Lookalike Audiences
When you run ads on Ad Platforms such as Facebook or Google Display Ads & Search, you place bids in a real-time auction for positions in the news feed and stories. Most advertisers start by bidding on users based on interest, age, gender, and location. Because this is the most popular strategy, it usually underperforms in the auction. This is also partly why Facebook and Google now suggest that advertisers target a broad audience and let the platforms optimize audiences.
Facebook offers a solution called Lookalike Audiences (and Google offers Similar Audiences). These are built by uploading a list of customer identifiers (such as Google ID, Facebook click IDs, email addresses, etc.) to Facebook or Google. Facebook creates a custom audience that matches current Facebook users to the identifiers. From there, Facebook can create a lookalike audience of similar users to the original list.
These lookalike audiences allow you to take advantage of the hundreds of user details Google/Facebook tracks, from post content, likes, shares, interests, and more. With these data points, the platforms create a much more detailed view of the type of person who buys your product, so you don’t have to rely on demographics like age and gender for targeting.
Shortcomings of Lookalike Audiences
Lookalike Audiences are not perfect because not all of your customers are the same. There are likely several different types of people who buy your product, which will confuse Facebook’s lookalike algorithm because it’s looking for a single type of user to target.
And, not all of your customers are equal in value. Your high-value customers spend repeatedly, while low-value customers only respond to steep discounts. Basic Lookalike Audiences don’t understand these differences in customer value.
Customer Lifetime Value Can Help
After calculating the lifetime value of each of your customers, you can unlock the following three strategies for better marketing:
- Value-Based Lookalike Audiences (Facebook Only)
- High/Low Value Audiences
- Value-Based Bidding
Connecting your CLV data to ad platforms can be done manually or via a Conversions API. This API is a practical way for advertisers to pull in external data to create custom audiences in the platforms.
Under the Data Sources tab in Facebook Events Manager, you can choose your partner under Settings. Facebook currently partners with a variety of customer data platforms, system integrators, attribution platforms, and more.
You can follow similar instructions for Google that are outlined in this article.
1. Value-Based Lookalike Audiences (Facebook Only)
Value-Based Lookalike Audiences leverage Facebook’s own algorithm to show your ads only to users with high CLV. It is the simplest way to leverage CLV to improve your customer acquisition.
To create a Value-Based Lookalike Audience,
- Generate CLV for each customer and match it to an identifier, such as email address.
- Upload that data to Facebook to create a Value-Based Custom Audience. Facebook will then match its users to the customers which you uploaded.
- Create a Lookalike Audience based on that custom audience. This will become your new Value-Based Lookalike Audience.
- Facebook applies their algorithm to look for what defines a high CLV customer and populates the lookalike audience with users of the same type.
- When running new acquisition campaigns, use this new lookalike audience, and exclude current customers from the custom audience.
Facebook reviews the users in the Value-Based Custom Audience to determine what high-value customers have in common, including posts, like, shares, interests, and more. Your Facebook campaign will then optimize ad-ranking to be shown to users who resemble your high-value customers.
This strategy reduces cost by reducing the number of times you need to display your ad (impressions) in order to generate a successful sale. It also increases your revenue by acquiring customers with higher and more frequent spending habits.
However, for some marketers, this type of audience may still not have the desired performance because it relies on Facebook’s algorithm. Facebook may be targeting users too broadly. In such cases, we recommend the second type of audience.
Here are example case studies by Facebook on Value-Based lookalikes: (1) YellowHammer and (2) Revolve.
Google will be rolling out lookalike audiences soon. With 12-months of conversion data, Google will bucket current customers into high, medium, and low-value segments for lookalike audiences. Marketers will have the option to manually adjust audience bids or follow value rules if they choose auto bidding.
This new feature from Google will provide advertisers with custom similar audiences based on conversions.
2: High/Low Value Audiences
To more tightly target high CLV customers and avoid low CLV customers, you can generate two audiences. This approach is slightly more complex, but it gives you more control over how to target your ads. It also reduces your dependency on Facebook’s own business logic that comes built-in with their algorithm.
To implement this strategy:
- Start with labeling the CLV for each of your customers.
- Divide your customers into two lists: high CLV and low CLV. It is best to use the top and bottom 20% (approximately) of your customers in order to emphasize the differences between the two..
- Upload each list to Facebook and create a normal (not value-based) custom audience.
- Facebook will then match its users to the lists you uploaded.
- Create a Facebook lookalike audience for each of the lists.
- When running new acquisition campaigns, use the high-CLV lookalike audience. Exclude both the low-CLV lookalike audience and your current customers.
The figure above illustrates how this strategy works. This strategy gives you more control by relying on your own definition of the type of customer you wish to acquire. It is ideal whenever your cost of acquisition is high compared to your average customer lifetime value.
If you are trying to maximize the number of customers acquired, we have developed a third strategy.
3: Value-Based Bidding
When advertising on Facebook and Google, there is a trade-off between the maximum amount you are willing to spend to get your ads displayed (bid cap) and the number of times your ads will be shown (impressions). If you spend more, you will win more bids, but your costs will increase.
The idea behind Value-Based Bidding (VBB) is that even low CLV customers are worth acquiring, as long as you don’t spend too much to acquire them. VBB uses multiple audiences at multiple bid caps. You acquire more customers while ensuring an appropriate margin. The illustration below explains how this works.
To implement VBB, start by defining your spend limit: your target CLV to CAC ratio. If your market is highly competitive, you may need to spend nearly as much to acquire your customers as you derive from their full lifetime value in order to win the bids. So, you may need a ratio like 1.5:1. If there is less competition for your customers, or if you prioritize profitability over quantity, you can maximize your returns by targeting a CLV to CAC ratio as high as 4:1.
To implement this strategy:
- Start with the labeled CLV for each of your customers.
- Divide your customers into 5 buckets by CLV, labeled A through E, each containing 20% of your customers.
- For each bucket, find the average CLV of those customers. Apply your target CLV to CAC ratio to come up with the maximum amount you would spend to acquire similar customers.
- Create a normal (not value-based) custom audience in Facebook for each bucket.
- For each custom audience, create a lookalike audience in Facebook.
- Create multiple tiered acquisition campaigns as follows:
- For bucket A, start a campaign with a bid cap set at the target CAC. Use the lookalike audience for bucket A and exclude your current customers.
- Next, for bucket B, start a campaign with a bid cap set at the target CAC. Use the lookalike audience for bucket B and exclude the lookalike audience from bucket A. Also exclude your current customers.
- For bucket C, do the same as the above. This time, exclude the lookalike audiences from both buckets A and B, and your current customers.
- And so on.
This strategy takes some attention to detail to set up. But, it will maximize the number of customers you acquire. Also, VBB ensures an appropriate level of profitability via your target CLV to CAC ratio.
Here is a Google published case study that achieved remarkable results using this strategy.
Note that the Facebook algorithm for lookalike audiences is not perfect. The CLV you start with for a given bucket may not exactly match the CLV you realize from the customers you acquire using that bucket. Also note that there will be overlap between the audiences, so it is important to set up the audience exclusions in step 6.
Now, Choose Your Strategy
These three strategies provide you with superpowers in the ever-competitive realm of Digital Ads marketing. You may have competitors trying to sell to the same consumers, but you can still win those display ad bids. Be smarter and more targeted in how you set up your audiences and you’ll be all set. The users you acquire will not only convert and make purchases, but will also have a high customer lifetime value.
- Value-Based Lookalike Audiences are a simple way to get started with using the power of CLV.
- High/Low Value Audiences give you more control over the customers you target.
- Value-Based Bidding maximizes the number of customers you acquire while maintaining profitability.