Facebook Advertising is fairly straightforward – one of the main reasons why it has attracted millions of advertisers over the past few years. You choose the group of people you want to reach and the message to reach them with. The difference between a successful and unsuccessful campaign may be as simple as a few tweaks. As a diligent Facebook marketer, you spend hours and hours combing through data, looking for patterns and tweaks that you can apply, with the hopes of increasing your bottom line. These are a few of the most successful strategies I apply on a regular basis.
If you’re looking for a “getting started” guide, this isn’t it. There are plenty of resources on the web to get you started. This particular guide is for advanced Facebook marketers who are looking to get some new tricks under their belts. I’m assuming that you:
- Are familiar with the Facebook Advertising interface
- Have run multiple Facebook Advertising campaigns (ideally, spent a few thousand dollars+)
- Have your tracking pixels set up and know how to use them
- Understand basic campaign types and their purposes
So, let’s get started.
Facebook Advertising Targeting Strategies
Your targeting strategy will make or break your campaign. You can get away with subpar copy (obviously, not recommended), but you definitely can’t get away with a subpar targeting strategy. Choosing a targeting strategy, or building an audience, on Facebook requires a basic understanding of demographics, user behavior, and psychographics. This is probably one of the biggest distinctions between Facebook Advertising and search engine PPC.
Search engine PPC is rather straightforward. If you are selling a protein shake, you would target keywords related to “protein shakes.” This is because you are catching customers towards the middle of the funnel. These customers have already expressed interest in protein shakes, therefore it will be easier to sell them your product. On Facebook, you do not have that luxury.
Facebook campaigns are generally built to attract top of the funnel customers/leads. You are either trying to make prospects aware of your product or you are trying to spark interest in them. Yes, you can run remarketing campaigns to catch bottom of the funnel leads, but you will be left with a much smaller audience. Overall, your main goal is to increase awareness and push new prospects into your marketing funnel.
Finding the right targeting formula takes time and testing. Don’t neglect the basic targeting functions as those will work for a lot of campaigns. As you begin to expand your campaigns, you may consider some of the following advanced strategies.
Use Lookalike Audiences
Lookalike audiences were a well-kept secret for a long time, and now the term has become a marketing buzzword as many pros swear by this targeting strategy.
Facebook gives you the option to create lookalike audiences from any of your other custom audiences. This may include:
- Website visitors
- Page likes
- Custom audiences (i.e. customers)
- Email Lists
Facebook uses an algorithm to create an audience that is similar to the inputted audience. This saves you a lot of manual work and gives you the advantage of incredible analytical power. As much as we all pride ourselves on being great analysts, it’s hard to beat a computer in certain scenarios.
Facebook can look at your audience a wide range of factors that would take you months to analyze, such as:
- Education Levels
- Category Interests
- Specific Interests
You just have to choose your Source, Country, and Audience size (I always stick with “1” as the audience size.
Create a few lookalike audiences and start running ads tests. I will usually create lookalike audiences for the following groups:
- All site visitors
- All fan page likes
- All site customers
- Customer email lists
- All site leads (based on conversion tracking)
From there, I will start testing the audiences.
I never target the entire audience, as it’s generally about 2 million users. I will break down the targeting even further by incorporating the other information I have about my audience.
For example, if I was selling recipe books and I knew most of my audience was stay at home moms, I may use a Lookalike Audience of all website traffic and restrict it to Females between the ages of 28-44. I may also narrow it down further with a specific interest such as The Food Network. Once I have done that, I have a good starting audience that can be optimized as more data rolls in.
Turn Your Email List into an Audience
If you are collecting fresh leads on a regular basis, you will want to turn your email list into an audience. This works best if you have at least 5,000 people on your list, but it’s never too early to start building the list.
You can consider email subscribers to be warmer leads, therefore you may push different advertisements to these leads.
For example, if I owned a nutritional supplement company, my first advertisement to cold leads may be offering a free resource, such as a basic nutrition guide. Once a prospect was added to my list, I may push more specific advertisements, such as an article on “10 Benefits of Fish Oil” that is designed to push a product on my site. You can take this a step further by segmenting your email list by the length of the subscription. So, if someone has been subscribed to the list for a year, you may push sales and promotions via the advertisements.
Of course, this is just a complement to your actual email campaigns and shouldn’t be used to replace them.
Create a Sequential Retargeting Campaign
I’m a big fan of sequential retargeting because I believe it mimics basic human interaction. In real life, you rarely meet someone new and immediately push for what you want. You wouldn’t meet a client and ask them to sign a check immediately, nor would you meet someone in a bar and immediately ask them to go home with you. There is a subtle trust-building process involved with any relationship; marketing is no different.
Sequential retargeting allows you to show different messages to different users over a period of time. You can use time as sole the basis for your triggers or you can use user behavior.
When time is used as a trigger, you show advertisements based on when a user left your site. For example, you may show one ad after 3 days, another after 7, and another after 14.
When user behavior is used as a trigger, you show advertisements based on specific user actions. For example, you may show one ad when a user hits the homepage, another if the user has hit the pricing page, and another if the user abandoned the checkout page.
Regardless of the triggers you use, the goal remains the same. You introduce a more subtle, natural, and funnel-like approach to your retargeting.
Here’s how this looks. Let’s assume you own a high-end coffee company.
Day 1: User visits
Day 2+: User sees ads for premium coffee
Day 1: User visits site
Day 2: User sees ad for tips on how to make better coffee
Day 3: User sees ad for what to avoid when buying coffee
Day 4: User sees ad for 10% off company products
TRIGGER: User adds product to cart but abandons
Day 5: User sees offer for 10% off + bonus product
TRIGGER: User checks out
Day 12+: User is shown ads from time to time about promotions and new products
Of course, this example is oversimplified, but the general point should be clear.
So, how do you accomplish this in Facebook?
First things first, think of your actual sequence. Which days and triggers are you going to use?
Once you have the sequence planned, start creating the audiences.
For time-based audiences, create remarketing lists based on a certain period of time. Facebook allows you to create remarketing lists for users who visited your site in the last X days. Create one for 3 days, 7, days, 14 days, or whatever you need.
For user behavior audiences, keep the targeting window at 30 days+, but choose specific URLs that show different user intent. Here are some examples:
- Users visiting a specific blog post
- Users visiting your product page
- Users visiting checkout
- Users completing checkout
- Users who visited a product page but NOT checkout
- Users who spent 5+ minutes on the site but did not view a product page
Don’t create lists just for the sake of it. Focus on what type of user behavior is actually significant and create the audiences accordingly.
Then you just need to create the ad sets using the targeting.
Basic time-based targeting would look like:
3 Day – INCLUDE 3 Day
7 Day – INCLUDE 7 Day EXCLUDE 3 Day
14 Day – INCLUDE 14 Day EXCLUDE 7 Day
Account for triggers by including/excluding audiences:
3 Day – INCLUDE 3 Day EXCLUDE “Checkout Completions Audience”
7 Day – INCLUDE 7 Day EXCLUDE 3 Day EXCLUDE “Checkout Completions Audience”
14 Day – INCLUDE 14 Day EXCLUDE 7 Day EXCLUDE “Checkout Completions Audience”
Retarget Video Viewers
Getting customers to leave Facebook and convert on your site can be tricky. Essentially, you are interrupting the user’s experience on Facebook, which may be difficult to do for colder leads. To combat this issue, you may consider using top-of-the-funnel content to start building a warmer retargeting list.
This can be done by promoting an interesting video. You can assume that anyone who watches the video has an interest in the general theme, which will give you a list of qualified leads to market to moving forward.
Facebook allows you to retarget video viewers. You can retarget all viewers or narrow the list down to people who were more engaged with the content (based on view time).
The main benefit of this approach is that it allows you to create a hyper-focused audience on a much smaller budget. You should be able to get video views for a few cents each, whereas you may pay 20-50x that amount for clicks to your site.
Additionally, this strategy allows you to ease a prospect into a sale vs. pushing them in.
Let’s look at an example of how this may work in practice. Assume you run a SaaS business designed to help increase website conversions. Using traditional methods, you would try to pinpoint your audience and start boasting about how great your product is. Using the video retargeting funnel, you would create a more refined audience first. You may start with a short video on 5 conversion optimization tips and then retarget everyone who watched at least 50% of the video. You already know those people have an interest in conversion optimization, which means you have a higher chance of converting them at a later point.
Building Facebook Ads Campaigns
Once you’ve chosen you’re targeting strategies, the next step is building out the campaigns. It’s important to note that your targeting strategy is nothing more than a hypothesis, and a hypothesis needs to be tested.
Building campaigns is pretty straightforward. I’m not going to break down any ad copy secrets because every brand and audience is different. Instead, I’m going to go over a few advanced tips for campaign creation.
Split Testing Properly and Using Breakdowns
When I first started learning about split testing, I wanted to split test every element. This made more sense when the Facebook Ads system was limited in its reporting features, but now you have more advanced tools at your disposal. That said, I still see clients do this when I take over their accounts. They have ad groups for males, females, devices, placements, etc. While this can serve a purpose at times, it’s overkill in most situations. Facebook gives you the option to break down your results based on a few criteria.
This will save you a ton of time because you can create a standard ad group and use the breakdowns as split testing data. The only time it may be necessary to split out these groups is when a certain target is not getting equal traffic (i.e. M/F, Instagram/FB, etc.)
So, here’s how split testing should look.
Step 1: Focus on Existing Customer Data
Think about everything you know about your customer without making too many assumptions. Are they married? Male/Female? Old/young? How old are they?
Incorporate data from other sources, such as Google Analytics, AdWords, your email lists, etc. Once again, you do NOT want to make any unjustified assumptions. For example, if you own a makeup company, it would be fair to say that your ideal audience is female. That is a fair assumption. Contrarily, assuming that your entire audience is under 28 years old is an assumption that may cost you money.
Think about what you know about your customers and test the rest.
Step 2: Targeting Strategy
Start with a very basic targeting strategy. If you split-test multiple variables, it will be harder to understand the effectiveness of your strategy AND you will spend more money. The goal is to manipulate a single variable so you can be sure that your results are significant. That one variable should be in the “Detailed Targeting” section and you should split test one interest at a time.
Let’s look at two fictional examples to show why this is important. Assume we are choosing the winning ad group based on CPA over 2 weeks.
- WINNER: Female, 25-34, Married, College Educated, Interested in MTV, Vogue Magazine, Sephora Makeup, and Shopping
- LOSER: Female, 20-44, Single, College Educated, Interested in VH1, People Magazine, Going Shopping, and MAC Makeup
We don’t get any real long-term takeaways from this do we? We just know that the first ad group is better. Compare that to the following example:
- WINNER: Female, 20-34, Interested in MTV
- LOSER: Female, 20-34, Interested in VH1
This split test clearly shows us that people interested in MTV convert better than those interested in VH1. This insight can be applied to future ad campaigns and we can start split testing data like relationship status, and age range at a later point. Remember, every variable you test will require more budget if you want the results to be significant. Choose your variables sparingly.
MAIN TAKEAWAY: Manipulate one variable at a time and start by testing detailed targeting.
Step 3: Analysis
Let your ads run for some time so you can collect data. I generally give each ad group at least 1000 impressions before analyzing it.
Don’t expect the campaign to be a home run right away. Your goal is to slowly narrow down what is and isn’t working so you can optimize your campaign.
At this point, you should use the “Breakdown” option within Facebook Ads. Break down your results based on criteria that will allow you to take action at a later point. For example, if males account for 80% of conversions, you can choose to target males only as you move forward. These are my most common breakdowns:
- Time of Day
Sometimes, you will need to pull this data into Excel to make more sense of it. For example, I like to plot out the “Time of Day” data onto graphs.
As you do this, you should be paying attention to statistical significance. For example if you have 10 conversions, and 5 of them happen at 7PM while the rest happen in the morning, you shouldn’t immediately create new ads that only run at 7PM (trust me, I’ve done that before). Contrarily, if you have 800 conversions and 600 of them come from the 18-24 age group, you may consider targeting that group exclusively.
The whole goal of this analysis is to start eliminating the groups that don’t work. Let’s go back to the fictional example from above:
- WINNER: Female, 20-34, Interested in MTV
Let’s say we run some breakdowns and see that
- 70% of Leads came from the 25-34 age range
- 64% of leads came from Instagram
- 90% of Leads came between 6-10PM
This provides solid rationale for an ad group that looks like:
- Female, 25-34, Interested in MTV, Instagram Targeting, Scheduled from 6-10PM
We could let that group run for awhile to confirm this hypothesis and then start testing other variables such as relationship status, similar interests, device type, etc.
Use One Post vs. Multiple Ads
As you continue to split test different targeting methods, Facebook will automatically create new ads for every ad group. When you do this, you are diluting the power of your ad.
For certain types of campaigns, you should create a page post and use the ads to boost this post. This makes your ad look better and increases its power. Picture you have 10 ads with 5 likes each or 1 ad with 50 likes; which would you prefer? Rhetorical question.
The only downside to this method is that it will restrict your bidding strategies. You cannot optimize a campaign for conversions (using FB’s system) if you are running a post boost campaign. That said, you may be able to achieve the same CPA without Facebook’s optimization system. Test it.
Scale Budget Slowly
Many intermediate (and advanced) Facebook marketers make the mistake of scaling campaigns too quickly. This is not a good idea. Although it may seem like hindering your growth defies all laws of business, there are other variables in play.
I’ve run into this issue with clients before. A campaign is going well so the client wants to scale to infinite. If you can get a 100% ROI on $1000 ad spend, why not put $10,000 into the machine? This works great in theory but fails in practice.
Facebook automatically optimizes the delivery of your ad to your budget. If you increase your budget from $5/day to $50/day, you are throwing off the balance and will see higher costs per result. It will balance out over time, but why waste the money?
Instead, take a more conservative approach and increase ad group budgets by 10% every couple of days. This will allow you to scale without compromising your results.
Facebook Ads Analysis and Optimization
I started to touch on the topic of optimization in “Step 3” of the targeting section. Optimization is what truly makes or breaks a campaign and the goal of setting up proper targeting is to make the optimization process easier. Here are a few tips.
Know Your Metrics
You should know which metrics are relevant to your campaign before you get started. You don’t want to waste time digging through metrics at a later point. Different campaigns will have different success metrics. Whatever metrics you use, make sure they give you actionable insights.
Here are how some metrics align with each goal:
- Brand Awareness – Impressions, Reach, Frequency
- Website Clicks – CTR, CPC
- Leads – CPA
- Sales – CPA, $ value of conversion
Focusing on the right metrics helps you ignore the noise. For example, if I’m running a campaign with “Sales” as the main goal, my impressions really don’t matter if I have a negative ROI.
Generally, I will only stray from my main metrics if I notice dramatic changes in a previously stable campaign. For example, if I see my CPA double on a campaign, I may dive deeper to see that my CPM has doubled, which may lead me to notice that my Frequency is doubled. I can then assume that my ads are being shown too much and I should either expand the audience, introduce new ads, or drop my budget. This leads nicely into the next tip.
Monitor Frequency and Performance Metrics
If you ever have a campaign that is performing well for a long time and suddenly starts delivering worse results, look at your performance metrics. Frequency is a metric that allows you to see how often each person has seen your ad.
Here are some takeaways:
Frequency – If Frequency has increased (at least over 2), keep a close eye on your ads. If the same people are seeing your ads, they may have ad blindness or may not be interested in your ad. This will lead to a lower CTR, which will lead to a higher CPC, and ultimately a higher CPA.
- Cap budget so you aren’t reaching people to frequently
- Expand your audience to reach more people
- Introduce new ads to avoid ad blindness
CTR – If CTR drops, CPC will generally increase (because you are probably paying for impressions).
- Test new ads
CPM – CPM dictates how much you pay for every thousand impressions. If you see the CPM rise, check your quality score. Your CPM may be higher if the quality score is lower because Facebook users are not responding to your ad very well. CPM can also increase during period of higher competition, such as the holiday season.
- Create better ads
- Target different audiences
- Modify budget to meet ROI needs if CPM inflation is due to higher competition
Understand Attribution Models
Attribution models are used to give advertisers some flexibility over their conversion reporting. By default, Facebook will claim credit for a conversion if a user converts within 1 day of viewing the ad or within 28 days of clicking the ad. So, what does this mean?
View-through conversions and click-through conversions are not created equal and you may be misattributing conversions. For example, if you are running a retargeting campaign and using view-through conversions with a wide click-through conversion window, you may be improperly assigning revenue to Facebook. Using this model, a visitor could go to your site, sign up for your mailing list, click on a Facebook ad a few days later, and convert from an email within a few weeks. The conversion belongs to the email, but it will be attributed to Facebook (within the FB ads platform).
It can be difficult to determine the exact path that a user took before converting, but you should incorporate attribution models in your analysis.