Experimenting with Facebook Ads
My experience with a (very small) test of Facebook's self service ad system.
I've been wanting to experiment with Facebook's self service ad platform for a while now. With that in mind, I thought I might as well finally do it.
A few facts about what I'm doing.
- Budget: £12 for a weekend period (this budget, if used quickly, will be increased)
- Advertising my services as a web designer and developer
- This site is the linked site
- I'm taking advantage of their (very customisable) targetting system and only targetting people who live within an approximate 20 mile radius of my hometown (more on this later).
So, armed with my tiny experiment budget, let's see what happened!
So, what's the ad?
Well, here's a screenshot of the ad:
Facebook lets you choose between an external link (which I chose to drive traffic to this website), or to an internal Facebook Pages, Events or App. Pages, Events and Apps have the added bonus of 'Social Reach', whereby Facebook includes the names of a user's friends who have liked or RSVP'd (etc.) with the Page, Event or App you're advertising. For obvious reasons this is not technically possible with external links.
As you can see, I've not used a logo, but instead a screenshot of the homepage. This is probably a mistake, but I don't really have a logo which I can use. Sure, I could use related imagery to do with what I'm advertising, but this is fairly abstract and unfocussed when it comes to what people imagine when they think of web design or development. As this is only a small test, I didn't put as much effort into this area of the advertisement. It's obvious, though, that an eye catching visual element to the ad is preferential and probably yields higher returns than a poor visual element.
The text element is interesting. There are obviously rules as to what you can and can't say in the ads, and the style you can use. You can't, for example, use too many exclamation or question marks (that's not to say that you can't use them at all). Similarly, you can't use a lot of capital letters (an ad can't be in all caps, for example, but business names such as 'BBC' or 'CNN' are acceptable). There are also character limits in place for the title text (in my case 'Need a Website?') and the body text.
Targetting is where Facebook seems to really shine. Because of the scarily enormous amount of data about where people live, how old they are, whether they're married, single, etc., you can target a very specific set of people.
I only wanted to target people in my immediate area as part of this test. You can target by country, by region (i.e. England, Scotland, Wales), or by town/city. I opted for town/city. You list towns and cities which you wish to target, and can limit it to exactly that town/city, or include a radius of varying degrees around them (16km, 40km, 80km). You can list seemingly as many as you want, and their database of placenames is unsurprisingly complete.
I also set the age of my target demographic to be above 20 years old, either sex. You can also set by education level, or a specific age range, or relationship status, or gender, or language.
Finally before submitting the ad for review, you need to set a budget and your CPC (these values can be changed before or during the campaign).
How did it go down?
Once you've created your ad, uploaded the visuals, targetted your potential clients, set a price and budget, and decided on a schedule, you have to submit the ad for review.
The review process seems to be at least partly automated. From creation to approval, the review process took approximately 1 hour, and I have to imagine that a site like Facebook processes a lot of self service ads per day, so this seemed relatively quick.
The ad dashboard is very easy to use. Ads are grouped by campaign, which you can drill down into. You get the Reach (the number of unique people who saw the ad), the Frequency (the average number of times the ad has been seen by each person within the Reach), the Social Reach (not applicable because I'm advertising an off-Facebook website), Clicks, CTR, Bid and Price.
My ad reached around 4,000 people, appearing around 13 times for each, and had 8 clicks (0.014% CTR). I should point out that this is a very specific subset of people I'm targetting who live near where I live. For those 8 clicks, I spent around £6 (~70p/click). They recommend a bid for you, which seems to be fairly high (although a factor in it being high may be due to targetting a small set of people).
I decided to cut this ad short and spend the rest of my budget (the other £6) on an advert for Gatekrash. The ad for Freshly Minted was done on a cost per click basis. Facebook also allows you to run a campaign on cost per 1,000 impressions basis, which is what I decided to use for Gatekrash (as a test)
With the Gatekrash ad, I targetted several large cities in the UK, and only targetted people at university.
The suggested bid was much lower than the previous ad. The ad itself reached around 16,000 people, who each saw the ad around 4 times each. This yielded 4 clicks (a CTR of around 0.007% - half of the CPC method).
Obviously the ad itself and the way it was worded will have affected the clickthroughs, but I was honestly expecting a larger number of clicks.
I have a feeling that advertising internal Facebook pages (i.e. Pages, Apps, etc.) will yield better results than advertising external websites. I have not actually done this (yet), but I assume that because of the social reach aspect of advertising internal Facebook sites (i.e. seeing your friends have Liked a page or are using an app in the ad itself) will persuade more people to interact with the ad.
I also think that Google/Yahoo/Bing ads are probably the best way of advertising normal websites. People are actively searching for something related to your site, whereas on Facebook they're very passive and usually not very related to what the person is doing - they're just bolted on to the side of the page regardless.
That's not to say you shouldn't use Facebook Ads. I fully admit that I could have created better ads to display, but this was only a small test. I would test it yourself to get a feel for it - you can use small budgets (minimum of £1!)