Testing a new idea requires a completely different approach to performance marketing than scaling a business. If you have the wrong approach with your Facebook ads, you’re doomed.
Building the right audience is the first step to creating a Facebook campaign. Let’s make sure that you’re building an audience that’ll reflect the goal you’re trying to achieve with your ads.
Testing new ideas
In pre-launch/ pre-revenue stage, it’s important is to validate your core business ideas rather than generate revenue.
This is one of the more unusual uses I have for performance marketing. If you think of how we validate ideas as part of positioning, the steps are about finding who responds to a message. They can be viewed as asking a series of one-question customer research surveys.
This is not what most people want from their performance marketing.
So it shouldn’t be surprising that when you let Facebook handle the optimization of these ads, you won’t get the results you need.
This is the same tool that everyone else is using—display advertising—but used in a different way.
What we’re looking for
- Targeted audiences by demography or experience
- Limited optimization from Facebook
- Measure on CTR (Click Through Rate) instead of CPA (Cost Per Acquisition)
- Build Ads to test and validate theories
So here’s a crazy fact about Facebook ad targeting. You have the incredible capacity to target your ads with laser-like specificity. When creating ad campaigns that you want to drive traffic that converts on a landing page, you probably shouldn’t use any of them.
But these aren’t ads that are designed to convert on a landing page. They’re designed to ask a question to a specific set of people. So we’re going to make use of a lot of the targeting features I’d normally recommend you stay away from.
You’re going to want to start by building a customer segment you want to target. Customer segments are what most people mean when they talk about a buyer persona or an ideal customer profile.
A persona offers very specific insights into a fictional customer and their problems and needs. A segment, on the other hand, groups customers based on big, broad characteristics. Marketers might make segments for customer demographics, behaviors, and where they are in the customer lifecycle. These segments are used for distribution purposes — from media buying up to emails.
Your segments should be pretty narrow. When you’re targeting on Facebook for new idea testing, you generally want an audience between 100,000 – 5 million people. Any smaller and you’ll find you can’t do much experimentation before your audience gets brand fatigue seeing your ads. Any larger and you’re probably not targeting a narrow enough market to learn a lot about preferences from the ads.
Once we know what segment to target, check out how to build a top of funnel test campaign on Facebook. That’s if you’d like a walkthrough on how to set up your ad campaign correctly.
What you want to see
When testing very different ideas to the same audience, you’ll see a clear performance difference in the Click Through Rate.
Here are identical ad units targeted to different audiences and receiving roughly the same number of impressions. It’s pretty clear that one of these audiences is much more responsive to this messaging than the other.
That said, none of these ads are performing well enough to make me happy that the overall audience response is high enough to make growth easy as the company really goes to market. For that, I target at least a 2% CTR.
⚠️ It’s important to note here: in almost every other instance than this kind of testing, CTR isn’t the most important metric for you to consider when you’re measuring ad performance. Again: we are making an unusual use of Facebook display ads here. In most cases, you don’t care nearly as much about CTR as you do here.
The other metric to consider here is ad frequency. If you see more than about 1.6x frequency for a 1-week test with a modest $1k – $4k budget, you may have problems with too small an audience size to count on it to grow your business long-term unless your average customer value is quite high.
Scaling your business
Once you know who you sell to and how to talk about what you do, your focus for your advertising changes. You don’t need to create many narrow audiences to try and manually target someone who might want your services, and you don’t want to eliminate all of Facebook’s own targeting AI anymore.
Instead, you’re going to create an ad campaign that is optimized for conversions (make sure your FB pixel is firing correctly!), and you want as broad a reach as you can make it. You’re going to want to set up two different audiences here for scaling purposes: one that brings you new people, and one that retargets visitors to bring them back.
New Traffic Targeting
This is as simple as setting up a geographic region like the US, and then setting an age target like 18 – 55 years old.
That’s it.
You let Facebook’s machine learning algorithm take it from there.
This is simple, but there are a few things you need to know before you can do this.
- You need to already have a creative that you know performs well to bring you traffic that converts into a lead. That’s the point of the GTM funnel that we talked about earlier: it teaches you how to talk about your business in a way that works reliably.
- You need to have all of the technical aspects of your Facebook account set up correctly. If your pixel isn’t reporting correctly, then Facebook can’t learn and you’re throwing your money away.
- You need to commit to iteration. When you’re running broad targeting like this, you want to test variations 2x or 4x a month & analyze what’s working and what isn’t.
As part of this broad reach targeting, you may well want to test out Facebook Dynamic Creative, where you create several different pieces of creative and text and let Facebook mix & match them to achieve maximum performance.
How do you know what creative performs best with Dynamic Creative Ads?
To see creative-level analytics, go to Facebook Business Manager and open up Ads Manager. In the Ad Manager, you’ll have to select either the ad set or ad level to view the data. Then click the “Breakdowns” menu and then select the “by dynamic creative asset” tab. You can see some results here.
Should you build a Lookalike audience to boost new discovery?
When you have created a very broad target audience as above, and you’re telling Facebook to find you the people who will most likely convert to your website based on its own ML processes, you actually don’t need to build Lookalike audiences. Facebook is already doing that for you invisibly because it will, of course find people who look like the best performing impressions on its own.
Retargeting
If you’re running ads on platforms other than Facebook, or if you’re getting substantial traffic from other channels than Facebook, it can be handy to create a retargeting campaign. This is because you specifically know that you’re bringing traffic back to your website instead of enticing it to come and visit you for the first time. When used in conjunction with a broad-reach discovery campaign and new traffic from other sources, it helps you create even more value from your Facebook ads.
But there’s a catch.
You can easily end up siphoning a lot of the new traffic you find from your Facebook broad reach ads into this retargeting campaign and thus end up teaching the algorithm the wrong thing.
Building retargeting correctly
To manage your retargeting to keep from damaging your broad reach Facebook discovery campaign, you simply need this. Make sure you’re running your broad reach ads to one landing page. Your retargeting ads should not include anyone who has visited that page.
- Go to Ads Manager and create your dynamic ad or open an existing one.
- In your Ad Set, find the Audience section. Choose Retarget ads to people who interacted with your products on and off Facebook.
- Select a retargeting option, update the corresponding number of days and choose the products to include.
- Then choose Custom Combination. Here you can specify what audience interactions to include or exclude. For example, you could exclude visitors to the Facebook Broad Reach Discovery campaign’s landing page in the last 14 days.
- Make sure your conversion attribution window on the Facebook Broad Reach Discovery campaign is set to less time than the exclusion window set for the retargeting ads. That way, you won’t confuse the ML training, but you can perhaps still recover that traffic after it has seen the retargeting ad.
These two different means of using Facebook ads have completely different goals. They should be used for different stages of a company, and they should have different outcomes for your business.
This is why sometimes you see me getting into Twitter fights with other Facebook Ads experts on social media
Sorry, Charley, but you need to answer those $0.05 questions before you can spend a million. 🙂
