Buying Facebook fans – apparently it works…

Few months ago I added in the index a number of websites that advertize on the possibility of 'Buying Facebook fans'.

why?
Because I am monitoring Facebook Fans, so I was asking myself if the game could be rigged...
Are all fanbase growth "natural", or are they "pumped" somehow?

Turns out one player has gotten a huge Facebook Fans growth in the last two months:

http://www.buyrealfacebookfans.com/

Check out the chart:


More than 20,000 fans gained in 1 month! This is huge. Better than AirBnb, one of the hottest startup right now.

Better than Groupon, TripAdvisor, and many other big players.

For more details check the graph by visiting our site at http://www.99like.com and search for 'buy facebook fans'.

On their site they propose "1000 Facebook fans for only $57 - limited time offer".
Which would bring the cost of a "real fan" at $0.057.

So it would seem that it works, or at least they themselves managed to up their Facebook fan base by a huge amount in a very short time frame, which tend to show that it worked..

As a side note, one could say that a huge advertizing campaign is a way to "buy" Facebook Fans, although not in a direct way.
I keep seeing ads for AirBnb.com, and it reminds me of Groupon a few months ago. Classic rush to be the biggest player...

AirBnB doubles its Facebook fanbase in 3 months

Tracking AirBnB facebook fanbase is interesting. Probably one of the fastest growth right now.
They went from 42,000 fans in end of June 2011 to 75,000 fans in beginning October 2011.
That's a 80% Facebook fanbase growth in 3 months.

In the mean time, Tripadvisor went from 153,000 to 164,000 fans, a 7% growth.

At this rythm AirBnb fanbase should overtake TripAdvisor fanbase by March 2012.

Lessons learned from 1 million Facebook API calls

The goal of 99like.com is to track Facebook link statistics (number of times a link was shared, liked and commented on) for as many websites as possible, and present charts with this data.

We started by indexing more than 100,000 URLs using Alexa list of top sites by traffic, as well as CrunchBase data. We then implemented a script that sent requests to the Facebook API asking for the current number of share, like and comments for each URL.

Over a period of one month, working off and on on our script, we have been sending over 1 million API calls to Facebook, and stored the data. Here are some lessons we learned in the process. Perhaps this will help others.

————————

Our first version of our script was using a single-query API calls (basically asking for the link stats for a single URL). But we could not get enough data to track all URLs every day.

Two weeks ago we started batching multiple queries per API calls. This way we were able to increase the rate of queries per day.

Lesson #1: Facebook multiple queries are great to increase the rate of queries per day.

Unfortunately we noticed that the data received from Facebook was completely out of whack: our top website by number of total share + like + comment was suddenly tak.ru with more than 18 millions, and google.com and facebook.com were getting ridiculously low numbers. Obviously something was wrong with the data.

After looking at the responses from Facebook we could see that Facebook API was actually sending a response with data (sometimes 0, sometimes just seemingly random numbers).

We decided to test different parameters and verify the consistency/integrity of Facebook responses by comparing with single-query calls for sampled data. The results were interesting. It turns out that Facebook responses for multiqueries with more than 10 queries start to have errors.

Here is a chart of Facebook API responses consistency rate vs the number of queries made per API call:

Background on the test done:

-          Total of 1400 API calls were made

-          Total of 16,200 queries were made

-          10% of the data was checked for consistency (multiquery vs single-query per API call), which means 1,620 checks were done

-          Tests were done across 2 hours, rotating several times the number of queries to get a fair comparison

The good news is:

Lesson #3: multiqueries return reliable data for 10 queries (or less) per API call

We could not find this limitation in Facebook documentation. There is some documentation on the multiqueries here:

http://developers.facebook.com/docs/guides/performance/

We would be interested to hear from others who have been using the Facebook API.

Facebook users top 10 popular sites

First surprise: the #1 is not Facebook, nor Google or Twitter. It is a emoticon service: bandoo.com.

#2 is causes.com

Then come the big boys: Youtube, Facebook, Twitter as #6, Google as #9.

You can check anytime the top sites at http://www.99like.com

Introducing 99like

Welcome to the 99like blog. This is your first post.

99like is a tool to search which websites have been shared the most by Facebook users.

We pull the data from Facebook, using the “share” stats. We then store the information together with the web address and a description of the site (using the html ‘Description’, ‘Title’ and ‘keywords’ tags).

The user can easily (hopefully) search through the index of sites in our database to see which sites match the given keyword and have the most “fans” (“fans” being counted as having shared the link with their friends).

We hope the tool will be of interest to others. Let us know what you think.