It’s a good general way to begin, especially when you are testing if one target works for you or not. But what a lot of people don’t realize is that once you know that various targets are working for you, it’s better to combine them into one giant Ad Set – because the system then has a larger pool of potential users to target, and will have an easier job finding cheaper eyeballs for you.
Let’s look at a bookish example.
Imagine I’m an author of epic fantasy, and my comp authors are people like David Eddings and Raymond E. Feist and George R.R. Martin and Janny Wurts and Robin Hobb. (Let’s just pretend they are all targetable interests on Facebook – I’m too lazy to check!)
I would usually start out with five separate Ad Sets, targeting each of these guys separately. That’s the only way I can measure the performance of each of those targets. Splitting them up in this manner also allows me to dedicate more of my budget towards George R.R. Martin, if I wish, but my primary concern right now is making sure that, for example, Robin Hobb is as good a target for me as David Eddings – if I just lump them all in the one Ad Set, I won’t know who is working for me and who isn’t.
So, I separate them. If Raymond E. Feist turns out to be a loser, I’ll cut him (with apologies to Mr. Feist). If his readers aren’t as responsive to my ads as all the others, I switch off that Ad Set and let the rest of the ads run as they are performing well – until my Campaign is over.
What do I do the next time I’m running a Facebook promo? Do I set it up the same way? Well, no. There’s a better way, now that you know which of those authors works for you, and which doesn’t.
You are far better off lumping all those authors together now, and having one giant ad targeting David Eddings, George R.R. Martin, Janny Wurts, and Robin Hobb as one. People might be reluctant to do this because the audience size might be quite large, but – and this is one of the counter-intuitive things about Facebook that makes it very different from any other ad platform – in this case, in these specific circumstances, with tested targets, bigger is most definitely better.
Again, assuming these are all authors you previously got good results with, throwing them all in the Ad Set might result in a huge audience, but that will actually allow the system to find you the most responsive users for the cheapest price. A large pool like that, again, I must stress, once these are tested targets, will give Facebook’s system much more latitude. It will serve the ads to people more likely to click and who cost less – everyone’s eyeballs are valued differently, because everyone has a different set of advertisers trying to reach them (authors, for example, can be incredibly expensive people to target because of all the people selling $1000 courses and the like).
This approach keeps your initial costs down, but also stops them spiralling out of control too. When you have four or five different Ad Sets, there is a danger of what’s called Audience Overlap. This is best explained by returning to our epic fantasy example above, but from the perspective of a reader.
It’s not the biggest stretch in the world that a fan of David Eddings might also be a fan of Raymond E. Feist, given that the big debate in fantasy when I was growing up was whether you preferred Magician or The Belgariad (and everyone had read both). It’s safe to say that these are two overlapping audiences.
The danger in targeting David Eddings and Raymond E. Feist separately is that anyone who Likes both pages might get overserved an ad. And this is bad for a few reasons.
You can see how many times users, on average, are seeing your ads with the Frequency metric in your reports. There are different approaches here, but I personally don’t like that going above 2. My threshold here is much lower than it would be with BookBub, for comparison. As I like to only run ads to Newsfeed, my ads are hard to ignore. They are literally front and center – not served up at the bottom of an email and possibly going unseen. If a given user sees my ad twice in their Newsfeed, I don’t think there is much point showing it a third time to them. They will either have acted already, or not. (Note: you might look at this slightly differently with long-term Campaigns pushing a permafree.)
Usually when the Frequency starts to rise, you will start getting what’s called “ad fatigue” – people tired of seeing the same ads again and again. Those who aren’t interested will ignore your ad again, dragging down the CTR, reducing the engagement on your Campaign. This will increase your costs because Facebook uses relevancy very heavily. If you have great CTRs, your costs drop. If everyone hates your ads, your costs will spike. So if people start ignoring your ads, this is bad for you, and that’s before we get into people actually Hiding your ads (which is really bad for you).
Even the people who liked your ads and clicked on them and bought your book can’t buy it a second time, so even those people will start dragging down the performance of your ads on repeated viewings.
Ad fatigue is bad. Seriously overlapping audiences lead to lots more of it because I might be watching my David Eddings Ad Set with a Frequency of 1.6 and my Raymond E. Feist Ad Set with a Frequency of 1.4 and think everything is fine, but not realize that is masking a bunch of people being double counted, getting served my ads three or even four times because they Like both authors. See what I mean?
There are a couple of ways to prevent this.
First is going for those monster Ad Sets, once your initial testing is done and you are happy that the bunch of authors you are targeting get reasonable results. If all the authors are in the same Ad Set, you won’t have Audience Overlap and you can easily control the Frequency which your ads are shown and prevent ad fatigue. This will stop your costs growing too fast over time. Facebook Campaigns tend to do that anyway, there’s little you can do to avoid that, but minimizing ad fatigue will considerably slow down that (more-or-less inevitable) decay.
But it doesn’t always suit you to throw everything in the same Ad Set. There are all sorts of reasons why you might want to have one of those author targets separated out, aside from early testing. For example, let’s say you actually met Janny Wurts at a conference once, and she was lovely and you became friends and she read your book and then obligingly provided you with a juicy quote.
Using that quote in your ad text would obviously be something that her fans would respond to. And probably Raymond E. Feist’s fans too as they are long time collaborators. In fact, let’s say all the above authors’ fans responded well to that quote… with the exception of George R.R. Martin’s demographically different fanbase.
Let’s further assume that after a little testing, you found that a different bit of ad copy worked better there. How do you prevent audience overlap now, given that fans of David Eddings and Robin Hobbs are quite likely to also be fans of George R. R. Martin?
Actually, let’s take a step back.
How do we know where audiences overlap? We might not be sure that David Eddings fans are George R. R. Martin fans – maybe the overlap isn’t enough to be sweating about. But it might be. How can we figure this out? Well, Facebook has something called the Audience Overlap Tool. Simply follow the process outlined on that Facebook help page and the system will tell you if there is enough overlap to be worried about.
For our purposes here, let’s assume there is. How can I prevent ad fatigue when George R. R. Martin’s discerning fans require a separate ad and Ad Set with different text?
It’s quite easy, in fact. You can do something called Exclusion Targeting. Remember when you are setting up your ad and you choose an author like George R. R. Martin as a target? You know the way you can Narrow that audience by choosing Kindle as a further interest to make sure you don’t hit fans of print books or users of Nook with your ad? Well you can also Exclude interests here too. So you can have your ad running to George R. R. Martin fans who own a Kindle, but exclude everyone who likes Janny Wurts, David Eddings, and Raymond E. Feist. Pretty neat.
Excluding audiences has other uses too. For example, let’s say you are running ads to boost sign-ups to your mailing list. One of those ads might be running to those who Like your page, but you don’t want to pay money to show an ad to people on your list already – so simply exclude them. Easy peasy.
I’m sure you can see how many uses this can have, just don’t go too crazy with excluding audiences or the system can trip out and you might have serving issues.
One last thing: I’ve seen some debate about whether you can bid against yourself on Facebook, and it was something I’ve been confused about myself, but I spoke to someone at Facebook about this recently and I can be categorical here.
You can never bid against yourself on Facebook, not from within the same account at least. If you run two different Facebook Pages and have two Campaigns targeting the same people, then, yes, you can bid against yourself and drive up the cost of your ads.
That’s never a thing inside the same account, even if you have 10 Ad Sets or Campaigns targeting the same set of eyeballs. The danger here is not bidding against yourself, but audience overlap, and the resultant ad fatigue – which will drive up your costs and harm the performance of your ads. And if you read my blog post last week on Campaign Budget Optimization, and the suggested workarounds of creating separate Campaigns for each Ad Set where you want to ringfence a certain budget, you can see that audience overlap and ad fatigue are going to become even bigger issues in the future.
But now you have solutions! And next time we'll look at some more ways to keep those ad costs down.