Google’s Asymmetric Social Network and Plus Authoring Patent

Sharing is caring!

When you write a post at Google Plus, the social network allows you to link a web page and display an image from that page to the post, as well as defining who your post should be seen by. A patent filing published last week describes some of the processes behind adding those features, and provides us with screenshots of interfaces from an early version of Google Plus. It also shows off some of the thinking that might have led to the use of circles in defining the idea of an “asymmetric social network.”

The posting interface is shown in many screenshots in the patent filing:

Some interface images from the patent that is similar in many ways to Google Plus, and the ability to add links and images to posts and to decide between sending posts to subscribers or making them public.

I find it interesting about reading through patents is that they introduce us to the vocabulary used by the inventors and concepts that they might have considered that we might not even be conscious of. Chances are, you may have never thought of or referred to Facebook or LinkedIn, or Myspace as an asymmetrical or bidirectional social network, or Twitter or Google Plus as an asymmetrical social network. The patent’s description starts off with a detailed definition of what an asymmetric social network is.

An asymmetric social network is a social network where a first member’s relationship to a second member is not necessarily the same as the second member’s relationship to the first member. Since the character of the social interaction between members in a member network can be defined by the nature of the relationship between those members, a first member in an asymmetric social network may interact with a second member in ways that differ from the social interaction provided for the second member to interact with the first member.

With the circles in Google Plus, the social network allows users to have several different social interactions, based upon whom they are attempting to interact with.

For instance, if you create a post and send it to one person, it’s a little like sending an email. If you create a post and send it to one or more of the Circles that you’ve created, it’s a little more like a newsletter, or a tweet that goes into the streams of people who have subscribed to you. (note: the patent doesn’t refer to “circles” anywhere though it describes a difference between sending something to subscribers or sending the post to your profile). If you create a post and post it to your profile, it’s more like a status update on a LinkedIn or Myspace.

The patent application focuses primarily upon authoring tools for posts where you might link to a page, choose an image from that page, or a thumbnail of the page to include with your post.

The patent application is:

Assisting The Authoring Of Posts To An Asymmetric Social Network
Invented by Samuel Shoji Fukujima Goto, Joseph Rideout, Braden F. Kowitz, and Todd Jackson
US Patent Application 20110197146
Published August 11, 2011
Filed: February 8, 2010


Methods, systems, and apparatus, including computer programs encoded on a computer storage medium, assist the authoring of posts to an asymmetric social network. In one aspect, a method performed by a system of one or more data processing devices includes receiving, at the system, an identification of an electronic document that is available on the Internet, the system identifying image content in the electronic document, the system filtering the identified image content, the system triggering presentation of the filtered image content to an author of a post to an asymmetric social network, the system receiving a selection of a first image from amongst the presented image content, and the system adding the first image to a post to the asymmetric social network.

Some of the features involved in the process of assisting authors of posts could include:

1) Filtering images, so that advertisement images aren’t shown.
2) Filtering image content not to include icons or decorative elements.
3) Possibly providing a screenshot of the page as an image to include with the post.

The patent goes into much more detail about the posting features and how the system might distinguish between images that are either advertisement or decorative images or neither, and is worth skimming through. For example, one of the ways that an advertisement image might be identified because it might be 3 times longer or taller than it is high or wide, indicating a banner advertisement of some type. An image that exists for decoration purposes or as an icon might be identified by being less than 30 pixels on any one side.

This patent doesn’t include the concept of Circles at this point, but it does include a posting selection feature that allows authors to decide whether a post is public or private. A private post would only go to the subscribers for an author and not be displayed on the author’s profile or subscribers’ profiles.


When I talk or write about creating titles for pages to be put in a title element on a page, one of the things I like to stress is that a page title is most commonly seen out of the context of the page. The words that you put in between an opening <title> and a closing </title> don’t appear upon the page itself, but rather at the top of the window that the web page is displayed in. If someone bookmarks that page, the title might be used as the name of the bookmark. A search engine usually (but not always) uses the title of the page as a title (and link) displayed in search results above the snippet for that page. It’s also not uncommon for people who might link to that page to choose to use the title for the page as the anchor text of that link.

When someone chooses to link to one of your pages at Google Plus, the process described in this patent filing allows them to include a choice of some images from your page, filtering out other images as described in this patent filing. I think that’s something to consider when you’re deciding upon an image or images to include with your article or post or products that might be linked to from the social network. Will thumbnails of one or more of those images make it more likely that someone may click through from Google plus to read your page? How well do those “out of context” thumbnails convey the ideas or meaning that might be found on your page? How effectively do they work out of context?

If you’re using Google Plus, have you experimented to see what images show up as choices for people to include when they post a link to your page? Facebook also provides a similar choice of images take from a page when you post a link to that page.

I’m not completely satisfied with the choices of images that show up in either Facebook or Google Plus with my pages because the books are shown in the left column of my page sometimes are shown as pictures to include with a link in a post, and I’m thinking of making some changes because of that.

Sharing is caring!

10 thoughts on “Google’s Asymmetric Social Network and Plus Authoring Patent”

  1. Hindsight is a funny thing. Seems to me that an asymmetric social network is the intuitive way we would prefer to share things online.

  2. Hi Bill,

    interesting patent – the Asymmetrical and Symmetrical (bidirectional) titles are a useful description. I do think that both types of network are useful though, in a kind of broadcast versus targeting messaging kind of way. (If I’m out for dinner and we’re having drinks afterwards – i might want to invite a few folks specifically for dinner, but then broadcast to everyone where we’re having drinks… Google Circles & Twitter would be my preferences.

    The imagine definition is much more interesting though. From memory the standard sizes are 468 x 60, 728 x 90, 300 x 250, and 120 x 600. I’m guessing the way in which the banner ads are served will mean that they won’t appear in the image selection option for Google+ though, but it’s worth thinking about how images are formatted to make the most of it. For instance I’m currently using a logo which has a much larger border than it should do to cheat some page formatting – in Google+ it looks horrible and when i find time i’ll promise I’ll tweak it….

    I had a quick go and choosing to share this page with LinkedIn and I’m sorry to report back that the images options available weren’t overly brilliant. Maybe time for a small SEO By The Sea logo? 🙂

    On the non-social booking marking side of thing – it’s worth remembering that the favicon.ico image appears when someone bookmarks a page. Perhaps an even smaller SEO By The Sea logo too 🙂

  3. Bill,

    Great research as always, thanks for this. It’s interesting to think about an asymmetrical social network. I have always felt that the best network is one that more accurately models desired human relationship. In life, we can talk to someone who doesn’t want to talk to us, but they generally have the choice of how much they want to listen to. I suspect that most people appreciate this fact about Google+.

    I was intrigued by the way that Google might differential an ad from a regular image – I can’t help but wonder where else this or similar methods may be used. Could Panda be using this as one of the signals to determine ad-to-content ratio?

  4. Hi Gyi,

    I think it’s fascinating how social networks are evolving, and yes, hindsight is 20/20.

    An asymmetric social network seems more like how social networks actually happen outside of the Web.

  5. Hi Tom,

    I agree, having both types of networks available to us isn’t a bad thing at all. It provides us with a chance to use those networks in different ways. Even the constraint on Twitter of 140 characters isn’t a problem, but rather a strength of the network.

    I’ve thought about how the images from this site are being used to provide choices to include in social networks, and would have hoped that the people providing those would be smart enough to distinquish between images that might appear in a post, and images that might show up in other content areas, like a sidebar. They haven’t gotten to that stage yet, though it’s something that I think they could and should be able to do. I may take some steps to make it more likely that better images appear in those situations, but it may take some experimentation to learn more about how the choices of images are made.

    I do have a favicon which I like that seems to work well, and an apple-touch-icon and an apple-touch-icon-precomposed icon as well.

  6. Hi Joe,

    One of the things that I like about the Web is that there’s so much opportunity for experimentation, and for people to try different things when they develop sites and social networks, and I find that there are things that I like about both symmetric and asymmetric social networks. I also think it’s possible that some people might like one type more than the other, so having a choice is great.

    The description of how Google might distinquish ads from other images probably just brushes at the surface. I suspect that Google could and may also look at if there is a link associated with an image, and check to see if it’s from an advertising domain they might be aware of, or even follow the link to see if it leads to something like a landing page for a particular service or product.

    Chances are that Panda may look at similar types of features that could be associated with things that it finds on Web pages, and as a machine learning process may even come up with new rules regarding those features and either suggest their use to the people running the upgrades, or in some cases even apply them without any human intervention.

    For example, Panda may work to identify everything that it can in its seed set of known sites about advertisements that show up on pages, and may determine that ads often tend to be certain sizes and have certain ratios of dimensions (you often don’t see people posting images on their pages that have the same dimensions of banner ads). Panda may also then look to see if the image has a link, where the link points to, and if the page it points to is a landing page.

    It may also crawl the page a number of times within a short period of time (a minute or two apart for instance), and check to see if the images it thinks may be ads change on each visit, and whether they point to a landing page each time. The process could involve a good number of steps to determine whether something is an ad.

    Once Panda has a good grasp on what might be an ad, and what might not be, it could then start examining where ads tend to show up on pages, how many ads appear, what ratio of ad to content is on pages, and more. The machine learning process would build upon features associated with what it finds on pages, and identify ways to score pages based upon those features.

  7. I had never considered the structure of social networks, so the concept of asymmetrical vs. symmetrical was really interesting. Clearly both types have value in different ways, but I’m sure it can also depend on how they are employed by the user.

    Reading your observations about image selection spurred me to play with what are my choices when attempting to share a post from one of the blogs to which I contribute. Needless to say, the choices were really underwhelming. Any idea on how you’re going to experiment with your images?

  8. Hi Bill,

    I hadn’t given too much thought to the structure of social networks from that perspective either.

    As for image selection experiments, I may move some images around that appear in my sidebar, maybe introduce some other images into those sidebars and try different sized images in posts, and see if some other ideas spring to mind in the future.

  9. Pingback: Matthew Effect: The Motivations of Google+

Comments are closed.