October 20, 2023

Use Generative AI for Facebook and Instagram Ads to Combat Fatigue


This post on Generative AI for Facebook and Instagram ads was published on February 17, 2023 and updated on October 20, 2023 date. 

Have you noticed your Facebook and Instagram ads drop in performance soon after you put a new campaign on the platform? You’re definitely not alone. The Meta platforms still have the largest audiences in social media and some of the most consistent performance. A crowded platform and distracted customers can consistently erode ad performance, however. Enterprise brands have a solution now using Generative AI for Facebook and Instagram ads.

How Can Marketers Use Generative AI on Facebook and Instagram?

Meta is counted among the large technology companies investing billions in developing Large Language Models (LLMs) to deliver Generative AI capabilities. The latest version of Meta’s LLM, Llama 2, launched in July of 2023. Meta has made Llama 2 available as an open source model. Its motivations include catching up with GPT-4, which has a significant head start in LLM development, and hopefully benefiting from crowdsourced innovation as other organizations use it and find ways to improve its core functionality.

Meta’s core LLM model is only one piece of its GenAI strategy. Inside the Meta platforms, the company has also launched new functionality in Ads Manager. Now marketers can leverage Generative AI to perform background generation, image expansion, and text variations.

This appears to be one of multiple steps Meta has taken to reverse the downward trend in ad revenue seen for multiple quarters in 2022. In response, Meta has also added a number of services to increase platform engagement. These include Broadcasting Channels and the immensely popular Threads (Meta’s version of Twitter/X) on Instagram. Reels also received a revamp for both platforms, with significant enhancements for advertisers. Those and other actions seem to have increased ad revenue in 2023. The new GenAI capabilities are intended to hold onto that growth by making it easier to build and revise ads in the platform.

These enhancements improve marketing efficiency by enabling brands to develop more ads faster. They do not automatically improve ad effectiveness, however—Generative AI can only do that if it is trained using marketing language proven to perform. As far as we know, Persado Motivation AI is the only enterprise Generative AI with that capability, since its language model was developed using 10+ years of enterprise marketing campaigns tested with real customers to identify the language that motivates customer action.

Before we explore how brands can increase ad performance leveraging Generative AI for Facebook and Instagram ads, let’s dig into the current challenges marketers face on the Meta platforms and suggested solutions.

Ad performance challenges and solutions on Facebook and Instagram

There are a number of systemic reasons why many advertisers are seeing lower engagement at higher costs with ads on the Meta platforms. One is that consumers have been choosier about their spending in response to the mixed economic signals. Advertisers also are diversifying their ad spend in response to younger demographics taking their activity to alternative platforms—such as TikTok for Gen Z specifically. Also, the continued loss of cookies coupled with the increased privacy protections for iPhone users with iOS14 have decreased Meta’s ability to target ads to the right audience. 

Beyond these industry-wide issues lies your ad creative. Ad performance can degrade because the creative text and images fail to engage the customer. To combat ad fatigue and tackle some of these issues, marketers can turn to Generative AI for Facebook and Instagram ads.

Here’s how:

Challenge: Demographics and Meta ad performance

The audience demographics on Facebook and Instagram is one well-documented factor that can affect ad performance. Adults between the ages of 25 and 34 are Facebook’s largest audience segment. In contrast, Instagram has about an equal share of adults in the 18-24 age group as between the 25-34 age group. Many advertisers looking to speak to segments within these demographic groups frame their ads to appeal to them. Yet pressure from regulators may make it harder for Meta to enable effective ad targeting. As a result, ads geared to specific demographics or customer groups may not reach the right people.

A related issue is ad reach on the platform, or how many customers see your ads on Facebook or Instagram. Meta reporting highlights the fluctuations in active users and new account openings. Some analysis suggest those numbers are rising less consistently than in the past. They may even be declining as the platforms get better at identifying false or duplicate accounts or those that violate platform terms.

Solution: Speak to the right audience

Every marketer knows to start with the audience. But, in today’s market of many options and fast-moving behavior, “the audience” may change faster than you sense it. Former brand advocates may age or stage out of your product, switch platforms, and/or re-prioritize their spending. Any and all of these factors can affect how well your ads perform.

The takeaway is for brands to make sure they have updated insights about the segments and sentiments they speak to. Demographic data is rarely enough. You need to go deeper and then develop dedicated ad messages to ensure you are engaging key audiences.

Challenge: Customer engagement with Meta ads

Demographic targeting is only one issue affecting how well ads reach millions of Facebook and Instagram users. Perhaps a bigger factor is, are they listening? Do they want to see your ads on Facebook and Instagram?

Research into the different engagement rates by industry suggest wide variations in the answer. Brands in the retail and apparel industries see some of the highest click-through rates on the platform. But the highest performance goes to legal advertisers. And the lowest performers? Ads for employment and job training. Maybe they’d have more luck on LinkedIn?

Solution: Refresh and optimize your creative using Generative AI for Facebook and Instagram ads

Making multiple ad creative options was cost prohibitive in the past, when all creative depended entirely on human creators. Using Generative AI for Facebook and Instagram ads can allow brands to cost-effectively develop multiple creative options for the same campaign.

Meta’s built in Ads Manager functionality offers one way to do that. Brands that want more certainty that their ad updates will perform better should leverage a GenAI model optimized for marketing campaign performance. Persado’s knowledge base is trained on the enterprise marketing language of scores of Fortune 500 retailers, apparel companies, travel firms, and financial services institutions, among others. Informed by the equivalent of 640+ years of digital A/B testing showing what works to motivate customer engagement, Persado-generated language outperforms human-written control copy by an average of 40%.

Marketers can also explore complementary tools for different aspects of their advertising creative on Facebook and Instagram. Omnecky leverages OpenAI’s DALLE-2 and GPT models to generate visuals for social media ads. This platform also delivers performance insights by analyzing the success of the images in their customer’s social media feeds. Using collected data in addition to the client’s brand guidelines, Omnecky produces dozens of AI-generated images that can boost variation and performance of online ads. Jobscan, a SaaS business, saw the benefits first-hand when it used Omnecky to identify which visuals correlated with high ad performance. Those insights allowed Jobscan to decrease costs by 30% and increase conversions 2x for their Facebook ads.

Challenge: Ad fatigue

Marketers need new creative to combat ad fatigue. This can manifest as banner blindness, which happens when customers have seen an ad so many times it fails to register. They literally can’t see it when it goes by because the brain tells them they don’t need to pay attention.

One reason this happens is excessive exposure to the same creative copy and visuals. Even people who like your brand will tune out if they aren’t in the market for what you’re offering.

Facebook’s ad platform will stop showing your ads to audiences after a certain number of views. Having alternative creative options allows you to swap out underperforming versions and replace them with new ones.

Solution: Personalize the message using Generative AI for Facebook and Instagram

The next stage along the progression to improve Meta platform ad performance is to personalize messages. There’s plenty of research showing that consumers develop a better view of brands that engage with them in a way that is informed about their needs and experiences, without crossing over into intrusive or creepy. In fact, the majority of consumers say it’s annoying when brands don’t personalize their messaging.

There are many levels to personalization. The Meta platforms have long allowed advertisers to choose the audiences for their ads, for example. Cookie degradation coupled with increased security features in iOS14 makes that harder for them to do, however. That’s one of the reasons advertisers have noticed higher costs per click.

There remain a number of underutilized personalization levers brands can use, however, that remain in your control. Personalizing how you position and focus the message is one of them. Typically, marketers think of personalization in terms of the offer—what product will they promote with what incentive. But taking the same offer and leveraging more effective language can also have an outsized impact on engagement. 

Persado saw that impact with a Facebook loyalty credit card ad campaign. We developed multiple versions to appeal to different audience segments. For example, the credit card company knew that some loyal customers like to use their points to pay for travel. Others like cash back. Persado generated personalized language associated with these offers for the different audiences and tested their impact compared with a non-personalized alternative. The personalized messages outperformed the generic message by 60% for the travelers and 44% for the cash-back customers, respectively.

The Meta platforms Facebook and Instagram aren’t going anywhere. They will be a key part of the advertising mix for the foreseeable future. As a result, marketers need all the levers they can use to combat social media ad fatigue and optimize performance. Make sure AI-generated language that is specialized for marketers is one of them.

Learn more about the results you can achieve using Persado to deliver the right message in our CMO’s Guide to Achieving Impact With Generative AI.

Related Articles