Advertising on Facebook
If you’re tired of paying high click rates on Google for PPC ads, then you should give Facebook a try. Let’s talk about some of the differences between the two and how you should approach each.
Don’t Spend a Dime Until You Do This…
First, and this is important… no, it’s UTTERLY, CRITICALLY, IMPORTANT… you should not invest one single penny in advertising unless you can actually measure its effectiveness. You may say “well duh”, but people do it everyday. Everyday people pay Google, or Facebook, or some other site, to display ads… and they may look at various analytics about traffic, but unless you have your analytics setup to actually track when a sale (or other targeted event) occurs from a specific ad source (and even the individual ad), then you have no way to know what actually caused the sale. Zero, zip, nada. You may as well flush your money down the toilet.
Once this is done, you can then look at your data in Google and view the goals met and the keywords that led people there from search engines (or other source). This will allow you to know if the money you’re spending on PPC on a particular ad from a particular source is responsible for any sales and for how many sales (or completions of whatever the goal is).
Why is This SO Important?
This is important because here in the real world more than one thing usually happens at the same time. And unless you track everything you can, you won’t know that the PPC ads had nothing to do with the little bump in sales you saw, that instead it was someone mentioning your product on a popular blog or web site. Or, you would know that for every $1 spent on Facebook you’re getting $2 in sales.
Now to my Facebook Tests…
I really like the way Facebook handles ads.
- They have a fast turn around on approval – I’m talking 10 minutes to an hour in my experience.
- They have a easy to understand, real-time, display that shows each ad your running along with stats like impressions, clicks, and click through rate… and how much you’ve spent.
- You can set low per-day limits – I set a $10 a day limit.
- They make it easy to clone an existing ad and make a few tweaks.
The one thing I didn’t seem to be able to do – and it may be possible, I just didn’t see it, was to specify that my ads run at certain times of the day. But you can manually pause and restart your ads or campaign at will.
But What about Keywords – How to Target?
Facebook does not have keywords because its users are not searching, and therein is the big difference between advertising on Google and Facebook. On Facebook you select various profile settings to target where and to whom your ad is shown. You can select age ranges, male vs. female, looking for (male, female), location, and interests groups, etc. You can really zoom in on the perfect demographic for your product.
One nice feature: as you tweak the demographic settings for an ad you see a real-time estimate of the number of possible people or profiles that could see your ad.
What About Facebook’s Suggested Bid Range?
I didn’t use their suggested bid range. I cut it a little more than in half. Once I had tweaked the demographics of my first ad sufficiently, the suggested bid range was between $1.02 and $1.80ish… I bid a maximum of $0.50, with a cap of $10 a day. I got as many impressions as I needed to reach my $10 limit – they do a good job of throttling the ads so that they don’t all get displayed in 10 minutes – they were spaced out around the clock. And, a lot of the clicks cost a little under my 0.50 bid…
Don’t be Pushy… A Picture IS Worth a Thousand Words!
Facebook users don’t really respond to pushy sales ads. You have to hook them with an interesting photo and then some copy that peaks their interest. I read a really good post on Troy White’s blog that his Facebook advertising tests showed that the most important factor in successful ads on Facebook is the image – and he is correct. I tested the exact same headline and body copy with several different images. Out of the 8 or so images tested only one had any click through at all. I got a few clicks (meaning literally 2 or 3 clicks) on one of the other images, but most received NO clicks at all. The same ad with the winning image received clicks – lots of clicks.
What about My Results, My Sales?
Please read the section above about tracking your ads. My weeklong test with 25 or 30 variations of my ad resulted in…. drum roll… Zero Sales. Yep… nothing. However, I thought after the first few days that it was. I had not looked at Google Analytics yet, just the Analytics on Facebook showing CTR, etc and was happy that I was seeing clicks and was making sales. My brain naturally saw the cause and effect. WRONG. Once I sat down and spent some time in Google (I had setup my goals as noted above before running the first ad). I could not find any sale that was directly related to Facebook ads. I did see a few where the person came from Facebook, but not from the ads. I did see where people came to the site from the URL from Facebook Apps (the ads) but that did not link to any sales. It’s possible that one or two sales were the result of someone clicking the ad and then coming back to the site later. There are always those things that skew the totals a little.
I will test Facebook again. It didn’t cost much to run this test. But, had I not setup analytics correctly, I would have mistakenly thought the ad was working and kept spending money on it… money that would have been wasted.
For this product, I didn’t come up with a winning ad. It’s a low cost product, so I didn’t have much margin to work with: I had to have low click cost and some sales… I only got one of those, not both. Maybe if I’d raised my bid I would have received better placement or an increase in some other factor on Facebook I’m not aware of, but with this product, even if the ads had resulted in sales, at the higher click price I probably would not have broken even, I would have lost money. My goal was to at least break-even on the ad.
Until next time,
P.S. Are you interested in a low cost tutorial on how to setup both Facebook ads and Google Analytics to do this? Leave a comment if you are.