# Instagram Carousel Ads: 12 Story Structures That Earn the Swipe > 12 Instagram carousel ad structures built on the psychology of curiosity and memory, with real long-running examples and a way to choose the right shape from what your viewer believes. - **Author**: Marxx - **Published**: 2026-10-24 - **Category**: Creative strategy - **URL**: https://marxx.ai/posts/instagram-carousel-ad-ideas --- **What makes a good Instagram carousel ad?** A story shape, not a set of pretty cards. The first card opens a gap the viewer wants closed. Each card closes part of it, and the last card ends on a peak and the offer. Pick the structure from what your viewer believes now: a myth bust for wrong beliefs, offer maths for price doubts, claim then proof for "does it work?" ## A carousel is the only ad that asks for permission to continue Most ads are seen or skipped in a moment. A carousel is different: the viewer decides, card by card, whether to keep going. Every swipe is a small "yes". That makes a carousel closer to a short story than a poster. Like any story, it has a shape. Most carousels that fail don't fail because of the design. They fail because they're **a set of slides, not a story**: five nice cards with no reason to swipe from one to the next. This guide explains why people swipe, using research on curiosity, memory and story shapes, then gives you 12 structures you can use, how to choose between them, and how to test them. Each structure is shown with real long-running carousels from the Marxx library where we have them. Six carousel plot shapes drawn as lines across cards: countdown, reveal, before/after, myth bust, problem ladder, offer maths --- ## Part 1: Why people swipe, and why they stop ### The information gap: Why card one matters most Economist and psychologist George Loewenstein described curiosity as a kind of deprivation. It arises when we notice a **gap** between what we know and what we want to know ([Loewenstein, *Psychological Bulletin*, 1994](https://stafforini.com/works/loewenstein-1994-psychology-curiosity-review/)). The information gap: card one opens a gap, the viewer feels a small itch to know what's missing, and each following card closes part of it while opening the next **How it applies to carousels:** The first card's job isn't to sell. It's to open a gap. "5 mistakes that damage your skin barrier" works because the viewer now wants to know the five, and whether they're making them. **Why it fails:** A gap only pulls if the viewer cares about the topic. "5 facts about our company" opens a gap nobody wants closed. And a gap that's too wide ("Everything about skincare") feels like work, not curiosity. **The design rule:** Each card should close part of the gap *and* open the next one. A carousel where card 2 answers everything has no reason to reach card 3. ### The peak-end rule: Why the last card matters too Psychologists Daniel Kahneman, Barbara Fredrickson and colleagues found that people judge an experience largely by its **most intense moment and its end**, not by the average of every moment (Kahneman et al., *Psychological Science*, 1993). The first and last cards do most of the work: open a gap with the hook, end on a peak with the offer **How it applies:** The last card is what people remember, so it should carry your strongest moment *and* your offer together. A carousel that peaks on card 3 and trails off into a logo wastes its best memory. **Limitations:** The peak-end research measured experiences like pain and pleasure, not ads. Treat it as a useful principle, not a proven ad rule. ### Stories have shapes In 2016, researchers analysed the emotional arcs of **1,327 stories** from Project Gutenberg. They found that most fall into **six basic shapes**, and that some shapes were associated with more downloads ([Reagan et al., *EPJ Data Science*, 2016](https://epjds.epj.org/articles/epjdata/abs/2016/01/13688_2016_Article_93/13688_2016_Article_93.html)). Six emotional arcs mapped to carousels: rise for step by step and offer maths; fall for a problem ladder; fall then rise for before/after; rise then fall for myth busts; rise, fall, rise for a reveal with a twist; fall, rise, fall rarely right for ads The same shapes describe carousels. "Fall then rise", the classic problem and solution, is before/after. "Rise then fall" is a myth bust: the belief, then the reality. Mapping your carousel to a shape before you design it is the fastest way to spot a flat one. > **The quotable version:** A carousel isn't five cards. It's one gap, opened on card one and closed on the last. --- ## Part 2: What long-running carousels actually look like Long-running carousels in October 2026: JYM challenge 779 days, Broadway museum list 200 days, Licious store reveal 193 days, Carro reasons list 81 days, Perfora before/after 79 days, Instacart problem then fix 44 days | Brand | Card one | Structure (our reading) | Live (25 Oct 2026) | |---|---|---|---| | **JYM Supplements** | "Still on single-sourced protein?" | Challenge a habit, then the alternative | 779 days | | **Museum of Broadway** | "Things in the Museum of Broadway that just make sense" | List | 200 days | | **Licious** | "The freshest store in HSR Layout is here!" | Reveal | 193 days | | **Porcellia** | "I grew my D2C brand to 30L in monthly revenue with 0 ad spends." | Claim, then proof | 125-150 days across versions | | **Carro** | "Reasons to sell my car to Carro (besides their best price guaranteed)" | Countdown / reasons list | 81 days | | **Perfora** | Purple whitening strips | Before / after | 79 days | | **Instacart** | "12 people show up to an author's book signing, but your bookstore only planned for 6" | Problem, then fix | 44 days | **What they share:** Every card one is **specific**. A number, a place, a habit, a named situation. None opens with the brand name or a generic benefit. **How to read this honestly:** Lasting a long time is evidence a brand keeps choosing an ad, not proof that the structure caused its success. Treat these as patterns worth testing. Porcellia's carousel, claim then proof, and Wonderla's carousel, problem then offer --- ## Part 3: The 12 structures Each structure below has the same parts: the shape, what it's for, how to build it, and how it fails. ### 1. Countdown **Shape:** "5 mistakes..." from 5 down to 1, with the most important last. **Use it for:** Lists people want to finish. **How it fails:** Putting the best item first. The countdown only works if the peak is at the end. ### 2. Reveal **Shape:** Curiosity builds card by card; the product or result appears on the last card (Licious: a new store, revealed). **Use it for:** Launches, openings and transformations. **How it fails:** A reveal that disappoints. If the last card isn't worth the wait, people feel tricked. ### 3. Before / after **Shape:** The problem, the switch, the result (fall then rise). **Use it for:** Anything visual: skin, teeth, rooms, spreadsheets (Perfora). **How it fails:** Results the product doesn't reliably deliver. That's a trust risk, and in some categories a policy one. ### 4. Myth bust **Shape:** A common belief, then "that's wrong", then why and what to do instead (rise then fall). **Use it for:** Categories full of bad advice. **How it fails:** Busting a myth nobody believes. ### 5. Problem ladder **Shape:** Each card names a slightly bigger version of the problem until the last card offers the fix. **Use it for:** B2B and services, where the cost of the problem isn't obvious. **How it fails:** Climbing too long. Three or four rungs, then relief. ### 6. Offer maths **Shape:** The deal, step by step: one ticket, two, three, and the fourth is free. **Use it for:** Bundles and discounts (Wonderla's "Buy 3, Get 1 Free"). **How it fails:** Maths that's hard to follow. One calculation per card. ### 7. Claim, then proof **Shape:** A bold, specific claim, then the evidence card by card: numbers, screenshots, reviews (Porcellia). **Use it for:** High-trust purchases. **How it fails:** A vague claim ("We help brands grow") gives the proof nothing to prove. ### 8. Problem, then offer **Shape:** A feeling the viewer recognises, then the offer as the answer (Instacart's under-catered book signing). **Use it for:** Leisure, lifestyle and business services. **How it fails:** A problem described from the brand's point of view, not the customer's. ### 9. Day in the life **Shape:** Morning to night, with the product in each card. **Use it for:** Habits and routines. **How it fails:** A product that only appears in one scene. ### 10. Comparison **Shape:** Two columns across the cards: the old way and your way (JYM: single-sourced protein vs the alternative). **Use it for:** Switching campaigns. **How it fails:** Unfair comparisons. Viewers notice, and some claims need substantiating. ### 11. Step by step **Shape:** A how-to, one step per card, with the product in one of the steps (rise). **Use it for:** Anything people need to learn to use. **How it fails:** Too many steps. If it needs ten cards, it's a tutorial, not an ad. ### 12. Reviews wall **Shape:** One customer review per card, in their words, ending with the offer. **Use it for:** Social proof at scale. **How it fails:** Reviews that all say the same thing. Pick reviews that answer *different* doubts. --- ## Part 4: Choose the shape from what the viewer believes The most common mistake is choosing a structure because it looks good. Choose it instead from **where the viewer starts** and **where they need to end**. Choose the shape from the viewer's belief: that's not true leads to a myth bust; not my problem to a problem ladder; too expensive to offer maths; does it work to claim and proof; how do I use it to step by step; who else uses it to a reviews wall | The viewer thinks... | What's really going on | Use | |---|---|---| | "That's not true." | They believe something wrong | Myth bust | | "Not my problem." | They don't see the problem | Problem ladder | | "Too expensive." | They doubt the value | Offer maths | | "Does it actually work?" | They doubt the result | Claim, then proof; before/after | | "How would I even use it?" | They doubt they can | Step by step | | "Who else uses this?" | They need reassurance | Reviews wall | **Then:** 1. **Write the last card first.** It's what you need them to believe at the end, plus the offer. 2. **Write card one.** Open the gap between their current belief and your last card. 3. **Draw the line.** Sketch the tension across the cards. If it's flat, the carousel will be skipped. **Where to get the beliefs:** Your customers' own words in reviews and ad comments. A comment that says "Does this actually work on oily skin?" tells you which structure to use. ### Worked example: A skincare brand picks its shape *Hypothetical example.* A skincare brand's five-card myth bust: your moisturiser isn't the problem; most dryness is a damaged barrier; three signs yours is damaged; what repairs it, ceramides; INR 499 barrier cream with a 30-day promise **The insight:** Comments show customers keep switching moisturisers and still feel dry. The belief is "I need a better moisturiser." **The shape:** A myth bust. 1. "Your moisturiser isn't the problem." *(Opens the gap.)* 2. "Most dryness is a damaged skin barrier." *(Partly closes it, opens "how do I know?")* 3. "3 signs yours is damaged." *(Self-diagnosis keeps them swiping.)* 4. "What repairs it: ceramides." *(The mechanism.)* 5. "INR 499 barrier cream. 30-day promise." *(Peak and offer together.)* **The test:** Run it against a reviews-wall version with the same offer and budget. If the myth bust wins, the belief was real. If the reviews wall wins, customers needed reassurance more than education. --- ## Part 5: How to test and measure carousels ### What to measure | Metric that matters | Why | |---|---| | Cost per result | The outcome you're paying for | | Results by structure | Tells you which *shape* works for your audience, not just which ad | | Card-level clicks, where your reporting shows them | Shows where interest peaks and drops | | Metric that misleads | Why | |---|---| | Swipes or engagement alone | A curious swipe isn't a purchase | | Click-through rate alone | A gap can attract clicks from people who were never going to buy | | Likes on the first card | Card one can be popular and still not lead anywhere | ### How to test well 1. **Test structures, not just designs.** A myth bust against a reviews wall teaches you more than two colour versions of the same carousel. 2. **Keep the offer and budget the same,** so the structure is the only difference. 3. **Keep the card count close.** Meta allows **2 to 10** cards ([AdManage summary of Meta's specs](https://admanage.ai/blog/meta-carousel-ad-specs)), but each extra card is a chance to lose people. Use as many as the shape needs, usually three to six. 4. **Change one card at a time** once you have a winner, starting with card one. Marxx trending tactics showing hooks and formats rising and declining in a category --- ## Six mistakes, and how to spot them | Mistake | Symptom | Fix | |---|---|---| | Slides, not a story | Engagement drops after card one | Draw the tension line; give each card a reason to swipe | | A generic first card | Low swipes from the start | Open a specific gap: a number, a habit, a situation | | The peak in the middle | Strong card 3, weak finish | Move the strongest moment to the last card | | The offer hidden early | People swipe past it and forget it | Put the offer on the last card, with the peak | | Choosing the shape by looks | Nice design, flat results | Choose from the viewer's current belief | | Too many cards | Long carousels underperform | Three to six, as the shape needs | ## What to take into your next creative review - **A carousel is one gap, opened on card one and closed on the last.** No gap, no swipe. - **The last card is what people remember.** Put the peak and the offer there together. - **Stories have shapes.** Draw yours before you design anything. - **Choose the structure from the viewer's belief,** not from what looks good. - **Test structures against each other.** Two shapes teach you more than two colour schemes. ## FAQ **How many cards should an Instagram carousel ad have?** Meta allows 2 to 10. Use as many as the story needs, usually three to six. Every extra card is a chance to lose the viewer. **What should the first card of a carousel do?** Open a specific gap: a number, a habit, a named situation or a bold claim that the following cards resolve. **What should the last card do?** Carry your strongest moment and the offer together. People remember the peak and the end. **Do carousels work better than single images?** It depends on the story. Use a carousel when the message needs more than one step to land, such as a proof, a comparison or a process. **What are good Instagram carousel ad ideas?** Countdowns, reveals, before/afters, myth busts, problem ladders, offer maths, claim and proof, problem then offer, day in the life, comparisons, step by step guides and reviews walls. --- ## Turn a story shape into finished ads **Trial the Marxx generation canvas:** Pick the structure, and get card-by-card copy, visuals and every placement from one brief. Marxx write node drafting copy from a brief **[Trial the canvas: Turn one brief into a finished ad ->](https://www.marxx.ai)** *Related: [Best Carousel Ads Examples](/blog/best-carousel-ads-examples) - [Visual Ad Creative Principles](/blog/visual-ad-creative-principles) - [Buyer Persona Statistics](/blog/buyer-persona-statistics)* ### Sources - Loewenstein, G. (1994). [The psychology of curiosity: A review and reinterpretation](https://stafforini.com/works/loewenstein-1994-psychology-curiosity-review/). *Psychological Bulletin*, 116(1). - Kahneman, D., Fredrickson, B., Schreiber, C. & Redelmeier, D. (1993). When more pain is preferred to less: Adding a better end. *Psychological Science*, 4(6). - Reagan, A. et al. (2016). [The emotional arcs of stories are dominated by six basic shapes](https://epjds.epj.org/articles/epjdata/abs/2016/01/13688_2016_Article_93/13688_2016_Article_93.html). *EPJ Data Science*. - AdManage. [Meta carousel ad specs](https://admanage.ai/blog/meta-carousel-ad-specs). - Marxx ad library, 25 October 2026. Days live were accurate at the time of writing. --- - [More Creative strategy articles](https://marxx.ai/posts/category/creative-strategy) - [All articles](https://marxx.ai/posts)