The Serendipity of Attention

In the recent post Forget Dunbar’s Number, Our Future Is in Scoble’s Number, commenter Adam Metz wrote:


Maybe I’m missing something, but where’s your definition of Attention? Can you add it in to the second or third paragraph? Good idea, but a little rough around the edges.

Calling me “H-Dog” is one way to get my attention. ūüėČ But back to the definition of attention. Putting it simply:

Attention = time + interest

Time being a real-world constraint. There are only but so many hours in a day, so attention is bound by that dimension. If I’m tied up with work or playing with the kids, I’m not going to give anything my attention. The second aspect is interest. Say, I do have some time. If I’m viewing something on the foraging habits of the scaup bird, my interest is quite low and I’m likely not to pay attention even though I have the time. I’ll find something else.

I will observe though, that while time is a concrete and unyielding dimension, interest is fluid and dynamic. Our moods, activities, friends and life events¬† affect what is interesting at any given point in time. It’s not like it’s totally random – there is a baseline of things that consistently interest us. While time is rigid, interest is a flexible dimension of attention.

Next question is how we find things that are of interest to us when we do have the time.

The Reducing Bands of Attention

I think I can make this statement with certainty:

You will miss the vast majority of information which would fit both your interests and time available to read

Anyone disagree? That’s probably a frustrating aspect of our information age. Am I finding the things I should know? How do I improve that? How can I be both more efficient and systematic in finding what interests me?

Technology is making it easier to be more efficient and systematic, but we’re nowhere near perfecting that. And we can’t get too perfect, because as I mentioned before, our “interests” are fluid and I don’t think we could possibly catalog all of what interests us.

Honestly, we have to accept a certain serendipity of attention. And realize we’ve got a much better system of discovery than we did just ten years ago. I’ve thought about my own experience. What’s my personal system for attention?¬† It’s a mix of ways, as the graphic below shows:


Let me describe the bands.

Dunbar’s Number: This is the theoretical limit on the number of individuals whom you can follow closely. The number is pegged at 150, a number of people which even Robert Scoble uses for his core basis of attention. My Dunbar’s number includes the 70 or so people I’m following each day on my Enterprise 2.0 List on FriendFeed. It then includes some other folks who fall outside Enterprise 2.0 but interest me in other ways.

With people in your Dunbar’s Number, you read what they create, share and talk about. My guess is that this is the core use case of Facebook members. Note that you expand the number of people you track via this group when they share content or talk with someone outside your core 150. The expansion is temporary though – based on what someone you follow has engaged with.

@replies: I use the Twitter @replies function as shorthand for the ways in which people reach out directly to you. This includes the @replies, the DMs, the Facebook messages, email itself,¬† etc. Now I’m not inundated with these, so they still get my attention. As you rise in the social media pecking order, apparently you get bombarded with these directed messages. Then they probably move to an outer band of attention for you.

Keyword tracking: This is how people, information and conversations outside my Dunbar’s Number most often get my attention. I track content that includes keywords in which I’m interested. This is the most systematic way I have for improving the efficiency and coverage of things that interest me. As I often write here, I use the Enterprise 2.0 Room on FriendFeed for this. Another good option is Filtrbox. I’m sure there are others.

Other groups: OK, you’ve got the core group of people you follow in your Dunbar’s Number. But there are others you like to keep up with as well. This is where the group functions come in to play. You can group people based on some characteristic, and check on those groups as attention allows. On FriendFeed, these are Lists. TweetDeck lets you group people.

Groups are great for when you’ve already seen your Dunbar’s List and @replies. And sometimes you just need a break from the usual topics and people on which you’ve put focus.

Random views: I do this as well. For some, it may be dipping into the public timeline of Twitter. Or FriendFeed’s everyone tab. Once you’re following a large number of people, checking out the tweets or FriendFeed entries of everyone you follow becomes a form of random views. Because you can’t possibly take in the full river of content all the time. You’d get nothing else done. But it is worth it to dip in occasionally.

Scoble’s Number Requires a System

In the graphic, I categorize all the bands outside Dunbar’s Number as the province of Scoble’s Number. To track people well outside your core 150, you need a way that aids the goals of better efficiency and more systematic coverage, while preserving the serendipity that accompanies the fluidity of our interests.

That’s where I am these days when it comes to attention. How about you?


See this post on FriendFeed:


Supply-Demand Curves for Attention

The basic ideas behind the Attention Economy are simple. Such an economy facilitates a marketplace where consumers agree to receives services in exchange for their attention.

Alex Iskold, ReadWriteWeb, The Attention Economy: An Overview

The attention economy. It’s a natural evolution of our ever-growing thirst for information, and the easier means to create it. It’s everywhere, and it’s not going anywhere. The democratization of content production, the endless array of choices for consumption.

In Alex’s post, he listed four attention services, as they relate to e-commerce: alerts, news, search, shopping.¬† In the world of information, I focus on three use cases for the consumption of information:

  1. Search = you have a specific need now
  2. Serendipity = you happen across useful information
  3. Notifications = you’re tracking specific areas of interest

I’ve previously talked about these three use cases. In a post over on the Connectbeam blog, I wrote a longer post about the supply demand curves for content in the Attention Economy. What are the different ways to increase share of mind for workers’ contributions, in the context of those three consumption use cases.

The chart below is from that post. It charts the content demand curves for search, serendipity and notifications.


Following the blue dotted line…

  • For a given quantity of user generated content, people are willing to invest more attention on Search than on Notifications or Serendipity
  • For a given “price” of attention, people will consume more content via Search than for Notifications or Serendipity

Search has always been a primary use case. Google leveraged the power of that attention to dominate online ads.

Serendipity is relatively new entry in the world of consumption. Putting content in front of someone, content that they had not expressed any prior interest in. A lot of the e-commerce recommendation systems are built on this premise, such as’s recommendations. And companies like Aggregate Knowledge put related content in front of readers of media websites.

Notifications are content you have expressed a prior interest in, but don’t have an acute, immediate need for like you do with Search. I use the Enterprise 2.0 Room on FriendFeed for this purpose.

The demand curves above have two important qualities that differentiate them:

  • Where they fall in relation to each other on the X and Y axes
  • Their curves

As you can see with how I’ve drawn them, Search and Notifications are still the best way to command someone’s attention. Search = relevance + need. Notifications = relevance.

Serendipity commands less attention, but it can have the property of not requiring opt-in by a user. Which means you can put a lot of content in front of users, and some percentage of it will be useful. The risk is that a site overdoes it, and dumps too much Serendipitous-type content in front of users. That’s a good way to drive them away because they have to put too much attention on what they’re seeing. Hence the Serendipity curve. If you demand too much attention, you will greatly reduce the amount of content consumed. Aggregate Knowledge typically puts a limited number of recommendations in front of readers.

On the Connectbeam blog post, I connect these subjects to employee adoption of social software. Check it out if that’s an area of interest for you.


See this post on FriendFeed:

Follow Everything by a Select Few, Select Content by Everyone

Item #1: Fred Wilson tweet:

@timoreilly i want to follow less people and more keywords in my twitter timeline. can’t wait for summize to get integrated into twitter

Item #2: Adam Lasnik FriendFeed post:

I switched over to reading mostly a ‘subgroup’ (“Favorites”) on FF, and was missing the serendipity of new voices. One way I’ve remedied that is to do searches on some of my favorite things (“a cappella”, “lindy hop”, etc.) and see who and what comes up.

Item #3: Steve Gillmor blog post:

A small number of Follows combined with Track produces a high degree of coverage on a daily basis.

The three items above share a common theme…limit the number of people you follow. At first, this sounds obvious. Isn’t that what people normally would do? Well no, it’s not. In social networks, there’s a dynamic whereby people tend to return the favor when someone follows them. This build up your follows over time.¬† As Louis Gray noted in a recent post:

While you might be following thousands of people and making new “friends” on Facebook, LinkedIn, Twitter, FriendFeed and all the other networks, you would likely hesitate before sending them an open invitation to your home.

“Thousands of people” I’m doing it: following 1,000+ people on FriendFeed, 600+ on Twitter. For seeing a broad range of information and opinion sources, it’s great to track so many people.

But there is a big downside. Much of what I see doesn’t interest me. The greater the number of people you follow, the more content you will see that falls outside your areas of interest. Putting this into attention terms, for any given minute you spend on a site, what is the probability you will see something that interests you?

It’s an odd phenomenon. I actually like that I’m following a lot of people, because it increases the number of instances where something that interests me will go by on my screen. But it affects the rate at which something interesting goes by. As you follow people that stretch outside your core interests, their streams do have a higher percentage of stuff that you don’t care about. And the overall probability of seeing content that interests you declines.

I want to differentiate this idea from Dunbar’s number, which describes limits on people’s ability to maintain inter-personal relationships. I’m not talking inter-personal relationships. I’m talking information foraging.

What Are You Trying to Get from Your Social Media

I enjoy following people that stream content outside my normal range of interests, such as Anna Haro on FriendFeed. It’s important to step outside the things that regularly occupy you, if you want to grow.

But the three items above show there is another rationale for people to participate in social media. Rather than seek content outside their interests, they want a concentrated dose. Personally, I’m finding I need this professionally. The Enterprise 2.0 space (my field) is fluid, and undergoing the stress of the global recession. Tracking the news, ideas, perspectives, trends and relationships is critical. For example, the microblogging trend (e.g. Yammer) is new and I’m interested in seeing how that plays out.

If you can see the point of that social media use case, you can understand the value of this idea:

Follow everything by a select few, select content by everyone

As I noted in my last blog post, I’m tracking everything for a select group of Enterprise 2.0 people, and keywords/tags for everyone.

In terms of the three items with which I started this post, Fred Wilson describes this approach. Adam Lasnik isn’t too far away. His manual searches for “a cappella”, “lindy hop”, etc. could be turned into persistent searches to find new content and people. Steve Gillmor is a little more of the social media whale philosophy, where he only wants to follow a specific set of users and then interact with the @replies on Twitter. But even Steve could add keyword tracking via a FriendFeed Room as a way to improve his daily “coverage”.

Will This Trend Grow?

I’m a fan of this use case. It fits my needs professionally. It’s almost like I have my 9-to-5 social media, and then my nighttime social media.

I suspect this use case will make more and more sense as social media expands its mainstream footprint. Information workers are the ones who will be most interested. The hardest part is figuring out which keyword/tags to follow, what sites to track and what mechanism to use for this tracking. I’d argue FriendFeed with its Rooms and Lists is perfect for this, but certainly there are other ways.

One final thought. If this trend takes hold out in the wider market, I can see people practicing a little SEO on their content. Get those hash tags in your tweets to make sure Fred Wilson will see your content (if he ever reveals what he tracks).

For kicks, I’m curious what you think of this idea. Please take a second to answer the poll below. If you’re reading this via RSS, click out to participate in the poll.


See this post on FriendFeed: