Tagged: Facebook RSS Toggle Comment Threads | Keyboard Shortcuts

  • Jasika Bawa 11:13 pm on November 22, 2011 Permalink | Reply
    Tags: Facebook,   

    Facebook: 6 degrees? Try 4.74 

    In a recent blog post by the Facebook Data Team, we can see how a joint study has shown that any single person is only 4.74 steps away from being introduced to any other person, instead of the commonly believed upon 6 steps. Some of the data brought forward by the study is pretty crazy, for example:

    While 99.6% of all pairs of users are connected by paths with 5 degrees (6 hops), 92% are connected by only four degrees (5 hops). And as Facebook has grown over the years, representing an ever larger fraction of the global population, it has become steadily more connected. The average distance in 2008 was 5.28 hops, while now it is 4.74.

    Thus, when considering even the most distant Facebook user in the Siberian tundra or the Peruvian rainforest, a friend of your friend probably knows a friend of their friend.

    The study also found that people are much more closely connected to individuals in their own country.

    Although this new “law” of separation applies only to Facebook itself, the number of Facebook’s active users is reportedly around 10% of the total human population today–do you think that is enough for the new law to be extrapolated to apply to all people in the world? Or do you think that the data just goes to show that “X knows Y” now means less than it did before?

    How many friends?

    The study also showed that Facebook's cumulative degree distribution is not nearly as skewed as earlier studies of social networks have suggested

    VN:F [1.9.22_1171]
    Rating: 5.0/5 (5 votes cast)
     

    • Mung Chiang 11:44 pm on November 22, 2011 Permalink | Reply

      This is interesting:

      Small world: whether it’s 6 or 5 degrees of separation, the real test is local information based discovery of the path, not just the existence of short paths. This is a common misunderstanding. So Facebook needs to run experiment of message passing, not just analyze topology.

      Scale free: strictly speaking it’s only about power law at the tail of the distribution, not the entire distribution. So it may still hold. The more relevant question is Dunbar’s number that truncates the tail of power law distribution: how can someone have 1000 or 5000 friends? That must be a redefinition of ‘friend’. A better definition of friend might be: if you call this person and say I am XXX, will she pause a while before realizing who you are?

      VA:F [1.9.22_1171]
      Rating: 0.0/5 (0 votes cast)

  • sna 6:07 pm on November 17, 2011 Permalink | Reply
    Tags: Facebook, ,   

    I am sure many of you have heard of the McRib. It comes around once a year for a limited time and seems to have a cult-like following. How does this mystery meat sandwich cause so much hype? McDonalds has learned to use social networks. In their 2011 commercial, McDonalds shows people texting their friends about the return of the McRib and advertises, “McRib is back nationwide, so get the word out because that saucy goodness is only here for a limited time.”

    If you go onto youtube there are thousands of videos concerning the McRib. Even one calling “How to Clone a McRib Sandwich” which details how to make your own McRib once the sandwich has gone (it has nearly 100 thousand views). The McRib has also made it to other popular sites like Facebook and Twitter. On Facebook there is a game called “The Quest for the Golden McRib” as well as many fan pages. You can follow the McRib happenings on Twitter or go to the McRib locator website to find McDonalds locations with confirmed McRib sightings.

    Whether you are a fan of the McRib yourself or it makes you want to gag, you cannot deny that McDonalds has done something right with this sandwich. In 2010 when the McRib returned, McDonalds saw a 4.8% increase in sales. This type of increase was made possible by the buzz that McDonalds was able to create in social networks.

    VN:F [1.9.22_1171]
    Rating: 5.0/5 (2 votes cast)
     

  • network20q 2:52 am on October 31, 2011 Permalink | Reply
    Tags: Facebook, , Online Influence, ,   

    How influential are you among your peers? Let’s ask Klout 

    With the ever-expanding presence of social networks in our daily life, users are becoming more and more conscious about increasing the extent of their social influence. Many companies now want their marketing, public relations, and publicity personnel to provide their Klout score at the time of recruitment. The Klout scores are a measure of an individual’s ability to drive other members of his or her social network to action, typically through retweeting content, replying, “liking”, and sharing the posted content of Twitter, Facebook, Google+ etc. Every time a person creates content or engages in conversation with others, they are in the act of socially influencing these peers. And it is this influence that Klout scores on a scale of 0 (least influential) to 100 (most influential). For example, Justin Beiber has a Klout score of 100, while Nature, one of the most prestigious scientific magazines, has a Klout score of 64 (no surprises there, science never stood a chance against Justin Bieber! :) ).

    The Klout Score uses data from social networks (although they don’t disclose the details, except that they consider some 35 metrics to determine the score) in order to measure:

    True Reach: How many people you influence

    Amplification: How much you influence them

    Network Impact: The influence of your network

    Many people have evaluated the benefits of Klout while others have argued against putting too much weightage on social influence metrics. Arguments made by holders of these two opposing views can be found at the following sites:

    1. http://econsultancy.com/us/blog/8026-klout-matters-but-not-for-business

    2. http://www.businessesgrow.com/2011/09/12/why-klout-matters-a-lot/

    3. http://tinyurl.com/kloutbenefits

    While the battle over usefulness rages on, Klout has been working hard to improve their scoring algorithm. In the past, Klout’s scoring algorithm had been questioned by many, see “The Trouble with Klout” for more- http://gillin.com/blog/2011/09/the-trouble-with-klout/ for details.

    So in the latest tweak to their algorithm, Klout is now evaluating an individual’s influence score by placing more weight on how many other influential people are following or getting influenced by this individual. The move has irked many, especially those who use Klout scores in their professional life. But whether the latest move was effective or not is yet to be seen. And this is where you can come in and explore Klout with a healthy dose of scientific skepticism, logical reasoning, and penchant for mischief to find out if Klout really works. It will be fun for each one of you to start looking after your Klout score and figuring out ways to improve it, perhaps by even tricking their algorithm ;-) .

    To get started with Klout, simply log in at http://klout.com/corp/kscore with your Twitter or Facebook account.

    ~Soumya

    Follow @SamSen4

    VN:F [1.9.22_1171]
    Rating: 5.0/5 (3 votes cast)
     

  • class 9:09 pm on October 2, 2011 Permalink | Reply
    Tags: CTR, Facebook, Nielson   

    Facebook Looking Beyond CTR 

    In a recent interview with Mashable, Facebook’s Brad Smallwood explained why he believes that the measurement of click-through rate is becoming increasingly irrelevant to most online advertisers. He elaborates by saying that as the Internet continues to grow and innovate, there is a lot that happens “beyond the click” with respect to brand advertising. What I found interesting was his mention of a Nielsen study that says that there is in fact a negative correlation between click-through rate and brand awareness.

    There are several factors that seem to be more important than a mere count of the number of times a link has been clicked on, including who is reading the ad, whether the ad made an impression on them and whether the click ultimately leads to a sale or profit. Smallwood says in the interview that Facebook is currently working with Nielson to move toward a more sophisticated measurement of click-through using a “gross ratings points” system.

    I would tend to agree that today, click-through rate seems to be an outdated measure of performance of an ad online. With the evolving state of the Internet, there are plenty of ads that make an impression on me each day without my having to click on them. However, the interview brings up a good point in that CTR, while primitive, is at least a form of measurement, and that it might be difficult for marketers to know whether their ad is effective without monitoring a measurement like this. Do you think that Facebook’s attempt to move toward a better form of measurement than CTR will be successful and understood by marketers? Or do you think that CTR is a good enough measurement of the performance of an ad online today?

    –Jasika Bawa

    VN:F [1.9.22_1171]
    Rating: 5.0/5 (4 votes cast)
     

    • edlee 2:58 pm on October 5, 2011 Permalink | Reply

      I’m curious to know what the real proportion of seen ads and clicked on ads actually lead to a purchase and which subsets of the web population correspond to certain success rates. In fact, sometimes I am surprised by the fact that a company like Google can make such a huge profit out of such a simple scheme. I use a lot of Google services, but I’ve never bought anything from Google and my CTR is probably well below 1/10000 if even that.

      Jasika makes a good point in that CTR is an “primitive” measure of judging how much to pay for an ad. CTR is an unfounded hypothesis about human behavior, and I would assume that the reason it was adopted was that it makes intuitive sense that CTR should correspond in some direct way with purchases and primarily because it is easy to measure. However, it could use vast improvement on the side of psychological analysis of people’s actions and beliefs that I think could lead to a disruptive business model.

      Here’s a classical example: classical economics or neoclassical economics. The presumptions about human behavior in these systems is completely wrong. They may lead to nicer equations and behavior that is easier to predict, but they in no way explain true market behavior. When we collect real data about how people respond to market movements, we see that psychological motives such as herding are incredibly important and define the landscape for the most significant actions.

      CTR, I think, is at a similar stage. There are most definitely some simple algorithms that can vastly improve the effectiveness of ads today. For example, one big reason I ignore all Gmail ads that I see is because they are somewhat related to what my emails are about, but they are untrustworthy. Sometimes, they suggest websites that seem like scams and by virtue of association, I am disinclined to select those links. If these ads were vetted through some process, the chances that I would click on them would increase sharply. Another example is the wording of these text ads. Depending on the demographic, certain key words will sound false, unappetizing or simply unworthy of attention. With the information that Google has today, I think personalized ads in the sense that they have researched what kinds of appeal or appropriate for certain populations could also vastly improve CTR and serious attention.

      I expect there is a reasonably simple algorithm that can account for these differences and learn how to tailor ads to an audience. Perhaps that will generate its own revolution.

      -Eddie

      VN:F [1.9.22_1171]
      Rating: 0.0/5 (0 votes cast)

    • Marphisa 7:09 pm on March 14, 2013 Permalink | Reply

      I got what you mean, thanks for putting up. Woh I am delighted to find this website through google.

      VA:F [1.9.22_1171]
      Rating: 0.0/5 (0 votes cast)

c
compose new post
j
next post/next comment
k
previous post/previous comment
r
reply
e
edit
o
show/hide comments
t
go to top
l
go to login
h
show/hide help
shift + esc
cancel