The Solution: Wrap the likebox in an extra div (with a border color set to your pages background color, and position the wrapper to overlap the likebox). Code below.

The HTML

<div id="likebox-frame">
     <iframe src="http://www.facebook.com/plugins/likebox.php?id=XXXXXX&amp;width=900&amp;connections=16&amp;stream=false&amp;header=false&amp;height=190" scrolling="no" frameborder="0" style="z-index:9;border:none !important; overflow:hidden; width:900px; height:190px;" allowtransparency="false"></iframe>
    </div>

The CSS

#likebox-frame{
 border:2px solid #F0F9F4;
 width:894px;
 z-index:10;
 height: 156px;
 overflow: hidden;
 position:relative;
 top:-1px;
 left:-2px;
}
#likebox-frame iframe{
  left:-5px;
  overflow:hidden;
  position:relative;
  top:-1px;
}

“The conversion rate is similar to paid search,” he says of Criteo’s retargeted display ads. Retargeted ads are typically clicked on four times more than a traditional run-of-site banner ad, he estimates.

This is all well and good, but when I going to be able to re-target the visitors to my competition?

e.g If I am Target.com, I can re-target visitors across the web who have visited BestBuy.com and offer them a competing offer on a product they have viewed at BestBuy.com or a coupon for $$ off their order.

Best Buy will not be pleased that the data they are sharing to power their own re-targeting campaign is being used against them, but the end user benefits with a better price, and net sales are increased as a result, creating a win for both the merchant (who makes the sale, the user and the advertising network).

So, in two days, that blog post generated 535 views of the Amazon page and 40 purchases. The affiliate fees associated with those 40 purchases add up to $6.50.

But those 535 views are pretty valuable. Those 535 clicks translated into a total of 118 orders in the past two days, including a Kindle. The total affiliate fees associated with those 535 clicks were $25.20.

But even including all the commerce that was generated from that link, that $25.20 is a cost per click of roughly 5 cents. I think that’s low for a bunch of reasons.

The post inspired some interesting conversation around how in a perfect world, each player in the chain of purchasing-intent generating events e.g blogs, social media pundits all the way to affiliate marketers should be compensated in some way, as opposed to the way affiliate links function now, where only the last attribution is compensated.

Inspired by the post, I decided to construct my own experiment in recommendation commerce to see if I could influence anyone to buy a product through an affiliate link, answering questions on Aardvark. I decided to only present an affiliate link as an answer when it was relevant to the question being asked, and after a series of un-monetizable questions, I was presented with the following.

The book I recommended was Mental Models: Aligning Design Strategy with Human Behavior (notice affiliate code in link again).

Notice how my answer was hyper-relevant to the question being asked of me. I would guess that in any other scenario, my affiliate link given as an answer would not have been treated with such gratitude. Also of note is the added benefit of my referral, Aaron purchased another two books, sweetening my take by $1.98.

Does this form of monetization scale? Maybe. Does it make sense at such a small one-to-one level? Probably not.

Therefore, I think it remains to be seen, which model (if any) is waiting to be unearthed to monetize the value that is generated in the form of social reviews and recommendations, but it is an interesting space, and one that I will be keeping my eye on.

A scary thought experiment…

Wikipedia is entirely supported by donations. So why isn’t it a good model for everyone else? Last year they raised $6.2 million from 125,000 donors. While that’s a lot of money and more than covers their operating budget for the year, Wikipedia gets 190 million visitors a day. Think about that for a second, one of the most import information sources in human history with daily traffic bigger than most nations only got 125,000 donors? How much is one of Wikipedia’s 5 billion monthly visitors worth when it comes time to donate? $.00124. That’s a fraction of a penny. If you’re providing a human experience enriching service on par with Wikipedia, multiply your total audience size by $.00124 to figure out how much you can expect to make from donations. That comes to about 8 million people to stay above the poverty level ($10,000). Wikipedia’s efforts worked well enough, but it’s a frightening proposition for anyone who doesn’t one of the most trafficked websites on the planet.

The New Paradigm: Put end-users first and profits will follow. (Google).

Much can be said about the corporate cultures (particularly Google’s) which have in part enabled this brand of user-centric thinking, but that is a topic for another discussion.

Time and time again Google has put the best interests of their users AHEAD of their own share-holders and short-term bottom line.  This isn’t to say that they are profit agnostic in the slightest, they happen to make very healthy profits, rather they believe that doing right by their customers will earn them more loyalty (read: money) in the long run.

To phrase it another way: Google is not successful in-spite of putting users interests first, to the contrary, they are successful BECAUSE they put users’ interests first.

Ok, so what are some examples of this seemingly “altruistic” behavior?

Example 1 (2007): Google get’s serious about ad quality by implementing a “quality score” for all of it’s search ads.

Example 2 (2010):  Google’s recent entrance into the mobile phone market, providing more openness and choice in a stagnant and typically user-unfriendly industry.

What remains to be seen is how long this type of user-centric behavior can last.  The danger is that with time, pressure to increase profit will lead to cost-cutting measures that affect their products, but in the mean-time it seems like Google and it’s management are still eating the “don’t be evil” dogfood.

You can paste the following code into you’re browser toolbar after clicking through an ad to you’re site to see if (any) referrer was passed. If it’s empty than Google Analytics will count it towards direct traffic if you don’t pass any campaign tracking parameters.

javascript:alert(document.referrer)

How do I fix this?

1. Realize that a chunk of the traffic that shows up in your “direct” bucket ~(10-15%) could be attributed to some other source e.g (ppc,display,link). This could be from click-through traffic or type-in traffic (where people see you’re ad and type in you’re url instead of clicking on it.) The latter case is essentially free advertising and while hard to measure directly should be considered into the ROI calculation for advertising. This is why it’s always important to display a url in you’re advertisement.

2. Always use tracking parameters in your urls when possible. http://www.google.com/support/googleanalytics

Case and point: Grilled Cheese Sandwiches have a whopping 212,686 fans who are willing to self-identify as Grilled Cheese Sandwich lovers on Facebook!

If most people are not willing to “self-identify” as being a fan or user of your product than you are not alone. It is the rare exception where a product is so fantastic or speaks to a certain demographic in such a way that the users feel compelled to broadcast their affinity to others in their social spheres.

A great example of a product that people self-identify with is the Apple iPod, not only is it an incredibly sexy product but it is made by a company with tremendous customer affinity and good will.

The iPod has 341,820 fans on Facebook.

By comparison Viagra has a flaccid (yes i said flaccid) 368 fans on Facebook.

While I picked an especially un-sexy (or sexy, depending on how you look at it) product to make the point, the same is true of countless other products and services. If people are not willing to self-identify as users of your product than you are going to have a tougher time promoting your product organically.

But don’t be dismayed if you are selling an unsexy product. The silver lining here of course is that there is usually a negative correlation between the propensity of people to self-identify with your product and the actual profitability or value being delivered. Products like the iPod, again, being the rare exception.

While few are willing to admit their affiliation with online dating sites like Match.com, 789 fans on Facebook.

A whopping 1,255,505 are fans of Mojitos!

I’m willing to guess that there is a strong correlation between people who have tried Match.com and are ALSO fans of Mojitos….but these numbers shed some light on the discrepancy in the willingness of people to self-identify with each product.

Note: Some of these methods will work on a shared host without root access, but ssh is required for some methods, like installing memcached.

Step 1: Offload Images to Amazon S3

Images generally make up a large percentage of the total page load (up to 50%, see below). By using Amazon S3 as a CDN, you can save on bandwidth costs with your host and speed up your page load times significantly.

Graph taken from the YSlow firefox plugin, a great tool to help diagnose page load performance issues.

Costs for S3 are reasonable, but can add up if you are hosting large files. Below is a report of costs for the first 10 hours of a blog that I run. Most requests are for a single image (2k in size) loaded externally.

Tantan S3 is a great plugin for wordpress that links your Amazon S3 account to your wordpress blog. You can configure it such that your media uploads are sent to your S3 “bucket” on upload.

Another useful tool is S3Fox, a plugin for firefox that you can use to manage and sync files in your S3 account through your browser. You will need your S3 security credentials to set it up.

Step 2: Apache Tuning

Edit your .htaccess file to set expire headers on static files

#Expire Header
<FilesMatch "\.(ico|jpg|jpeg|png|gif|js|css|swf)$">
ExpiresDefault "access plus 2 hours"
FileETag none
</FilesMatch>

Step 3: Configure some kind of caching

For many smaller wordpress blogs WP Super Cache and other MySQL performance tuning measures will suffice, but if you still aren’t seeing much improvement, you may want to consider a more heavy duty option.

Installing Memcached on a MediaTemple DV (3.5) (requires root access)

In conjunction with Batcache Manager, you should see a real boost in performance.

Step 4: Sit back, relax and enjoy your speedy new blog

Further Reading:

• Using Amazon S3 for image hosting with a WordPress blog
• 4+1 Ways To Speed Up WordPress With Caching
• High Traffic Tips For WordPress