Optimizing signal vs noise on the Internet

Based on questions I am asked repeatedly, a lot of people think I must spend all day reading online. Trust me, this is squarely not the case. Several years ago, I used to read a lot more RSS feeds and visited a long list of sites daily to stay up-to-date. One day I decided I’d had enough of this information overload and became aggressive in improving the signal-to-noise ratio in my daily reading.

Since then, I’ve been converging toward a low information diet. Here’s how I keep up with all the information I track, and how I stopped drinking from the fire hose.

I’ll list the keys and then how they apply to every major service I use.

  1. Batch process every information source and flag interesting content for follow-up.

  2. Ruthlessly kill any information source that doesn’t nearly always add value to your day, or which provides very little value in return for the cost in mining it.

  3. Read all the great content you flag during the best time for you.

  4. Iterate until your information diet converges.

RSS. At one time I followed close to 300 feeds in Google Reader, including a lot of sites like Lifehacker, Slashdot, and BoingBoing which updated over 20 times per day. I seriously read every single Lifehacker post for three or four years. I learned a great deal from this experiment but obviously this approach is unsustainable. I was basically mining these high-volume feeds for interesting content.

Then I decided I was tired of this process. I wanted to only follow feeds that consistently (as in almost every post) add value to my life. How did I decide what to cut? I found the easiest way is to go a few days without checking RSS, letting content build up to a massive unread count. Then, review your list of feeds and unsubscribe to any feed you instinctively desire to “Mark as Read.” Some high volume feeds can be replaced by “Best of” feeds from the same source which only include outstanding editorial works from those sites. If there’s no “best of” feed, unsubscribe anyway; there are all kinds of other ways to find the content you want (see below).

You have to be pretty ruthless here, but it’s totally worth it once your feed list converges to something awesome. This process keeps the reading load in Google Reader under about 15 minutes per day. I now open Google Reader in the morning and maybe have 30 unread items, most of which I immediately look forward to clicking on.

Twitter. In some ways Twitter has started replacing RSS for me. I follow lots of friends, but also many folks I trust to curate great content. This is one reason I can get away with not subscribing to a great site like Lifehacker via RSS anymore – any truly exceptional content on LH will be shared by people I follow on Twitter. There are many oustanding curators out there like Jason Kottke, etc., who consistently point you to the good stuff.

Whenever I encounter a new site or blogger who looks interesting, I start following them on Twitter. Previously, I would add their feed to Google Reader for a probation period, but I don’t do this anymore now that I use Twitter. If they are consistently writing posts that make me click through to their site (meaning I wouldn’t want to miss a single post from them), they pass the test and go into my RSS reader.

I love Twitter in this sense because I don’t have an unread count. If I get busy, I just jump out of the community for a while and start listening again when I can. The other thing I love about Twitter is that it is efficient to maintain. I don’t use Twitter much on my computers at work. I process Twitter in line at Chipotle, while my simulations run, etc. I find it very easy to keep up with during down time.

Instapaper. Instapaper is a great service that allows you to store longer things you’d like to read later, but which perhaps don’t merit a permanent bookmark. This service has revolutionized how I handle Twitter and RSS. I usually batch process Twitter and RSS, meaning I don’t take the time to read every article when I see the link. Instead, anything that looks interesting gets saved to Instapaper. Then, I read pieces from Instapaper via the website, iPhone, or Kindle when I have time.

Facebook. I have been especially ruthless in cutting down the noise on Facebook. I have lots of friends, but of everyone in my friend list, probably less than 75-100 still show in my News Feed. I realized my daily reading was filled with updates from people I haven’t been in touch with in years and have no ongoing relationship with. For most of these friends, I enjoy being able to click over to their profile when I start wondering what they’ve been up to. But, I honestly only care to read the day-to-day commentary of people I am very close to in real life: my closest friends and family.

When scanning Facebook, I often ask myself: “whose statuses am I reading here that I don’t really want to see?” Anything meeting that criteria gets hidden. It’s nothing personal – I assume lots of people hide me, and would be surprised if they didn’t. The end result here is that I can read all of a day’s updates in Facebook from people I care to read in about 5 minutes.

Email. Honestly I don’t receive enough email for it to be much of a problem, but there are a number of filters I rely on that keep things going smoothly. I have one filter called “Kill File” in Gmail that nukes email from accounts that tend to be spammy. Another example: I set up a filter to auto-archive all emails from the Abaqus Yahoo group that don’t contain the word “concrete,” as this is the only topic on that massive list that I care about.

Other. News and other things I check on an “as-interested” basis. Occasionally when I have a few minutes I jump over to Slashdot or Digg, but this is very rare.

Bottom Line. The end result of everything listed above is an extremely high signal-to-noise ratio in all my online information processing. I don’t feel like I miss important things I care about, and I can do all this in 30 minutes of my day. It’s taken a lot of trial and error to get to this point, but it’s well worth the effort.