The Secrets of Human Behavior Applied to Social Ads

Struggling to create Facebook, Instagram, or TikTok ads that people remember long enough to buy from you? Wondering how consumer psychology can tell you why your winning ads work and how to repeat that success?

In this article, you'll discover how to use behavioral science and consumer psychology to create social ads by deciding how to deliver a message, who to aim it at, and where it fits in the customer journey.

Why Social Ad Creative Needs to Build Recall, Not Just Conversions

Sarah Levinger says the biggest misconception about ads is that every ad is supposed to sell. Conversion matters, especially in e-commerce, but for most brands, the core job of advertising is to build recall. A customer who doesn't remember a brand exists has nowhere to convert to and ends up buying whatever shows up in their feed. Sarah recommends designing ads to target both recall and conversion simultaneously.

​Behavioral science makes that possible because it explains why a piece of creative works. When marketers can identify the psychological reason an ad or organic post resonated, they can replicate that success rather than guessing. Sarah sees the results reflected in lower-cost customers, longer-lasting ads, access to audiences a brand has never reached before, and less decision fatigue about what to create next.

​The way ads get distributed raises the stakes. TikTok built its feed on interest graphs, and the approach now extends across the major platforms. An interest graph serves content based on whether the platform predicts someone will be interested in it, rather than on whether they follow the creator or have searched for a term. The algorithm learns from behavior such as watching, liking, commenting, sharing, and saving, and it updates in real-time as moods shift. Sarah's own feed serves intense content during anxious weeks and cooking and sewing tutorials on relaxed Saturdays. Because marketers can't dictate who sees their content, every concept, organic or paid, has to be built for interest and taste rather than topics and keywords.

​Sarah's process for building that kind of creative follows three decisions in order: how the message is delivered, who it targets, and where it fits in the customer journey. All three start from a defined core message. If a brand like Nike has a core message of pushing through and doing it, the three steps determine how that message gets expressed in each ad.

#1: Set the Emotional Delivery of Each Ad With the Valence and Intensity Model

The first decision is whether the message should be positive or negative and how intense it should feel. Sarah makes this call using the valence and intensity model, also called the valence and arousal model.

​Valence is the emotional range a message activates, from positive (happy, comforting) to negative (sad, frustrating, disappointing). It's a range, not a binary. Intensity is how strongly the message is delivered, and it affects how the message's valence is interpreted and how important it feels. 

A reminder to take out the trash can be positive and low-intensity (“I'm so sorry, I need your help with this, I completely forgot”) or negative and high-intensity (“You've missed the trash five times. I need you to step it up”). Each version prompts a different response depending on the listener's mood.

These choices matter because emotional messages are encoded in the brain more strongly than logical, factual ones. The chemicals the brain produces in response to an emotional experience, positive or negative, tell it what to remember and what to ignore. That's why emotion drives recall.

How to Choose Valence Based on Category Psychology

Sarah sets valence based on category psychology: what the product is and what's being sold. Some categories respond best to negative messaging. A carbon monoxide detector or fire alarm ad that promises “protect your family, and everything's going to be great” won't land. The negative version is far more salient: there's a silent, invisible danger in the home, and without a detector, people risk harming their own family by accident.

How to Set Intensity Based on Brand Identity

Intensity is a brand decision rather than a category decision. Brands whose positioning is supportive, essentially “we're not trying to trigger you; we just want to help you feel better,” should stay in low intensity all year. Sarah says those brands rarely change intensity, and it works because that's how the brand is interpreted.

Other brands are primed for intense messaging. Liquid Death's core message is that its customers are misfits who don't subscribe to corporate culture, and it delivers that message through intense sarcasm. A brand with a rebellious, funny, or provocative identity can push intensity higher than a wellness brand can.

​Pro Tip: Single-word swaps can shift valence and intensity enough to move revenue. For about 18 months, the watch brand Original Grain ran ads telling customers they “deserved” to treat themselves to a watch. Sarah's research showed customers didn't feel they deserved anything; they felt they'd earned it. Changing “deserve” to “earned” across all of the brand's messaging was a six-figure change. “Deserve” is passive. “Earned” is active.

How to Audit Ads and Reviews for Valence and Intensity With AI

Sarah runs this analysis in AI. She usually uses Claude, but Gemini, Perplexity, or any other LLM works because valence and intensity are a standard matrix with four zones that AI tools already understand. Her process:

1. Gather Customer Reviews: Sarah pulls the full range of reviews, from five stars down to one star. Most reviews skew positive, so including the negative ones gives a more accurate picture of how customers feel.
   
   
2. Gather Current Messaging: Next, she collects a good batch of current ads or organic posts.
   
   
3. Ask AI to Categorize Each Batch: The AI sorts every review and every message into valence and intensity zones.
   
   
4. Find What the Winners Share: The key question is what the top-performing ads and posts have in common. If the winners always sit in zone two or typically land in zone three, that pattern guides new creative.
   
   
5. Compare Reviews Against Messaging: Comparing the review data set with the messaging data set shows whether the brand's language is activating what the brand thinks it's activating and whether it matches what customers need in terms of emotional response and intensity.

#2: Target the Right Version of the Customer With the Three Selves Model

The second decision is who the message targets. Sarah doesn't use customer avatars. She calls them an overused crutch: a fabricated customer who doesn't exist, used to make large business decisions. An e-commerce brand might sell to an 18-year-old and a 57-year-old, people of every gender and political affiliation, and people who own five dogs alongside people who hate dogs. One avatar can't describe that.

Instead, Sarah targets a version of the self. Her model is based on self-discrepancy theory, which holds that every person has three versions of themselves and switches between them throughout the day. Each one is tied to a different time period.

​Ideal Self: The future-focused, aspirational self. Messaging about a better version of the customer, doing better tomorrow, or changing their life targets the ideal self.

​Actual Self: The present-day self, which is biased and wants its existing beliefs validated. Actual-self messaging doesn't ask anyone to change; it affirms who they already are. Sarah often sees eco-friendly and environmentally focused brands activate the actual self.

​Ought Self: The past-focused, guilt-driven self that thinks, “I should have fixed this, but I didn't.” It shows up often in parenting and baby brands because of the pressure parents feel. The ought self responds best to lower-intensity, slightly negative messages, especially when someone has been sitting on a chronic problem or is sensitive to messages about the person they should be but aren't.

How to Map Brand and Competitor Messaging to the Three Selves With AI

The three-self audit follows the same pattern as the valence and intensity audit. LLMs are built to analyze language, so this is fast, and Sarah has tested it in Portuguese, Dutch, German, French, and Spanish.

• Collect Current Messaging: Sarah gathers a batch of ads plus organic posts, landing pages, and emails.
  
  
• Ask AI to Categorize by Self: Claude or another LLM labels each message as targeting the ideal self, actual self, or ought self and reports which one is most common.
  
  
• Run the Same Analysis on Competitors: The same prompt goes to competitors' landing pages, emails, ads, and organic content.
  
  
• Look for Overlap: If a brand and its competitors are both talking to the same self, they're saying the same thing to the same person, and recall drops. That makes it hard to compete long term.
  
  
• Shift to an Open Self: When the market is crowded on one self, Sarah moves messaging elsewhere. A brand stuck in ideal-self messaging alongside its competitors might shift this month toward the actual self or run more ought-self messages.

The ad and its landing page don't need to target the same self, but the landing page must match what the person wants most. 

An ought-self ad for soccer cleats (“soccer season is coming and you need to get on this”) can lead to an ideal-self landing page (“be the star soccer mom this year”). A parent buying cleats runs two states at once: resolving the guilt of not having bought them yet and wanting to be the parent whose kid comes home saying the whole team loved the shoes. The ideal-self page solves that secondary need.

How to Write One Message for Each Self

Before writing, Sarah decides where the message will run (landing pages, emails, or organic channels) and what it needs to accomplish there: conversions, customer activation, or building recall. 

The goal determines how she combines valence, intensity, and the three selves. Nike shows how one core message can flex this way. “Just Do It” has run since the 1980s, but Nike often delivers it as “motivation comes from within,” a secondary message with a different valence and intensity.

To see how the self changes a single ad, here's Sarah's carbon monoxide detector message written for each self, all with negative valence and a silent-killer framing.

Ought-Self Version: “This has been on your to-do list for years, and you really should have gotten it done. It's your duty to make sure your home is safe for your kids.”

​Actual-Self Version: “You're the type of person who organizes your home around the technology in it, and you're good at spotting detectors that won't work for you. This one has all the specs you need. Buy it today.” This version barely touches the danger. It activates the buyer's identity as someone who's good with tech and takes care of their home.

​Ideal-Self Version: “You don't want to deal with this in the future. Set it and forget it. Peace of mind is what you're looking for.” This version frames the purchase as something the buyer's future self will be grateful for.

#3: Assign Each Ad to a Stage of the Customer Journey Using Google's Messy Middle

The third decision is where each piece of creative fits in the customer journey. Sarah doesn't use awareness levels because they don't tell marketers what to make, and awareness doesn't equal buying intent. Sarah is fully aware of Coca-Cola but drinks soda two or three times a year.

​Instead, she maps creative to the Messy Middle, a Google study of how people shop in digital ecosystems. It identified four behaviors people move through when they purchase: trigger, exploration, evaluation, and purchase. Every piece of content fits into one of the four.

How to Create Content for Each Messy Middle Stage

Trigger Content: Something in a person's environment prompts them to notice a problem and seek a solution. The problem is either acute (sudden, unexpected, and irritating, like a burst pipe) or chronic (something they've avoided for a long time). Trigger content should speak to one of those two problem types and address the person's experience at a high level rather than the product. 

Point-of-view content is one of the best formats here (“Your house just flooded because a pipe burst. What are you going to do? You're going to call Hank,” followed by the pitch), and day-in-the-life content also works (a stay-at-home mom trying to get the plumbing fixed before school pickup at 3 p.m.). Both rely on storytelling about the trigger moment.

​Exploration Content: In this stage, people explore the industry, not the brand. They're asking what creatine is, how protein works, or whether hydroxyapatite matters in toothpaste. Across the more than 150 brands Sarah has worked with, the exploration content that performs best covers what to avoid (“don't believe the hype about this product”) or things people don't know. 

Long-form, documentary-style content in the vein of How It's Made fits here. Deep researchers can stay in this stage for a long time, so skipping exploration content overlooks a large portion of the potential audience. Sarah says this is where organic content does critical work for e-commerce and DTC brands that otherwise lean on ads, because exploration is where people decide whether they like a brand.

​Evaluation Content: Shoppers compare options, usually two. This content should deliver facts, proof, stats, numbers, ingredients, and how those ingredients work, all in service of the brand's unique selling proposition (USP). A brand that doesn't clearly differ from competitors has to win on personality by being interesting, fun, or likable. 

Duolingo competes with many language apps and wins evaluations largely through content built around its characters and mascot. Sarah expects AI to change this stage, since shoppers who use AI to evaluate products may compare more than two options.

​Purchase Content: The shopper is waiting for an excuse to buy from a specific brand. Purchase-stage ads give them that excuse.

How to Allocate Creative Across the Four Stages

Sarah starts with a fairly even split of creative across trigger, exploration, evaluation, and purchase, then shifts the mix based on the goal:

• When a brand needs more eyeballs, she swaps most of its creative to trigger-stage content to activate people who are just coming in.
  
  
• When a brand needs to move people through the middle of the journey, she runs more exploration-style and evaluation-style ads.
  
  
• When a brand needs straight conversion revenue, she concentrates on purchase-level ads.

How to Build and Map a Content Library With AI

To start applying the Messy Middle, Sarah recommends building a swipe library, which is how she began.

• Save Content That Stops the Scroll: Any ad or post with an interesting hook, a good story, or the ability to stop the scroll goes into the library.
  
  
• Review on a Schedule: Once a week or once a month, Sarah runs the library through AI.
  
  
• Ask AI to Map Each Piece: Using the same approach as the valence and intensity and three-self audits, the AI labels each piece of content as trigger, exploration, evaluation, or purchase based on the customer behavior it serves.

Other Notes From This Episode

• Connect with Michael Stelzner @Stelzner on Instagram and @Mike_Stelzner on X.
• Watch this interview and other exclusive content from Social Media Examiner on YouTube.

Where to subscribe: Apple Podcasts | Spotify | YouTube Music | YouTube | Amazon Music | RSS

✋🏽 If you enjoyed this episode of the Social Media Marketing podcast, please head over to Apple Podcasts, leave a rating, write a review, and subscribe.

---

Stay Up-to-Date: Get New Marketing Articles Delivered to You!

Don't miss out on upcoming social media marketing insights and strategies! Sign up to receive notifications when we publish new articles on Social Media Examiner. Our expertly crafted content will help you stay ahead of the curve and drive results for your business. Click the link below to sign up now and receive our annual report!