Monday, December 26, 2011

Gumbo Shop's First Facebook Ad


Ok so there are quite a few sites with how-tos about setting up a facebook ad... this one in particular is a little dated but is extremely detailed:

http://subliminalpixels.com/facebook/how-to-set-up-and-track-facebook-ads/

Rather than go into the mechanics of setting up an ad in general, I'm just going to talk about some of the rationale behind the very first ad created for the Gumbo Shop.

I'm actually pretty glad that it took me forever to write this post, because now there's a really interesting article discussing what people look at when presented with a Business / Brand page.
http://mashable.com/2011/12/14/eyetracking-facebook-brand-pages/#38865CocaCola-Visual-Attention-Level
My main takeaway from this article is that you need pictures and you need some decently compelling posts on your wall.  My goal for the very first ad is to get people to interact with your page, regardless of how populated it is.  Therefore, they'll populate the page and later on, people who WILL make their decisions based on the how populated the page is, will see a page with Likes, Posts, and Post Responses.

So what demographic will have people who not only know of the Gumbo Shop, but are the likely people who enjoy the place and will click that "like" button without a second though?

At first I thought of the regulars - people who work in the vicinity of the Gumbo Shop and come on a weekly basis.  The absolutely easiest way, of course, is to simply have a little sign at the restaurant "Like us on Facebook!" or something silly like that.  However, in terms of filtering on Facebook, the biggest problems with that were the following:
  • Will have to get a listing of all the businesses in the area: even a decent representation would be pretty hard.  There aren't that many ginormous employers in that area where picking a couple will be good enough.
  • Not that many people actually post where they work on Facebook anyway.
The next group of people that came to mind were people in St. Louis who likely sought out places like the Gumbo Shop: people who lived in states where Cajun food is much more prevalent.  Unfortunately, Facebook doesn't have a filter for "where people used to live," so education was the next best thing.
All the LSU guys I know in STL are Lawyers...
So after doing a quick Wikipedia lookup of  colleges where Cajun food is more prevalent, I went with pretty much all the Gulf States between Texas and Florida.  As for the ad, itself, I simply wanted to convey "Hey, remember the Gumbo Shop? We have a Facebook page now."  For that, I put up a pic of Gumbo Shop's logo and took my favorite Yelp quote on the place:
So I felt that this accomplished a couple things:
  • Identify the Gumbo Shop.  Duh.
  • Show a picture of yummy Gumbo.  The people this is aimed at, I'm assuming, have already had gumbo before.
  • Point out that there's a positive review on Yelp.com by an (awesome) Lisa C.
Ok.  Now all that's left is to set a budget and maximum bid for clicks.  This means that every single time someone clicks on the ad and visits the Facebook page, I will be charged however much the bidding went down.  I don't really know how the mechanics all work, but essentially, the higher you place your maximum bid, the more likely your ad will be shown.  When there are many advertisers going after the same demographic you are, the more expensive this will likely to be.  I have recently found out that Women who are in a Relationship are quite expensive... but that's another post.
Self Explanatory.
http://sprott.physics.wisc.edu/pickover/pc/lifeexplained.jpg
At first, I set a daily budget of $3 and let Facebook have its way with my max bids.  In the end, I feel that this isn't the way to go.  I personally don't think that Facebook really does have a decent way of figuring out how to "maximize" the value of your clicks and it doesn't take any other input for your goals.  For example, your bidding and budgeting strategy can vary quite a bit if you goal was to get as many impressions (total times an ad is displayed) as possible in one month, as opposed to ensuring that you can get the best bang for the buck with at least so many impressions a day.

Also, the daily budget idea is a bit silly.  This severely limits how often your ad is going to be shown, as once that budget is hit, your ad won't be shown for the rest of the day.  Sometimes, on a Facebook heavy day (when people are home for the holidays, for example), you're going to hit that budget limit pretty quickly and then your ad will be pulled.  That's just pretty lame.  That, coupled with Facebook's relatively high automatic bidding, you're paying more for people to click on your ads while limiting how often your ad will be displayed.

So here're some of the metrics in the first month of that ad.

A "Social Impression" means that this ad was shown 1003 times where the target has a Facebook Friend who "likes" the Gumbo Shop.  Initially I was pretty satisfied with the results of this... with some fiddling we go through September's results...


WHOA what just happened... numbers are WAY higher for some crazy reason and yet my (yet undisclosed) costs were practically the same.  What the heck happened?  (aside from that fact that these screen shots are from Google Docs and Excel, respectively?)

Well... hopefully it won't take me more than a month to post the next post to explain some of that crap.

2 comments:

Sweet Kat's Creations said...

Great job Allan. This is all really interesting!

David.Capino said...

Were you manual bidding for month 2?
It would seem like you basically captured back a ton of slack that the auto-bidding was giving away. It would be funny if the auto-bid increment grain size was unusually big. Good writeup