Twitter has to be the “noisiest” social media platform I’ve experimented with. I realize I’m not a typical Twitterer - I tweet to mine contacts and information relevant to my career, and to experiment and play with how Twitter can be used for community building and intel-gathering purposes. To me - and many else on Twitter I’ve chatted with - you have to read through a lot of irelevant personal minutia to pull out the occasional gem. It’s also obvious that you need to hit a certain critical mass threshold of Followers to get real value from Twitter - and that threshold requires an ever-expanding level of management, attention, and dealing with noise.
What is “Noise”?
Obviously that’s subjective - noise is anything you aren’t interested in or don’t find of value. Sometimes, that “noise” gives the human connection you need to build a personal connection / relationship with another Twitterer - so even irrelevant minutia can sometimes be turned into value by changing an impersonal, tenuous Twitter Follower/Followee connection into a friendship. To make the problem worse - noise isn’t “binary”, it’s a continuum - the degree of value you find in a tweet or other piece of conversation depends on context, and sometimes you find value you would never think to look for through an unexpected link, fact, news, or other tidbit. Regardless, noise (we can call it a “relationship cost” - the effort and time necessary to maintain a connection to somebody) increases with the number of people you follow / connect with, and everybody has a maximum limit to the noise they are capable of handling.
Given all that, Twitter posts tend to be less focused and categorized than even general blogs - e.g., higher noise. I think part of the high percentage of noise is just cultural - most users of social media expect to use the platform to share all the little personal details of their daily lives and seemingly get a lot of satisfaction from doing so. However, I also have to wonder if the prompt at the top of the Twitter home page was different, would it change the nature of most tweet content?:
vs.
Given that we’re stuck with what Twitter.com is offering, here are the strategies for dealing with noise I’ve seen or speculated about:
Only Follow High-Value Twitterers
Fortunately, not everybody has the same twitter style. I like to review the recent posts by Twitterers that catch my interest - and if I don’t see one or two posts potentially relevant to me, I pass them by and do not follow them. This mostly works for me, but even the most “on-topic” Twitterers occasionally divert into off-topic discussions about their latte or their nephew’s piano recital. Pros: if you keep the number of others you follow to a manageable minimum, you can reduce the relative noise-to-signal ratio of your Twitter feed, and leave yourself open to unexpected gems. Cons: noise level still grows as you add people you follow, and you limit your own growth (the number of people that follow you seem to be loosely associated by the number of people that you follow).
Filters and Pipes
Another solution is to filter your entire Twitter feed, or feeds from individual Twitterers, or even search terms fed into a Twitter search engine like summize.com into Yahoo Pipes or some other tool for filtering and applying rules to RSS feeds. It’s easy to do, and can provide a really good solution for making sure the content is almost 100% on-topic. Pros: low noise, broader coverage, more focused categorization of the conversations you are following. Cons: you lose Twitter’s mobile device and IM support, posts are broken out of conversational context, and you are guaranteed to miss out on all the “unexpected gems” you didn’t know to look for.
When it gets down to it, the default process is to have you do the work of filters, pipes, and automated rules manually. I find that pipes/filters work well to create specialty RSS feeds to catch categories and topics of special interest to me, that I might not otherwise catch using a Follower/Followee twitter strategy. Some great tutorial links on this subject:
- Using Pipes to extract links from Twitter
- Using Pipes to aggregate and filter on keyword content
- Using Feedrinse.com to filter Twitter RSS feeds
Dashboards
You can implement a Radian6.com or other dashboard to set up pre-defined search term profiles which will mine into Twitter, and in the future other microblogging sites. There are probably other dashboards already available which can tackle Twitter too. Pros: same as “filters and pipes”, great graphs and charts and user interfaces, easy to use and monitor. Cons: monthly license fees, and generally the boolean logic rules are less sophisticated than Yahoo Pipes or desktop RSS rule systems.
Content Categorization
Ok, this is wishful thinking right now, but it’s not necessarily far-fetched for future microblogging services. Twitter would be great from a signal-to-noise perspective if when I followed (subscribed to) a Twitterer, I checked which categories of posts would interest me, just like a regular blog - personal, professional, world news, etc. Of course, the nature of Twitter (140 characters, mobile device support, quick updates) is antithetical to adding a “select which category applies” step to the person submitting Updates.
However, in the mysterious far future when text mining technology becomes far more sophisticated, I can dream of telling my browser / reader to apply machine learning, pattern recognition and fuzzy logic, and auto-categorize the Tweets and posts for me based on what I have marked as useful in the past.
Give Up, Go Back to Blogs or Some Other Service
So if the analog process of manually reading through all the noise gets to much from you, you can stop participating in Twitter. I was at an academic conference that pointed out that no social media platform provides an option for users to “quit”, so there’s no way to know how many people just had enough and bailed out. Instead, we only find people who haven’t contributed in a long time. While indirect, it’s still a way to “vote with our feet” and let market forces cause Twitter or other social media platforms to evolve to better address the problem of information overload. There is already a strong exodus of Twitterers moving to FriendFeed, although I haven’t seen how that improves things from a “noise” level yet.
What’s the Goal?
Of course, the goal is to achieve a subjectively manageable level of noise (the number of tweets you have to read for the percentage of valuable content you find). If you only tweet with your closest friends, then maybe this is never a problem. If you are trying to develop a Twitter community building strategy, however, you have to be more strategic. You also have to consider goals of reaching some critical mass of active and involved followers, and try to reach that goal in a reasonable timeframe.
I think a couple of useful Twitter ratios could include the following:
Velocity = Followers/ Updates, or Followers / Days Since First Update (which is not reported by Twitter)
Velocity looks at how many Followers join each time the person updates. This could be influenced by the quality of update content, social factors (referalls by other Followers), or external factors (very high prominence or popularity in other blogs, social media, or other media). I think any positive Velocity ratio is a very good sign. Here is some top Twitterer Velocity ratios (rankings thanks to TwitWho.com):
- Jason Calacanis (over 30,000 Followers): 6.0
- Chris Pirillo (14,500 Followers): 3.2
- Robert Scoble (over 29,000 Followers): 2.3
- Guy Kawasaki (over 15,000 Followers): 1.8
Efficiency = Followers / Following
Looking at several “normal” people I follow, the Efficiency ratio stays pretty close to 1.0. A good ratio seems to be about 2.0 - stating that for every one person you listen to, two people listen to you. However, I think that there is an effective upper limit of the number of others any one person can Follow without using filters or just outright ignoring them (using Twitter as a one-way publishing mechanism). People who are following thousands of others do so merely to encourage reciprocation, and to increase their own visibility - and don’t really factor into our discussion of noise. However, for comparison sake, here are some sample Efficiency ratios of top Twitterers:
- Leo Laporte (Earning a “Super Picky Follower” label from TwitWho): 100.9
- Robert Scoble:1.7
- Barack Obama (47,000 Followers): 1.0
- Guy Kawasaki: 0.9
- Patrick Curl (5,252 Followers): 0.3
ROI = ?
OK, that’s nothing that can be easily measured using Twitter stats, and is the subject of a whole other discussion. I’d love to see a collection of white papers on how Twitter was used strategically, and some form of ROI was demonstrated; and if I ever come across such a list or compile one I’ll be sure to share.
Bookmark This Post: del.icio.us | Reddit | Slashdot | Digg | Facebook | Technorati | Google | StumbleUpon | Tailrank | Furl | Yahoo | BlinkList | Bloglines
Previous Post: 
Next Post: 