What is creative similarity on Meta?
Meta uses the term in two places, and they agree. The Account Insights help page: “Creative similarity occurs when the images or videos in your ads appear too visually identical. This can lead to creative fatigue and increase your cost per result.” The insight is read from static ads active in the last 28 days, and the fix Meta gives is “creating new ads with unique creative”.[3] Its December 2025 note says why it matters: “Our ads system also analyzes creatives and groups together ads that share key visual and thematic attributes. When multiple ads look or feel alike, these are seen as variations of the same creative”, and “learnings and delivery optimizations are shared at the creative level”.[1]
Upload five crops of one image and Meta reads one creative with five files. A third page describes the metric Meta is building to measure it. Creative diversity “is calculated by comparing the images and video thumbnails in your ad set to each other to determine similarity”, labelled High, Medium or Low, and marked “estimated and in development”.[4]
What does Andromeda group together?
Andromeda is the retrieval stage of Meta’s ads system. Meta’s December 2024 engineering post gives its job as “selecting ads from tens of millions of ad candidates into a few thousand relevant ad candidates”, which ranking models then narrow to the ads a person sees.[2] The 2025 note ties it to your account: Andromeda “quickly scans tens of millions of ads” so “only the most relevant and distinct creatives move forward”.[1] The word that matters is distinct.
Meta draws the line with examples. Iteration “might produce two ads with identical visuals, but different text CTAs”; diversification is “a Reels ad tailored for GenZ and a separate partnership ad featuring a creator who connects with new moms”.[1] Agencies are sharper. 303 London says similar ads are grouped under one Entity ID and treated as a single candidate, and that headline or color swaps do not create a new one.[9] Marketing Brew’s September 28, 2026 report quotes Olivia Cripps of Billion Dollar Boy: one video with “Text A” and the same video with “Text B” “would actually be considered the same creative”. Meta’s spokesperson: “We don’t get into the specifics of how our ranking models represent individual creative assets.”[8]
| Same creative, by these accounts | A different concept |
|---|---|
| Identical visuals, different call-to-action text[1] | A Reels ad for one audience and a creator ad for another[1] |
| The same video with “Text A” and “Text B” on screen[8] | Ads “truly different in look, feel, storyline, and message”[1] |
| A headline swap or a color swap on one image[9] | A testimonial, a demo, a story and a joke on one product[9] |
What happens to budget when two ads are near-identical?
Three things, and Meta documents each. First, the pair shares one set of learnings, so the second ad adds no information: look-alikes are “seen as variations of the same creative”.[1] Second, the second ad still costs a learning phase. “Adding a new ad to your ad set” is a significant edit that sends the ad set back into learning,[6] and an ad set usually needs “about 50 results in the week after the ad set’s last significant edit” to leave it.[5] Third, every ad divides delivery: “When an advertiser runs too many ads at once, each ad delivers less often”, so fewer exit learning and more budget is spent before the system can optimize.[7]
So a near-duplicate buys a restarted learning phase and thinner delivery, in exchange for nothing the first ad had not already told the system. Meta’s remedy: “Decrease ads per ad set, but maintain diverse creative assets per ad set”, since one ad can hold up to 10 creative assets.[7] Five crops go inside one ad as placement assets, not into five ads.
Creative fatigue or creative similarity: which one do you have?
They look alike in a report (rising cost per result, falling engagement), and Meta separates them on purpose: “fatigue is about overexposure, while similarity is about a lack of visual diversity”, and “High similarity can also drive fatigue, as users perceive ads as repetitive”.[1]
| Creative fatigue | Creative similarity | |
|---|---|---|
| What Meta says it is | The audience “has seen the same creative (image or video) too many times”[3] | Images or videos that “appear too visually identical”[3] |
| What it is about | Overexposure[1] | Lack of visual diversity[1] |
| Where you see it | Account Insights, active ads of any format[3] | Account Insights, static ads active in the last 28 days[3] |
| Meta’s fix | Add another image or video; duplicate the ad and swap the media[3] | Create new ads with unique creative[3] |
Fatigue arrives over time on one ad as its frequency climbs; similarity is there on day one across several ads. At what frequency does ad fatigue start? covers the first case.
What counts as a different concept? A worked example for a med spa
A different concept passes 303 London’s test: “If you removed the text overlays, would they all look the same?”[9] Jon Loomer says the same from Meta’s examples: “Unless they’re distinctly different enough, any variation is unlikely to make a meaningful impact on performance.”[11] Meta’s standard is ads “truly different in look, feel, storyline, and message” that “speak to different audience motivations or use cases”.[1]
Take a med spa with one offer, a first-visit facial. Five ads that fail: the same treatment-room photo under five headlines. Strip the words and they are one picture. Four that pass:
- 1
The problem
A close crop of tired skin in bathroom light, with a headline about what she sees in the mirror. No product, no room. - 2
The proof
A real review in the client’s words, set as text on the brand color, her first name underneath. No photo. - 3
The process
Three small frames in one image: consult, treatment, the walk out. A different layout and story. - 4
The offer
The first-visit price as the headline on a clean card with the logo and the booking line. The only one where price leads.
Four pictures, four layouts, four reasons to book. Meta’s advice is to “Run no more than five ads per ad set for optimal delivery”,[12] so four concepts fill the set. An HVAC company does the same with a leak, a review, a van at the door and a price.
How different do your ads need to be? The numbers that circulate
Meta has not given one. Its creative diversity metric returns High, Medium or Low with the thresholds unpublished,[4] and its spokesperson told Marketing Brew the guidance is “deliberately outcome-based (not model-specific)”, with a “small number of agencies/GCR resellers” testing creative insights metrics in the Insights API since June 2026.[8] Every percentage comes from somewhere else.
| Figure | Who says it | What it rests on |
|---|---|---|
| Keep similarity under 40% | Segwise, crediting 303 London[10] | 303 London’s own guide gives no percentage[9] |
| Above 60% triggers retrieval suppression | Segwise, crediting admetrics[10] | A vendor claim; no Meta source |
| 8 to 12 concepts, 2 to 3 variations each | Segwise[10] | Vendor guidance for larger accounts |
| Similar creatives raise costs 20 to 30% | 303 London FAQ[9] | Cites data the page does not show |
The safe reading is Meta’s own: five ads per ad set, each passing the words-removed test.[12][9] Loomer adds that “more isn’t necessarily better”: more ads help only if they are a more diverse group.[11] How many ad creatives to test at once turns that into a count for a small budget.
A checklist before you launch a batch
- 1
Strip the words
Lay the images side by side with no text. If two look the same, they are one concept.[9] - 2
Name the motivation on each ad
Problem, proof, process, offer, origin. Same motivation and same look is a variation.[1] - 3
Put crops inside the ad, not beside it
The 4:5 and 9:16 versions of one concept are placement assets of one ad; Meta allows up to 10.[7] - 4
Cap the set at five
Meta’s guidance: no more than five ads per ad set.[12] - 5
Launch the batch in one edit
Every ad added later restarts learning.[6] - 6
Read Account Insights after 28 days
If creative similarity is flagged, Meta’s fix is new ads with unique creative, not more crops.[3] - 7
Make the next batch unlike the winner
Look-alikes share one learning path.[1] Showfolk returns four varied versions of every ad, each a different concept rather than a recolor. How the AI ad generator does it.
Questions people ask
- Does Meta penalize duplicate ads in the same ad set?
- Not with a published penalty, but it treats them as one creative. Meta says its system groups ads that share key visual and thematic attributes and shares learnings at the creative level. Each duplicate still counts as a new ad, which restarts the learning phase and thins delivery across the set.
- What is the difference between creative fatigue and creative similarity?
- Fatigue is overexposure: one audience has seen the same image or video too many times. Similarity is a lack of visual diversity: several ads that look too alike, even as separate files. Meta says high similarity can itself drive fatigue.
- How different do Facebook ads need to be to count as separate creatives?
- Meta gives examples, not a number. Identical visuals with a different call to action are an iteration; a Reels ad for one audience and a creator ad for another are diversification. A working test is whether the ads still differ with the words removed.
- Is there a creative similarity score in Ads Manager?
- Meta flags creative similarity in Account Insights, based on static ads active in the last 28 days, and is building a creative diversity metric that labels an ad set High, Medium or Low, with thresholds unpublished. Percentage scores come from third-party tools.
Sources
Every figure above points at one of these. Facts were last checked on .
- [1]Meta for Business, December 16, 2025: Demystifying creative diversification. The Andromeda sentence, “groups together ads that share key visual and thematic attributes”, “seen as variations of the same creative”, learnings shared at the creative level, iteration vs diversification, the Gen Z and new-moms example, creative fatigue vs creative similarity
- [2]Engineering at Meta, December 2, 2024: Meta Andromeda, supercharging Advantage+ automation with the next-gen personalized ads retrieval engine. Retrieval selects a few thousand candidates from tens of millions; the post does not itself tell advertisers to diversify creative
- [3]Meta Business Help Center: About Account Insights in Meta Ads Reporting. Definitions of creative fatigue and creative similarity, the 28-day static-ad window, and the fix for each
- [4]Meta Business Help Center: Creative diversity. The metric compares images and video thumbnails in an ad set, is scored into High, Medium or Low by thresholds, and is “estimated and in development”
- [5]Meta Business Help Center: About the learning phase. About 50 results in the week after the last significant edit; avoid high ad volumes
- [6]Meta Business Help Center: Significant edits and learning phase. “Any change to ad creative” and “Adding a new ad to your ad set” are significant edits
- [7]Meta Business Help Center: About managing ad volume. Too many ads at once means each delivers less often and fewer exit learning; decrease ads per ad set but keep diverse assets; one ad can contain up to 10 creative assets
- [8]Marketing Brew, September 28, 2026: How advertisers are adjusting to Meta’s ad creative diversification best practices. Agency accounts of same-look and same-hook ads being treated as the same ad, the Text A / Text B example, Meta spokesperson Alisha Swinteck’s statements, and the Insights API test since June 2026
- [9]303 London: Complete guide to creative diversity for Meta Andromeda. Agency guidance: the text-overlay test, Entity ID grouping, the fifty-similar-images line, 3 to 5 concepts to start, and a 20 to 30% cost claim without a shown source; no similarity percentage
- [10]Segwise: Meta Andromeda update 2026, creative strategy playbook. Vendor guidance: 8 to 12 concepts with 2 to 3 variations, similarity under 40% (credited to 303 London), suppression above 60% (credited to admetrics)
- [11]Jon Loomer, October 20, 2025: Meta Andromeda and creative diversification, 7 examples explained. “More isn’t necessarily better”, two images that look very similar, distinct formats, up to 10 creatives in one ad
- [12]Meta for Government, Politics and Nonprofits, January 21, 2026: Smarter campaigns: how machine learning drives ad performance on Meta technologies. “Run no more than five ads per ad set for optimal delivery” and “High ad set volume is a leading contributor to ad sets not exiting the learning phase”
