12.14.2011

Google+: Why Circles Fail and a Possible Solution

“I fell through a crack for years. Historically, I am nothing because I fit in no category.” – Howard Hodgkin

I must admit, I've been using Google+ for a few months, and I'm a huge fan. However, as with anything, it could be better. The biggest issue I have with Google+ at the moment is circles, and how they have failed me.

Circles
With g+ circles have a few different purposes, or at least seem to have been created with that intention.
1. Circles allow you to control who sees what you share
When you make a post you select your audience by selecting specific people, specific circles, all of your circles, the enigmatic "extended circles", or Public. If you add your mom to your "family" circle and then select your family circle when you post - she and everyone else in the circle can see it.
2. Circles allow you to control what you see
If I add you to a circle (maybe the "people who've visited my blog" circle), everything you post to Public should show up in my main stream. I can filter what I see by selecting circles on the left - so I could select the "people who've visited my blog" circle and see only shares from people in that circle.
3. Circles allow you to share themselves with other g+ users
If you're new to g+ a good thing to do is find a post where someone has shared a circle. Some people have/make the time to categorize g+ users into circles based on what they tend to post - or by shared interests. For example, one of the groups that seems to share circles often is the g+ photography subculture. People like to share their artistic endeavors and photographers seem to be finding g+ ideal for that purpose. Someone adds photographers to a circle and then shares that circle. You can add it to one of your existing circles or make a new circle for it.

How Circles Fail
G+ circles fail me in purpose #2, controlling what I see. The reason:
I have no idea what you're going to share.

Human beings are a multifaceted species with each individual having a variety of interests. A photographer might also like to cook. When he decides to share original content it could very well be a new picture or a favorite recipe. Maybe he just watched the Republican debate and wants to share his opinion about that.

Despite my having added him to my photographer circle so I can see the pictures he shares, when I filter my stream by my photography circle there might very well be recipes in there. Now, there are work arounds. You can let the photographer know you're interested in photography, and he can put you in a circle for "people that like my pictures". Then he can share his pictures with that circle and his recipes with his recipe circle, etc.

However, that's a giant pain in the rear. I have around 1300 people circled at present and contacting them all to let them know what kind of content I want to see (what circle they should put me in), is pretty unrealistic. My current situation is that I can't predict what categories of content these 1300 will be sharing, so I can't correctly put them in a certain circle. Similarly, even if I tell you that I like your photos when I first circle you, you might decide later to share a recipe. You'll have no idea if I'd like to see that too.

As a result of all this, I have a very rapidly scrolling stream that I can't filter by type of content. Therefore, I propose the following solution:


Crowdsourced Categorization
I think the best way to achieve the failed goal of content control is via Crowdsourced Categorization. At least that's what I'm calling it.

 - Setting Categories
I'd like to see the option to set categories when you make a post. It could start with a fairly diverse list of commonly seen categories that will be added to periodically. We'll use the 3 categories: Photography, Humor, and Politics for this example.

Once implemented, whenever I make a post I will be allowed to checkbox categories from the list and/or add a new category. I might post a picture of the mayor eating a doughnut and check the categories of Politics, Humor, and Photography - and use the write-in category of doughnut (only one write-in per post per g+ user will be allowed).

 - Why Crowdsource?
If we stop there though, we run into the problem of people being bad at categorizing. Everyone has a different perspective. You might not think there's any humor in a picture of the Portland Mayor eating a doughnut. When you filter by Politics and see my post, maybe your opinion dictates it be categorized as Local Politics, or Portland Politics.

With Crowdsourced Categorization implemented, everyone who views the post would have the opportunity to weigh-in on how the content should be categorized. Perhaps you'd uncheck Politics and write-in Portland Politics. Then when the future g+ users find the post in the stream, both Portland Politics and Politics would be options they could check - at this point, each category would effectively have 50% of the category vote.

If left unchecked in this form, the potential category list for each post could grow exponentially. Each person might quibble that none of the selected categories match appropriately and write in their own. If 100 people each do that, the category choices would probably take up more screen real estate than the post itself.

 - Category Limits
To prevent too many category options, I'd allow each user a limited number of checks to put in boxes. The final number of checks per poster and each post viewer would likely be different. I'd personally give the poster 5 checks to use to categorize (or 4 and a write-in), and each g+ user viewing the content would be allowed to choose 2 (or 1 and a write-in).

Additionally, I'd establish a percentage threshold that would determine if that category option appears to future users that would kick in once a certain number of g+ users had selected categories for the post - let's say 10 for this example.

Once 10 users indicate their category for the choice, any category option with 10% of the category vote would no longer be displayed as an option to future viewers of the content. The more users that categorize the content the more appropriately categorized the content will become.

 - General Category Consensus 
Over time, both posters and viewers of content will get a feel for how content tends to be categorized. If this post ends up with the categories Portland Politics, Picture, and doughnut with the most category votes, when I post a blog about obese mayors I will have a better feel for if I should categorize the post as Politics or Local Politicians. Similarly, viewers of the content will (as time passes) be more confident whether that blog post should be categorized under obesity and blog post, or humor and politicians.

Category Circles
If adopted by the majority of g+ users, the next step would be to allow us to assign categories to circles (in addition to circles as they exist now). I could add politics and humor to a circle and my stream would be filtered so that a post with one (or both) of those categories from anyone in any of my circles would appear.

Similarly, you could add categories to a 'never show' category circle and any posts with a majority category vote for that category would never show in any of your streams.

As Google+ continues to mature and more and more Google products get integrated, I'm expecting the way things work to be improved and refined - I hope Crowdsourced Categorization can be a part of that, or at least that I'm eventually better able to control how the content of my stream is filtered.

Thanks for reading,
GG

3 comments:

  1. My take on #2 is that interest circles should only be outbound circles only, not inbound. And my interest circles are actually disinterest circles. I just have 3 or 4 of them, if I discover you have a profound disinterest in say, Bollywood, I will take you out of my >Bollywood circle. But to most people, it will just be something that scrolls by: a mere disinterest that is easily ignored. That way I am not pigeonholing people into particular interests, I just try to assist in helping them not see what really annoys them. Hope that makes sense.

    ReplyDelete
  2. As for #3, I've found no benefit yet in blindly adding a whole shared circle that someone has shared. The most useful shared circle I've seen was a circle of recognizable news sources, allowing me to cherry pick the ones that looked most useful.

    ReplyDelete
  3. re: Interest Circles - by daddydave
    I wouldn't complain if your method was implemented, but I like the idea of being able to 'tune in' to various interest circles based on my mood. If I need a laugh, I sort my stream by my humor/funny/satire interest circle.

    If I'm feeling activisty, I tune in to my politics/conspiracy interest circle.

    ReplyDelete