Six way to make your Lookalike Audiences work harder for you

If you have ever bought or set up a Facebook ad, you have probably heard about or used Lookalike Audiences (LALs). LALs s one of the most commonly used Facebook targeting features, and its value proposition is easy to understand: With Lookalike Audiences you are able to reach new people who are likely to be interested in your product or business, since they are similar to your existing customers. 

When creating a Lookalike Audience, you choose a so-called seed (or source) audience, that the LAL will be based on. Facebook then identifies what the users within the seed audience have in common (for example demographics, interests or behaviours) and subsequently deliver your ads to an audience that look similar to your seed. 

Your seed audience can be based on website or app users (through your pixel or SDK), the fans of your Facebook page, or a list of people from your CRM database, for example. 

I think the appeal of using Lookalike Audiences is so strong because it is very simple, yet very clever. That being said, it is not rare for me to come across clients who have disregarded LALs based on past performance, and choose to focus on demographics or interest targeting for their acquisition-campaigns instead. When this happens, I sometimes challenge them not to give up on LALs for reaching new customers just yet, but to try to optimise their Lookalike-strategy first. Many times, the difference in performance by following the six simple tips below can make a big difference. 

Choose the right source for your seed audience 

The source to choose for your seed audience depends on what the objective for your specific campaign is. To create a lookalike based on people who have engaged with your Instagram-content or watched one of your videos on Facebook is a good idea if your campaigns objective is engagement or video views. If you are looking for online sales, however, your seed audience should containing past online customers. 

Make your seed audience high-quality

As mentioned, using previous customers as a seed audience is a good idea, but also consider the quality of those customers. Whether you create a seed audience based on pixel- or CRM-data, instead of including all purchasers in the seed - try to only include the ones that has the highest value to you. Depending on the data you have at hand, a good rule to segment on is:

• Recency: How recently was the purchase made? Someone who bought something from you five years ago might not have much in common with a potential customer of today.

• Frequency: Has the customer made repeat purchases, or was it a one-off sale?

• Value: How much was the basket size of this customer? What kind of products did they buy?

Of course, another excellent way to take value into account is to make use of Facebook Value Based Lookalikes-feature (read more here). You can use data from your SDK, pixel or list with lifetime value as the source for your lookalike, and the lookalike audience will automatically consist of the people similar to your highest value customers. 

Limit the size of your seed audience

This is of course partly related to the previous point, as an audience that is very large most likely contains all customers and not only the really valuable ones. Also, a very large seed audience will probably mean that it is harder to pinpoint what the users within it have in common. A good rule of thumb is for the seed audience to be no larger than 50,000 (but it depends on market and type of business, of course). 

Consider the size of your Lookalike Audience

Another decision to make is how large you want your Lookalike Audience to be. A 1% lookalike will consist of the 1% of the population most similar to your seed audience. What that means, is that a 1% audience will be the most similar/of the highest relevance, while a 10% audience will get you a higher potential reach, but less similarity. Depending on your market, budget, conversion rates and desired reach or amount of conversions - select the audience size that makes the most sense for you. 

Use a nested Lookalike strategy (maybe…)

If you see that your 1% lookalike produces the best performance, but you also want to take advantage of the scale of the larger percentages, a nested lookalike strategy might be for you. What this means, is that you group the total lookalike into different segments (5% and 5-10%, for example). That way, you are able to bid more for the more valuable audience (the 5%, in this case) and less for less valuable audiences (the 5-10%). That being said, if you see no dramatic different in performance and/or are not planning on using different bid strategies for the different segments, there is most likely no major benefits on using such strategy and you will only fragment your audience/budget which might have a negative impact on overall performance. 

Make use of international Lookalikes

I already mentioned international lookalikes in a previous post about cross-border business solutions (read more here). In short, if you are expanding internationally, this feature is a great way to find future customers in new markets by creating a seed audience based on the best customers in your home market.