th, the day the article was posted, we received 31,218 page views. This beat our previous all time high of 35 page views on November 5th (which was driven by Bob’s equally enthralling piece on correlated subqueries in MySQL). This yields the 89,094% traffic bump mentioned in the title. The November 6th record was quickly shattered on November 7th, when we received 75,484 page views. This level of traffic subsided and the following days came in at 5,739, 2,827, and 3,376 page views. It should be noted that WordPress keeps track of all dates and times in UTC, so the first two days would be much closer to each other if we were looking at them from the perspective of the east coast of the US (-5 hours).

This is the graph of traffic from our WordPress stats page:

ver those two days, our biggest sources of traffic in descending order were:

*aggregated from multiple pages on the same domain It’s also interesting to see which links people clicked on from within our blog post. We are very pleased that over 2,500 people clicked on over to our corporate site. This table shows the top links people clicked from the blog:

It’s also worth noting that this chart does not show all of the traffic our company website got from the post. There was also a direct link to www.rjmetrics.com in the Slashdot blurb, and that sent us as much traffic as the blog itself, which already includes a lot of people who came via Slashdot. The graphic below shows the top traffic sources for November 6th and 7th at rjmetrics.com from Google Analytics.

We were surprised at first when comparing the Google Analytics data (about how many people visited RJMetrics.com) to the WordPress stats data (about how many people visited the blog). The Google Analytics chart for the relevant period is below:

Google Analytics showed more traffic on our corporate website on November 6th than on the 7th, while WordPress showed our blog had more than double the traffic on the 7th than on the 6th. There are a few explanations for this. First, there is the time zone difference between Google Analytics (-5) and WordPress (UTC). Neither WordPress nor Google Analytics allows you to look at your historical data from a different time zone, but I was able to download our Google Analytics data in hourly intervals and manually re-orient the data to UTC so it synched up with WordPress. This UTC-normalized data shows that we had 4,036 page views on the 6th and 7,116 page views on the 7th (as compared to 6,744 page views on the 6th and 4,624 page views on the 7th before normalization).

The Google Analytics numbers converted to UTC yield a jump of 76% from the 6th to the 7th, while the WordPress blog stats show an increase of 142% over the same period-still a huge gap. Here’s my theory on what caused the difference: The biggest referrers of traffic to our blog were Slashdot, Digg, and Hacker News (news.ycombinator.com). When I think about the demographics of those three sites, a typical reader of Hacker News is probably the most likely to be interested in business intelligence and software companies, followed by a typical reader of Slashdot, and then Digg. Therefore, I think those readers who came from Hacker News and Slashdot were much more likely to be curious about our website and click the link.

One way to test this is to see which referrers sent more traffic on different days. There is a poorly documented feature in WordPress where you can see your historical stats dashboard for any day by adding “&day=2008-11-07” or the equivalent date to the URL string. When we look at the referrer information for specific days, Hacker News and Slashdot make their biggest contributions on the 6th, while Digg makes its biggest contribution on the 7th. So, it seems plausible to believe that the average visitor referred from Digg is not as likely to check out our site as the average visitor from Slashdot or Hacker News.

11/6/08

11/7/08

Also, we appreciate that many of you left comments and requested information on our beta test over at RJMetrics.com. Please keep those requests coming.

Finally, we want to throw a quick plug to WordPress.com (even though they don’t pay us, in fact we pay them $55 per year for domain mapping, the ability to customize our css, and no ads). While we have to give up some flexibility to host with them (plugins, flash, and JavaScript, to name a few), at least we didn’t have to worry about handling “The Digg/Slashdot Effect” on our own server. Their servers didn’t even flinch (not surprisingly), making us confident that our $55 was money well spent.