November 3, 2023

Engage with the machine: top machine-learning hacks for brands

Andrew Kanaly

Paid Social Strategist

Meta’s push for machine learning promises easier campaign set-up, more efficient results, and greater audience segmentation for brands. Despite this, many in the advertising world remain skeptical of using AI and machine learning in their marketing strategy. The good news? Implementing these strategies is far from scary or difficult. That’s why we crafted this handy cheatsheet to help brands – big and small – use Meta’s new tools to boost efficiency and fuel even bolder campaigns.

What is machine learning?

To put it simply, machine learning systems use data points provided by users to learn organically. For example, If you’re a person who enjoys window shopping through Instagram ads but doesn’t actually buy anything, Meta’s machine learning technology will put you in a similar pool of users. This group will likely be served ad campaigns optimized for landing page views or website visits. A user’s on-platform behavior is just one data point that the system takes into consideration when giving you a consumer classification. Additionally, data gathered through tracking pixels and third-party measurement tools provides information from outside Meta-based platforms. In other words, machine learning helps us media folks work smarter, not harder.

No cookies? No problem.

The digital landscape is always transforming, and brands shouldn’t feel like they’re grasping at straws when dealing with industry-wide changes. Major shake-ups like Apple’s iOS 14 update and its ability for users to opt out of tracking may have you wondering if you’ll have a job tomorrow. All of a sudden, users opted out of those golden nuggets of information used by advertisers. Meta’s solution to this is their Conversions API tool (CAPI). Traditional tracking methods stemmed from browser-side tracking, which relied heavily on the soon-to-be-phased-out cookie. Meta’s workaround is based on tracking events on the server side. Instead of relying on a Google Chrome, Safari, or Windows Explorer cookie jar, all events can be tracked via your website’s server. Best of all, it’s consumer privacy compliant

Optimize without heartburn

Meta also offers their Advantage+ Shopping Campaign (ASC), a system that uses privacy-aware machine learning designed after iOS 14. Advantage+ takes into account user behaviors and preferences to deliver effective in-platform shopping experiences. Adjustments for new customer prospecting and current customer reacquisition are also possible with ASC. Advertisers can rejoice knowing that the system accounts for scheduled budget and creative changes during a campaign. Gone are the days of hopping onto Meta Ads Manager just to activate a single ad and increase your budget by $50. 


Audience targeting just got awesome

Most businesses have an idea of who their ideal customer is. Maybe it’s 25–34-year-old male Deadheads who are into cooking and the outdoors, or maybe it’s 50–65+-year-old women who love westerns. Either way, Meta’s newest push is to keep your audience targeting as broad as possible (within reason). To all the advertisers reading this, put down your pitchforks. Under Meta’s new plan, the best strategy is to cater creative to your ideal target audience. For example, if your goal is to target gamers, new-tech enthusiasts, and engineering students, you should create three separate creative assets geared toward these audiences. Then, you’d put them into one, broad ad group rather than three segmented groups. Meta’s machine learning algorithms will determine who interacts with which ad. Consequently, relevant ads will be shown to the right people all while maintaining scale and cost efficiencies. 

What now?

Let’s be honest, the words “artificial intelligence” and “machine learning” can be intimidating, but if you know enough to be dangerous, your marketing efforts will be endlessly more efficient and effective. There’s no need to adopt all these fancy new tools at once, it’s all about finding what’s right for you. For more ways to stay up to date on Meta’s Machine Learning Suite, check out their business news site. Oh, and be sure to subscribe to Question Everything.

You're in! No secret handshake required.
Oops! Something went wrong.

Feel that? That’s curiosity.

Let’s solve something together.