Last Monday, This Or That's Rebecca Kelley created a ToT comparing rappers Lil Kim and Nicki Minaj. The two performers have recently traded barbs in their song lyrics and Rebecca wanted to see who our users were pulling for. Full disclosure:  I'm a huge fan of Nicki Minaj but don't care much for Lil Kim.

As usual, we shared the ToT on Facebook and Twitter to bring votes and comments to the debate. We were thrilled to see  the traffic to this ToT take off in a hurry. On Monday, the ToT had 4,046 page-views from 3,266 users and thousands of votes. We also gained a bunch of new registered users who commented on the tot and explored the rest of the site. One of our top users even emailed us to ask about the surge in votes and views not only on the Nicki/Kim page but also on most of the other ToTs featured on the site. I dug through our Google Analytics account and everything looked normal.

We continued to get a ton of traffic on Tuesday, with very high engagement. Our new fans were voting, commenting, visiting lots of pages on the site and creating their own comparisons. The votes on the Kim/Nicki ToT were split about equally, as were the comments (as well as the comments that were in ALL CAPS, something we hadn't seen much since the Insane Clown Posse fans stopped by for a visit a few months ago).  

 

On Wednesday I started to notice something odd. What caught my eye was that Lil Kim started to race into the lead. Nicki Minaj has among the most passionate fans on the ‘net and Lil Kim is an old has-been (sorry, I'm biased remember!). So, I went back to Google Analytics. Clicking on the "Visitors" page, something immediately struck me. This surge of voters on Wednesday – the surge that massively tipped the scales for Lil Kim, were 75% Chrome users. Now, I'm not one to stereotype, but my image of a Chrome user is an indie-music-loving hipster. Not a Lil Kim fanatic. Something was clearly wrong.

A little further digging in Google Analytics and I knew we had a problem. Under Visitors > Network Properties > Service Providers > you can see which ISPs deliver traffic to your site. On Tuesday, this looked normal for us, with Comcast producing the most visitors at just over 9% of our traffic. On Wednesday, over 70% of our traffic came from AT&T internet services. That made it pretty clear. Matt, our Director of Technology, parsed through the site logs and found a single AT&T IP address hitting the site thousands of times – we clearly had a script kiddie hitting the site.

As we went through the logs in more detail, it became clear that we had a user who had written a script to repeatedly visit our site every 4 seconds, vote for Lil Kim, delete the cookie, rinse and repeat. The team here at ThisOrThat has been involved in building websites for a long time, so we've seen our share of bots. But a bot to skew a just-for-fun poll in favor of Lil Kim? That's really baffling to us. We've had a lot of discussions about how to ensure that votes are pure when the votes count for something meaningful – like an iPad contest or a presidential election. But, honestly, we never anticipated someone writing a script to determine who the internet's favorite sassy rapper would be. Now we know better!

So, I ask this question to everyone out there with a modicum of programming talent. Would you go through the effort to skew the vote for your favorite recording artist?