Marketers can take advantage of first-party data in paid media campaigns by learning how to use lookalike audiences on platforms. Lookalike audiences populate a larger group of potential customers based on your own uploaded target audience. This increases your reach and ability to target audiences of similar demographics and interests.
Marketers around the country spent the last few months preparing for an increased need for first-party data. This effort is driven by consumers’ increasing demand to have control over their own data, and major industry players’ recent shifts towards privacy for their users. Individuals want the ability to know who is seeing their online activity, and provide permission for them to do so. It’s now easier than ever to give that power back to the user.
In particular, Apple decided to ask consumers if they’d like to opt-in to in-app behavior tracking in their new iOS 14 update (and subsequent 14 iterations). Since iPhones and Mac computers are prolific, major advertising platforms were forced to adjust their own data-collection capabilities to adhere to this new privacy standard.
The most notable adjustment for marketers is the decreased visibility into user demographic information. There’s no longer a way to pinpoint particular users and follow their online journeys. Instead, broad targeting of potential customers will now rely on “lookalike audiences.”
What is a lookalike audience?
A lookalike audience is a machine-learning-generated audience built on first-party data gathered from your existing target audience lists. Lookalike audiences on Facebook, LinkedIn, and Google differ slightly, but the concept remains the same: An audience of users similar to the people you’ve targeted in the past.
Using a lookalike audience provides a larger pool of individuals who share similar interests or browsing behavior as your existing audience. For instance, if you want to advertise a running shoe, uploading a list of shoe-purchasing customers would help the algorithms serve ads to individuals who may have similar interests or behaved similarly in the past.
How do lookalike audiences work?
At the very simplest level, you upload your own first-party user data into the platform of your choice. For Google, they then apply their proprietary methodology to filter out unusable individuals to match with their broader database. If enough users make it through filters, they’ll create an audience of people similar enough to the uploaded data to provide a promising group. Because of privacy updates like Apple’s App Tracking Transparency, some of the factors in creating a lookalike audience are more difficult than before.
Google’s methodology also provides a great example of how to use lookalike audiences.
You must upload at least 1,000 contacts. Then, Google filters out anyone who isn’t a Google user, then removes anyone flagged as a violation of personal advertising policies. Finally, it now removes anyone who opted out of tracking or ad personalization.
As the final layer of filtering, if any remaining individuals aren’t daily active users, they’re removed. After applying these layers of qualification, a 1,500-person list might only yield 100 users. With such a small sample size, the list is rendered useless and won’t provide any data.
A complicating factor is the match rate. While in the example above, the final product didn’t produce enough data for a similar audience. Yet, this example may produce a 90% match rate. This is because 1,500 of the users on the list qualified as usable, but after Google’s aggressive filtering, only a slim amount of the 90% remained. However, Google itself has less information available to them with new privacy controls in place, so match rates will start to fall and become a larger problem.
This methodology is similar across other popular platforms, like Facebook and LinkedIn. For Facebook, you can use the data collected from a pixel already active in a campaign. Otherwise, you can upload a list of your own, similar to Google’s process.
Because of LinkedIn’s reputation as a professional gathering platform, keyword search terms don’t fact in at all. Much like the other requirements, users upload a list of their own to generate a lookalike audience up to 15 times larger than the original source. Demographics are a particular area of change for LinkedIn’s lookalike audience. Factors like company attributes and personal characteristics aid the algorithm. However, they will not use sensitive information. For instance, you cannot create an audience exclusively of female professionals with “manager” in their titles.
Why marketers should learn how to use lookalike audiences
With marketers soon to have fewer targeting capabilities on social media platforms, lookalike audiences will be key to expanding your reach beyond site visitors. This is one feature that will likely remain intact once third-party cookies are completely deprecated. Even without visibility into individuals’ activity across platforms, lookalike audiences will provide a similar set of prospective customers.
However, the limitations of iOS privacy changes put marketers’ focus on new ways to advertise to target customers. While you’ll still be able to account for ad fatigue controls, you’ll want to refresh your creative faster than ever to better ensure you’re appealing to a broader selection of interests.
Additionally, the original source of data needs to be robust. A first-party data strategy is critical in ensuring you’ll be able to provide enough volume of contacts to use a lookalike audience. The more contacts you upload, the more likely it will survive the multiple layers of privacy-focused filtering.
Once you’ve fulfilled those conditions, you’ll have a lookalike audience operating similarly to campaigns of the past.
Remember to build your campaigns and creative around the attributes used in creating the lookalike audiences. If you’re targeting individuals who make vegan-oriented searches, don’t serve content outside of that dietary parameter.
Above all, lookalike audiences work to give you more people to reach outside of your owned contacts. Use them to your advantage by creating compelling creative and helpful content to fulfill the search queries or keywords they use.
Learning how to use lookalike audiences will be critical in an age of more privacy and less information for marketers. There won’t be opportunities to track particular site visitors’ movements through different sites to retarget. By using the best information you can collect with a first-party data strategy, you increase your potential to use a lookalike audience full of truly qualified people. However, without enough of your own quality data, lookalike audiences could be hard to come by.
Silverback Strategies is offering a first-party data consultation to help you understand your data strategy. We’ll explain how you can create strong lists for lookalike audiences. Schedule yours today to use lookalike audiences with confidence.