Experimenting with Facebook’s advertisement system [part 1 – Basic Ads]
|| 5/15/2008 || 2:47 pm || Comments Off on Experimenting with Facebook’s advertisement system [part 1 – Basic Ads] || ||

Facebook ad showing a link to my Geospatial Art on-line store

Screengrab from Facebook showing my advertisement being served
___names whited out___

On Monday evening I noticed that Facebook was promoting it’s internal advertisement system on my notification panel. As a curious person I decided to check it out and see why I receive so many stupid ads for dating websites. I remember reading a long time ago that at the height of MySpace’s popularity their servers had trouble keeping up with the demand of the user traffic, and while those times have passed, Facebook seems like a clear alternative for social marketing.

So what should I link to? This blog or my on-line store? I chose the on-line store because if I am paying money for a service I might as well get some money back. I also don’t actively promote the store because it’s become too cumbersome to manage and I haven’t added any new maps in ages.

But what to specifically link to? This is where is gets fun. The Facebook advertisement system allows for hyperlocal advertisements based on multiple user defined categories. From the very broad to the very specific, the scalable nature of the advertisement system is very impressive. For example, I can make an ad for all of Facebook (22 million people) or I can make ads for 27-year-old single women in Washington, DC with college degrees that are interested in men (under 500 women). For the first series of tests I used a bunch of different criteria for the advertisements.

“Fine Map Art For Your Walls” was the first ad that I created as a nationwide “campaign,” and it linked to my Geospatial Art on-line store front page.

“D.C. Map Art For Your Walls” was the second ad that was created for all men & women in Washington, DC, Maryland, and Viriginia and linked directly to my Washington, DC page on my Geospatial Art on-line store.

“American Decorative Maps” was targeted at women in 24 and up in Washington, DC, New York City, Los Angeles, Boston, Philadelphia, and Chicago and linked to the Diamond Quilt Projection page on my Geospatial Art on-line store.

The first two deleted ads were the earliest ads that were text-based and without an eye-catching graphic. I deleted these because I was bored with the lack of traffic and because I was being impatient.

“A Unique Map of Dallas” was sent to all men & women in Dallas, Texas, and linked to Dallas Quilt #2

“A Unique Map of Saint Louis” was sent to all men & women in Saint Louis, Missouri and linked to Saint Louis Arch Quilt #3

“A Unique Map of Harvard” was sent to all women in Boston and linked to Harvard Quilt #2

The summary really says it all. I received 173,750 impressions, which is how many times the ad was served and people could have looked at the advertisement. A measly 91 clicks was generated from all of those advertisements, yet I could have paid for the services as Cost-Per-Impressions route, which you pay per 1,000 impressions, and would have skewed the numbers differently. I could have also put a higher bid in place to show my advertisement more often or keep my bids arbitrarily low to ensure that some ads might run and I don’t spend too much money each day. However doing some quick math:

173,750 divided by 91 clicks equals about 1,909 or 1 out every 1,909 impressions resulted in a click.

All together the average cost per click was 44 cents, which is more of a result of having higher bids on my advertisements and not cost-per-impression.

But after spending $40 on 91 visitors I must ask: was it worth it?

Not really.

The day before I started the campaign I received about 200 visitors free of charge thanks to a new incoming link from a popular Buenos Aires-based art magazine’s website, genuine curiosity, and search engines. However, people visiting my website are not being directed toward my on-line store immediately. So in that respect I see the importance to pushing people to my store for sales and if they want more information they’ll visit my website. If one person of 91 people would have bought a decent sized map from the on-line store the whole experiment would have been paid for. Alas no one did, and I’ll write it off as a learning experience.

Related Facebook Entries: