I was quoted quite heavily in an article this weekend by Vancouver Sun new media writer Gillian Shaw on the recent problems Twitter is having with spammers. Like any newspaper articles there is always a lot said in the interview that gets culled due to space constraints so I thought I would follow up with a quick post to fully present my thoughts.
The issue in the article revolved mainly around the idea that Twitter has seen a rash of spam accounts and spammy messages lately and that this would pose a serious challenge to Twitter’s continued popularity and growth. I think, in a rather perverse kind of way, that being spammed is a good thing because it means your attracting enough online attention that spammers find you worthwhile to go after.
(A caveat before I continue with this line of thinking: often websites will get hit with spam not because they’re popular, but because the security on the site is out-of-date and/or weak and the investment in time for the spammers is so low that they can afford to pummel a small site that’s easy to infiltrate.)
I remember the day two years ago when one of the blogs I manage, DeSmogBlog, got hit with a Denial of Service attack (DoS). While it was a headache to deal with at the time, it was also an indicator that someone out there thought we were getting a little too big for our britches - our message was permeating which is the whole point of a blog.
Another thing that the Denial of Service did was build up our security systems on DeSmogBlog, so when the next one came along (we’ve had several) we were able to deal with it with very little trouble.
But here’s the real kicker.
Go to the homepage of Digg.com right now and see how many of the links listed go to a crashed site. The servers for the average website cannot handle the large spike in traffic that a site like Digg can send your way and fail, sadly at the very moment most website owners wait years for. Shortly after our second or third DoS attack on DeSmogBlog, we had a story on the homepage of a number of very large sites all at once - Digg.com, Fark.com (which, by the way, is a monster site that many overlook), Reddit.com, DailyKos.com, ThinkProgress.org and HuffingtonPost.com. We were taking about 200 page loads a second and thanks to our server having dealt with similar traffic volumes when hit with DoS attacks we didn’t even break a sweat and DeSmogBlog ran in a fully functional state that allowed us to take advantage of the influx of traffic to increase our newsletter and RSS sign ups.
Now I know a DoS attack is quite different in its intent and form than regular spam, but many of the things you do to prevent and deal with a DoS just so happen to very good at dealing with high-volume spam attacks.
In other words, the DoS made DeSmogBlog stronger, more resilient and ultimately a better web service and I predict that Twitter will benefit in the same way. I am going on the good assumption that Twitter will actually deal with spammers, because if they don’t their major growth curve will quickly resemble an unhappy smiley face.
Dealing with spam on a social community platform is no small feat given that many of the measures Twitter will have to consider to defend against spam come in the form of limiting the free-flow of information. For example, when it comes to the issue of porn spam, what is considered porn and what is not? For most of us, including myself, the line is pretty clear, but drawing a line in the sand as to what is “indecent” and what is not is always a tricky task.
One thing I think Twitter could do right away is start to crack down on mass follow behavior. The idea here for spammers is to run an automated script that allows them to follow hundreds of people very quickly in the hopes of building a large number of followers who are set up to auto-follow them back. Twitter could easily do what sites like Digg.com have done, which is hit users with a cooling off period when they repeat the same behavior too many times in a row. Facebook has the same kind of thing when you post too many times to Facebook Walls in a short amount of time.Twitter already runs something that tops out the overall number of people you can follow based on a ratio of the number following you so it would be easy to ratchet this up.
As for the porn, there’s hundreds of family filter modules they could adapt.
Bottom line, spam is everywhere, Twitter should be flattered that they’re being hit so hard and they should embrace it for what it actually is: an opportunity to improve on what is already an amazing social community platform.
About the writer:
Kevin Grandia is the Director of New Media for Hoggan and Associates, a Vancouver-based public relations firm and writes on the Vancouver Social Media scene and the latest news. Drop me a line on Twitter or email me at kgrandia@hoggan.com
