AdWords has intent. But Facebook ads have custom audiences.
These are people who’ve already looked at your website. Viewed your products and services. Subscribed to your newsletter. Or purchased previously.
And you can segment or target each specifically with a campaign that perfectly matches their interests, personalities, fears or motivations.
Audience targeting is easy when you’re dealing with a pre-built custom audience full of people who already know all about you. Clicks, leads, and conversions are easy.
But what if you’re starting from scratch? What if you can’t rely on custom audiences to bring in ‘top of the funnel’ people?
Here are three steps to get started.
Why Custom Audiences Convert Best
The best Facebook ad will suffer when targeting the wrong audience. Improve the targeting, and even an average-to-middling quality ad will perform better.
The fastest, easiest, most powerful method of audience targeting is to use custom audiences.
These are tailor made, pre-built audiences full of people who already know who you are, have checked out your stuff, and have possibly even purchased from you in the past. They’re your past website visitors (over the past 30-60 days). They’re people who’ve viewed a specific product page. They’re email addresses from your newsletter. They’re customer phone numbers from your database.
If you’ve got this data to work with, you’re gold. You can (and should) rely heavily on custom audiences as much as possible in the ‘middle’ and ‘bottom’ of your marketing funnel to target the best audience (and thus, reap the lowest Cost Per Clicks and Cost Per Leads).
Don’t take my word for it.
When Facebook introduced their Relevance Score, AdEspresso ran a test using the exact same advertisement to two different audiences.
The first audience, below, was generic. Targeting was off, as you can see by the dismal 2.9 Relevance Score.
(image source)
As a result, website clicks were costing $0.142 a piece.
Now contrast that with the same ad (and same budget), but shown to a custom audience (comprised of website visitors from the past 90 days).
(image source)
Relevance Score is up dramatically to 8.0 Comments, Likes, Clicks and Shares are also WAY up too.
But the best part? The Cost Per Click has fallen to only $0.03.
(Yes. That’s three pennies. Which is like basically free.)
That’s why message match matters in online advertising. Especially on Facebook where audience targeting can mean all the difference between campaigns that convert from those that don’t.
All of this sounds great. But here’s the problem… What if you don’t have a custom audience to rely on? They’re the perfect solution to nurturing and closing existing customers. But they can’t help you attract, acquire, or bring in fresh meat.
What if you’re starting up a new site? Launching a new business? Introducing a new product or service to a different vertical? You gotta start from scratch. And you ain’t gonna have no past visitors, newsletter leads, or customer database to rely on.
You should watch this webinar for more (which informs the points made in this post). But here’s the CliffsNotes version.
Step #1. Start by Finding Your Audience Demographics
The best audiences should understand the problem you solve. They’re knowledgeable (to a point) about their pain points and how to fix them. And they should be informed by basic information that you already have.
This all goes back to the fundamentals. The basics and the boring stuff, like buyer personas. The Marketing 101 stuff like demographics, psychographics, and other important stuff that defines who we are and why we do what we do (or don’t).
(image source)
This sounds trite. Whenever the phrase ‘buyer personas’ pops up in a blog post, all I hear is “Blah blah blah blah blah blah blah”, too. But it’s impossible to get audience targeting right without taking into account the differences between these people, the value propositions, and the stuff they care about.
The easiest place to start – unsurprisingly – is your own Facebook Audience Insights.
Let’s check out one client example:
96% of their fans are women. Almost all of them are located in the U.S., with a majority in the South West (with a few large cities sprinkled in).
Boom. Done. And it took all of ~1 minute to find.
Google Analytics will also provide a Demographic Overview (located under the Audience section) which can complement similar data.
This, along with any other secondary research (like association or research reports) can help you start building out different customer segments and personas that will eventually guide your Facebook ad audience targeting for different campaigns.
That’s a good start. But you’re gonna need to know a little more information first in order to create successful ads later on.
These are the main problems and pain points, along with how you can help or overcome these challenges in someone’s daily life.
Where do you get this stuff? From your customers. (Hey, Brad, you mean, we actually have to talk to people to do marketing?!) Yes! So go outside. Attend an event. Pick up a phone.
Fire up SurveyMonkey, Google or Gravity Forms. There’s also the Hotjars and Qualaroos of the world if you want to get fancy. But you don’t have to. People won’t hesitate to talk about themselves or bitch about what’s preventing them from accomplishing XYZ.
You just gotta ask.
Step #2. Begin Interests Research
If, like me, you’re looking for the fastest (read: laziest) way to target a new audience, you probably just select a few competitors or blogs at random. That’s OK to get started. But more than likely, you’re going to need a tiny bit more effort to get a large enough sample to start with (read on to find out the magic range).
So, how?
You can start by searching directly on Facebook. For example, a query like: “Pages similar to HubSpot” will give you a list of similar brands (which are also influenced by your social graph).
You can also spy on what interests these pages have previously ‘liked’ to get a better understanding of related topics as well.
From there, head over to the Google.
Right now we’re looking for ‘top of the funnel’ audiences to start targeting. That means you need to keep ‘searching intent’ in the back of your mind while pulling up example interests from a search engine. For example, “organic skin care blog” is great for opening up your potential audience to the widest amount possible, while “organic face cream” is a specific query that contains more purchasing intent.
Amazon is another perfect place to start. A query like “men’s chukka boot” will instantly provide a bunch of top brands that you can go back to Facebook and search for (selecting their name as a new interest and targeting all those who like them).
Beyond competitors and alternatives, is an information source. These are the places you just uncovered in the last step, where these segments and personas get their news and information from.
Specifically, these are typically individual influencers, media, blogs, or some variation of all three (like Tim Ferriss / Four Hour Work Week). And like the crazy, polarizing difference between Fox News and CNN, each caters to their own specific type of audience. In the business world, that translates into Bloomberg = old baby boomer dudes and Wired = tech-savvy Gen X/Y.
Let’s use this example now that we’ll carry over into the next step. We’ll select a few media properties to find relatively tech-savvy, educated, business owners.
We’re off to a good start.
The 720,000 audience size falls in the suggested range of around 500,000 thousand to a million users. (Generally speaking, the bigger the budget the larger the audience. And vice versa. So if your budget is under $200, target 100,000-200,000 range instead.)
The reason for these size ranges is that Facebook needs enough space to find the cheapest ad most likely to convert users.
Only problem?
Right now it’s a little vague and generic. We might be targeting a certain kind of person (demographics), but we haven’t really taken into account what these specific people might be interested in (psychographics, et al.)
So we need to refine it a bit, by including and excluding other criteria.
Here’s how.
Step #3. Interests Intersection & Exclusion
People who eat at Michelin starred restaurants aren’t coupon clippers.
They’re mutually exclusive because these people know they’re gonna drop at LEAST $75-100 per person before even stepping foot inside. The mere mention of that price per check makes your average Groupon buyer weak in the knees and break out into a cold sweat.
That’s the Catch-22 of targeting people in the ‘top of the funnel’.
You want to keep the audience size big enough to give yourself a fighting chance. Without enough reach, you will never scale a profitable business.
BUT…
You also want to target specific people who aren’t going to balk at your offerings or price point too. Otherwise, you’ll pay for a ton of useless, low-value clicks.
The solution is through interests intersection (or exclusions). Here’s what I mean:
Earlier, we started with a good (albeit, overly generic) audience based on a few different media properties. Let’s be a little more specific though to make sure we’re only getting targeted people.
So let’s add another interest that segments this initial audience by other criteria (like what they might do for a living).
Now we’re targeting relatively tech savvy, educated consultants. But our audience has dipped below the ~500,000 minimum. So let’s add a few more related interests that are similar (like trainers or coaches) and keep ‘any of the above’ selected so that ALL potential people are included.
But wait a second, because this might open the door to some people with conflicting interests.
For example, if we’re trying to sell marketing-related products and services, other marketers most likely aren’t gonna buy this stuff. (And if they are, they should probably have different messaging to say the least.)
So while we’re adding additional interests, let’s also add a few exclusions to make sure these people with conflicting interests don’t slip through the cracks.
Here’s the result:
Perfect. We’ve managed to segment our audience much better through interests intersection, and then reduced the chances for wasting ad spend on people who’re unqualified or uninterested.
Optional Step: Advanced Targeting
Advanced targeting can help you further refine the audience selection if you have a very niche product, service or business. For example, these are advanced behaviors or criteria like relationship status, income, net worth, etc. So if you’re selling investment services to accredited investors, targeting based on income or net worth is perfect.
The only problem?
People don’t always willingly provide this data to Facebook. So, in many cases, they’re relying on third party sources to help supplement their information.
The result is that it will reduce your reach a lot. For example, if we just went into behaviors and tried to select a specific industry or vertical, it would bring our audience size down tremendously (to an almost uselessly low range).
You can, however, use some of these advanced options to come up with creative workarounds. For example, you might not have a lot of data on income or net worth, but if someone went to Harvard then it might be a safe bet they’re close.
Conclusion
Facebook’s custom audiences give you one of the most powerful targeting options for any advertising medium.
You can target people who are brand aware, all the way to past leads and customers who’ve already paid you big bucks in the past.
The problem is, that in the beginning you don’t have access to these people. You don’t have the website visits, lead email addresses or customer ID’s to put your Facebook targeting on autopilot. Instead, when you’re creating a new business, expanding your product offering or entering new markets, you have to start from scratch.
You need to target a larger ‘top of the funnel’ audience that’s both broad enough to give you enough potential customers, while specific enough to rule out those who’ll never buy from you in the first place.
It’s a delicate balancing act. But using demographics and psychographics can give you an idea of which interests to start targeting. And using interests intersection with exclusions can help you sift the gold from the rubble.