How many slides should an Instagram carousel have? What 655,385 carousels say
We analyzed 655,385 Instagram carousels from 12,159 creators. More slides win, long captions help, timing barely matters, and the carousel-vs-photo war is a wash.
We analyzed 655,385 Instagram carousels published by 12,159 creators over the last 12 months. The clearest finding: more slides win. Carousels with 11 to 20 slides beat their author's usual numbers by 2x roughly 30 to 32% of the time, versus 23 to 25% for carousels with 2 to 4 slides, and the climb between those endpoints is almost perfectly steady. That is the biggest per-post lever we measured, larger than captions, hashtags, and timing combined. The honest surprise runs the other way: measured against the same account's own baseline, carousels and single photos break out at nearly the same rate. The format is not the edge. What you do inside it is.
Best-practice guides mostly repeat each other on carousels: post them because they get more engagement, keep them under 10 slides, stack hashtags. Almost none of it comes from data, because almost nobody has posting data across thousands of creators. Eden indexes Instagram alongside five other platforms for its Discover research feed, which is the corpus this study runs on. This is the same index behind our Substack timing study and our YouTube thumbnail study, pointed at a different question. Here is everything we found, including the parts that contradict the usual advice.
Key findings
- Carousels with 14 slides break out 32.1% of the time; carousels with 3 slides, 23.4%. The lift is nearly monotonic from 3 slides upward.
- The median 18-to-20-slide carousel performs 13 to 15% above its author's baseline; the median 2-to-7-slide carousel sits exactly at it.
- Only 16.7% of carousels use 11 or more slides, so the highest-performing length is also the least crowded.
- The slide effect holds at every account size: under 10,000 followers, 11-to-20-slide carousels break out 23.5% of the time versus 18.0% for 2-to-4-slide ones.
- Within the same creator's account, carousels break out 28.1% of the time and single photos 27.6%. The format switch alone is worth about half a point.
- Carousels earn 1.8 comments per 100 likes at the median, 50% more than single photos (1.2).
- Captions of 400+ characters break out 28.5% of the time versus 25.4% for captions under 50 characters.
- Hashtags barely move outcomes (25.4% to 28.0% across every band), and 65% of carousels use none at all.
- Timing is the weakest lever on the board: the best posting hour beats the worst by 3.4 points, and weekends beat midweek by about 2.
A breakout, throughout this study, is a post that earns at least 2x its own author's median. The full method is below.
How we ran it
- Sample: 655,385 Instagram carousels from 12,159 creators tracked in Eden's index, published between August 2025 and August 2026. Every post was at least 7 days old at measurement time, so like counts had settled. The median carousel in the sample earned 523 likes; the top 10% crossed 16,000.
- Metric: each carousel's like count measured against the same author's median feed-post performance. We call a post a breakout when it earns at least 2x its author's median. Measuring against each author's own baseline means a 3,000-follower account's hit counts the same as a 3,000,000-follower account's, and it prevents a few giant accounts' habits from becoming "what works."
- The comparison is honest where it matters: our pipeline grades carousels and single photos against one pooled per-author baseline, so when we say the two formats break out at the same rate, they were judged by the same yardstick. Reels are graded on views against their own reel baseline, so reel comparisons are "beats its own baseline" only, never raw numbers.
- Slide counts come from the full ordered slide set our index mirrors for each carousel, covering 99.2% of the sample. We validated those counts against a second metadata source on a random 5,000-post sample: 87.4% matched exactly and 92.8% within one slide, so read any single row plus-or-minus a little; the trend across the table is far larger than the noise.
- Timezone: all times are US Eastern (ET), the same convention as our Substack study. Instagram does not expose author timezones.
How many slides should a carousel have?
The question this study exists to answer. Breakout rate by slide count:
| Slides | Carousels | Breakout rate | Median vs own baseline |
|---|---|---|---|
| 2 | 65,554 | 25.1% | 1.00 |
| 3 | 54,599 | 23.4% | 1.00 |
| 4 | 50,854 | 24.0% | 1.00 |
| 5 | 107,221 | 26.8% | 1.01 |
| 6 | 56,522 | 26.4% | 1.00 |
| 7 | 57,201 | 26.9% | 1.00 |
| 8 | 50,720 | 27.4% | 1.00 |
| 9 | 44,764 | 28.4% | 1.01 |
| 10 | 47,033 | 28.8% | 1.00 |
| 11 | 23,694 | 29.6% | 1.02 |
| 12 | 18,076 | 29.7% | 1.02 |
| 13 | 13,233 | 31.0% | 1.04 |
| 14 | 10,895 | 32.1% | 1.09 |
| 15 | 8,535 | 30.7% | 1.06 |
| 16 | 6,678 | 31.4% | 1.04 |
| 17 | 5,002 | 31.7% | 1.05 |
| 18 | 4,509 | 32.2% | 1.13 |
| 19 | 4,421 | 29.9% | 1.12 |
| 20 | 13,762 | 32.0% | 1.15 |
Three things stand out. First, the climb is remarkably clean: from 3 slides to 18 the rate rises from 23.4% to 32.2%, with only minor dips along the way. Second, the last column says the same thing a stronger way: a typical short carousel performs exactly at its author's baseline, while the median 18-to-20-slide carousel runs 13 to 15% above it. The long deck doesn't just break out more often; its ordinary outcome is better. Third, the market hasn't caught up. Only 16.7% of carousels use 11+ slides, and volume collapses as length grows (23,694 eleven-slide carousels versus 107,221 five-slide ones). The best-performing length is the least-used one.
The likely mechanism is Instagram's own distribution logic: a carousel can be reshown to the same viewer starting from a later slide, each swipe counts as engagement, and dwell time scales with deck length. A 15-slide carousel simply gives the algorithm more chances than a 4-slide one. Whatever the cause, the old "keep it under 10" advice is not just unsupported in this data; it is backwards. (One small wrinkle: 2-slide carousels slightly beat 3-and-4-slide ones, 25.1% versus 23.4%. The before/after pair is its own genre and it works. The dip is real, but it is a detail; the story of this table is the climb.)
Does the slide effect depend on account size?
Yes, in the best possible way: it holds everywhere. Breakout rate by slide count within each follower band (follower count recorded at the moment of posting):
| Followers when posting | 2-4 slides | 5-7 slides | 8-10 slides | 11-20 slides |
|---|---|---|---|---|
| Under 10,000 | 18.0% | 18.7% | 21.3% | 23.5% |
| 10,000 to 100,000 | 21.7% | 25.4% | 25.4% | 26.8% |
| 100,000+ | 25.8% | 28.3% | 29.9% | 33.0% |
Unlike the timing edges in our Substack study, which flipped direction with account size, the slide effect is uniform: every band's best bucket is 11-20 slides and every band's worst is 2-4. For accounts under 10,000 followers the gap is 5.5 points, about a 30% relative improvement, from a change that costs nothing but effort. This is the rare growth lever that works the same for a 2,000-follower account and a 2,000,000-follower one.
Do carousels actually beat photos and Reels?
Here is where the data undercuts the standard pitch. Breakout rates by format, each measured against its own baseline:
| Format | Posts | Creators | Breakout rate | Median vs own baseline |
|---|---|---|---|---|
| Carousel | 655,385 | 12,159 | 27.1% | 1.00 |
| Single photo | 429,286 | 8,614 | 27.4% | 1.00 |
| Reel | 1,645,407 | 13,135 | 26.4% | 1.05 |
To rule out the "different kinds of accounts post different formats" objection, we also restricted to the 3,828 creators who posted at least 5 carousels AND at least 5 photos in the window, judged against one shared baseline: carousels broke out 28.1% of the time, photos 27.6%. Half a point. The claim you see everywhere, that switching to carousels multiplies your engagement, does not survive contact with 655,385 carousels. Formats are not cheat codes; on likes, the average carousel from an account performs like the average photo from the same account.
Two real format differences survive, though. Carousels are conversation engines: the median carousel earns 1.8 comments per 100 likes, versus 1.2 for single photos, a 50% higher comment rate (Reels lead at 2.7, likely because video reaches more strangers). And carousels hold a solid quarter of all feed-and-reel volume in our index (24%), because they are the natural container for the thing that actually drives the numbers above: depth. The honest synthesis of this table and the slide table is that "post carousels" is weak advice, and "post deep carousels" is the strongest advice in this study.
Does caption length matter?
More than hashtags, less than slides:
| Caption length | Carousels | Breakout rate |
|---|---|---|
| Under 50 chars | 82,224 | 25.4% |
| 50 to 149 | 116,051 | 26.3% |
| 150 to 399 | 149,215 | 25.4% |
| 400 to 999 | 209,600 | 28.5% |
| 1,000+ | 98,727 | 28.6% |
Long captions win, by about 3 points over one-liners. Notably this is the opposite of what we found on Substack Notes, where short posts dominated. The difference makes sense: a note IS the content, while a carousel caption rides alongside it, and a substantial caption gives the post a second surface for dwell time and a place for context, sourcing, and the comment-bait question. Write the deck first, then give it a real caption, 400 characters or more.
Do hashtags still matter?
Barely:
| Hashtags | Carousels | Breakout rate |
|---|---|---|
| None | 423,715 | 27.0% |
| 1 to 3 | 77,391 | 25.4% |
| 4 to 10 | 138,803 | 28.0% |
| 11+ | 15,910 | 28.0% |
The full spread is 2.6 points, the "optimal" 4-to-10 band beats none by exactly one point, and 65% of carousels in the sample use no hashtags at all. Whatever era hashtags were a growth lever in, it has ended. Spend the energy on slide 12 instead.
When should you post a carousel?
We ran the same hour-and-day analysis as our Substack study so the two are directly comparable. The one-line answer: timing matters even less on Instagram.
| Day | Carousels | Breakout rate |
|---|---|---|
| Monday | 100,414 | 27.0% |
| Tuesday | 100,475 | 26.4% |
| Wednesday | 99,842 | 26.1% |
| Thursday | 97,899 | 26.6% |
| Friday | 94,947 | 26.9% |
| Saturday | 78,098 | 28.4% |
| Sunday | 84,153 | 28.4% |
Weekends win by about 2 points over midweek, on noticeably lower posting volume: the fewest carousels are published on Saturday, and Saturday ties for the best breakout rate. By hour (ET), the spread runs from 25.0% at the worst hour (4am) to 28.4% at the best (10pm), with evenings from 4pm onward mostly a point above the midday hours. That is the whole story: about 3 points, hour to hour, with none of the dramatic quiet-hours structure Substack showed. Pick a weekend or evening slot, schedule it, and put the reclaimed attention into the deck.
The compounding lever: cadence
| Carousel cadence | Creators | Carousels | Breakout rate |
|---|---|---|---|
| Under monthly | 5,145 | 23,648 | 23.0% |
| Monthly to weekly | 4,113 | 109,534 | 22.2% |
| 1 to 3 per week | 2,050 | 180,227 | 24.6% |
| 3+ per week | 861 | 342,540 | 30.1% |
Creators posting 3 or more carousels a week break out 30.1% of the time, 7 to 8 points above everyone below weekly cadence, and those 861 creators produce over half the carousels in the entire sample. Read the causality both ways, because it runs both ways: high-cadence accounts are professional operations that got good, and high cadence is how they got good. Either way, per-post odds and volume compound together. The same pattern held on Substack; it appears to be a law of feed platforms.
How to act on this
- Make deeper carousels. The single clearest lever in 655,385 posts: get past 10 slides. If your decks are 4 slides, going to 8-10 alone moves you two to three points; the 11-20 band is where the full lift lives, and only 1 in 6 carousels competes there.
- Give it a real caption, 400+ characters: context, the story behind the deck, a question worth answering. It is worth 3 points and it feeds the comment rate, where carousels already lead photos by 50%.
- Stop optimizing hashtags and posting times. Together they are worth less than the jump from 4 slides to 10. A weekend or evening slot is a free point or two; take it by scheduling carousels in advance rather than posting live. One honest note: Instagram's publishing API caps scheduled carousels at 10 slides for every scheduler, so the 11-to-20-slide monsters have to be posted in the app.
- Build toward 3+ carousels a week. Cadence was worth 8 points in this data. Deep decks at that pace is a production problem, which is exactly the part Eden automates: image post and carousel generation turns an idea into on-brand slides, and the Discover feed shows you which carousels are breaking out in your niche right now, this same analysis running continuously per creator.
Frequently asked questions
How many slides should an Instagram carousel have?
More than you think: in 655,385 carousels, breakout odds climbed almost monotonically from 23.4% at 3 slides to around 32% at 14 to 20 slides. If you want one number, 10 is the floor where the data says you stop leaving lift on the table, and everything from 11 to 20 outperforms it. The old "keep it short" advice is backwards; only 16.7% of carousels use 11+ slides, so long decks are also the least crowded competition.
Do carousels get more engagement than single photos?
Not per post, once you control for the account. Among 3,828 creators who posted at least 5 of each, carousels broke out 28.1% of the time and photos 27.6%, judged against one shared baseline. The real carousel advantages are comment rate (1.8 vs 1.2 comments per 100 likes at the median) and depth: long carousels DO outperform, short ones don't, and photos can't be long.
Do carousels perform better than Reels?
They break out at nearly the same rate against their own baselines (27.1% vs 26.4%), but the comparison is imperfect: Reels are measured on views and carousels on likes, because Instagram only exposes view counts on Reels. Nothing in our data supports picking a format as a strategy. Depth, cadence, and captions moved outcomes; format switching didn't.
What is the best time to post a carousel on Instagram?
Weekends, and evenings around 10pm ET, but hold the advice loosely: the entire best-to-worst spread is about 3 points (28.4% vs 25.0% by hour, 28.4% vs 26.1% by day). Timing matters even less on Instagram than the 3 points we measured on Substack. Schedule a sensible slot and spend the energy on the slides.
Do hashtags help carousels in 2026?
Effectively no. The spread across every hashtag band was 2.6 points, 4-to-10 hashtags beat zero by exactly one point, and 65% of carousels in our sample used none. Hashtags are the lowest-leverage field on the compose screen.
Where does this data come from?
From Eden's own social index, the same corpus that powers its research features across Instagram, Substack, X, LinkedIn, TikTok, and YouTube. We took every indexed Instagram carousel published in the 12 months to August 2026 that was at least a week old, 655,385 posts from 12,159 creators, and measured each against its own author's median. Slide counts come from the full mirrored slide sets (99.2% coverage, validated against a second source). No third-party estimates, no surveys, no vendor data.
We ran this study because carousel advice is everywhere and carousel data is nowhere. If you want the deep-deck strategy without the production grind, Eden generates on-brand carousels, studies the top performers in your niche, and schedules the result. The free plan includes 100 saved items and 50 AI credits with no card required. Start free.