Meta’s Andromeda system changed how ads get selected before they even reach the auction. And one pattern keeps showing up: ads that look too polished, too scripted, too “produced” are struggling to get seen. Marketers have started using the term Authenticity Score as informal shorthand for this shift. It’s not an official metric Meta has published, but it captures a real, observable pattern in how creative performs under the new system.
For years, the advice was simple: shoot clean, light well, script tightly, look professional. That advice worked when Meta’s older ad system leaned on audience targeting and pixel data to decide who saw what. Creative quality mattered, but it wasn’t the main lever.
Andromeda flipped that. Now the creative itself is read, analyzed, and matched to users before targeting even comes into play. And under that system, ads that feel like real content, not like ads, are pulling ahead. This article breaks down what actually changed, why the Authenticity Score idea matters, and what to do about it.
What Changed Under Meta's Andromeda Update
Meta announced Andromeda through its engineering blog on December 2, 2024, describing it as a next-generation AI-powered ads retrieval engine. It rolled out in phases through 2025 and was operating globally across most placements by October 2025.
Here’s the core shift: Andromeda uses computer vision, natural language processing, and audio analysis to read your creative directly, including the visuals, the copy, the pacing, and the tone. It compares this against real-time user behavior instead of relying mainly on historical purchase data and audience definitions.
One documented mechanic worth knowing: Andromeda groups similar-looking ads under something called an Entity ID. If your ads score too closely on visual and messaging similarity, they get clustered together and compete for the same limited delivery opportunity, instead of expanding your reach. This is sometimes referred to as a Creative Similarity Score by industry analysts covering the system.
Meta has also confirmed that creative fatigue windows have compressed. Ads that used to perform well for six to eight weeks are now burning out closer to two to four weeks, because the system finds the best-fit audience for a creative faster, and exhausts that audience faster too.
None of this is officially labeled “Authenticity Score” by Meta. But the pattern advertisers are seeing, where real, unscripted, human-feeling content outperforms polished studio ads, is consistent enough that the term has become useful shorthand in the industry.
Why Authenticity Score Matters
Think of Authenticity Score as a way to describe how “real” your ad reads to both the algorithm and the person scrolling past it.
It matters because Andromeda’s retrieval stage happens before the auction. If your ad doesn’t get flagged as relevant and engaging early, it may never get a real chance to compete on bid or budget. Creative that reads as templated or overly commercial risks getting deprioritized at this earlier stage.
It also connects to Meta’s documented Ad Quality Score, which penalizes ads more aggressively for bait-and-switch creative, misleading headlines, and low landing page relevance. Authenticity, in practice, is one way to avoid tripping those penalties, because genuine, transparent ads are less likely to look manipulative to the system or the user.
User engagement plays a role too. When people interact with an ad, such as watching longer, saving it, or clicking through and staying on the page, that behavior teaches Andromeda what “good” looks like for your brand. Ads that feel authentic tend to hold attention longer than ads that feel like interruptions, which feeds the engagement signals the system is built to reward.
Signs Your Ads May Have a Low Authenticity Score
Watch for these patterns in your account:
- Frequency climbs quickly while CTR drops, a sign the algorithm has quickly found and exhausted your audience
- Cost per result rises even though nothing else about the campaign changed
- Ads Manager flags “Creative Limited” or “Creative Fatigue” earlier than expected
- New ad variations that are just small edits of the same concept show flat, near-identical performance
- Engagement (comments, saves, shares) stays low even when impressions are healthy
None of these confirm authenticity is the specific problem on their own. But together, they point to creative that isn’t resonating, which is the exact signal Andromeda is built to detect.
Common Reasons Highly Polished Ads Perform Worse
A few patterns show up again and again in how advertisers describe this shift:
Users scroll past anything that looks like an ad. Studio lighting, stock-photo energy, and scripted delivery are easy to spot instantly, and people tune them out.
Polished ads often lack variety. If your production process only allows for one “hero” version, you end up with fewer genuinely distinct creative concepts, and Andromeda rewards diversity, not repetition of the same polished formula.
Overly commercial framing reduces trust. Ad copy that reads like a sales pitch, rather than a real recommendation, tends to generate less of the on-platform engagement (saves, comments, shares) that signals relevance.
Fatigue hits polished ads just as hard, if not harder. Because production is expensive and slow, brands often run the same few polished ads longer than they should, right into the compressed 2 to 4 week fatigue window Andromeda enforces.
None of this means the algorithm is penalizing “quality” itself. It’s rewarding relevance and engagement, and authentic-feeling content is simply winning more of that engagement right now.
Practical Ways to Improve Your Authenticity Score
- Use real people, not actors reading scripts. Employees, real customers, or founders speaking naturally tend to read as more credible.
- Shoot in natural light and real environments. A kitchen, a car, an office desk, settings people recognize from their own lives.
- Feature genuine customer experiences. Real testimonials, unscripted reactions, and honest before and after moments outperform manufactured claims.
- Loosen your messaging. Drop the salesy language. Speak the way you’d explain the product to a friend.
- Add behind-the-scenes content. Show how something is made, packed, or used. Process content reads as inherently more honest.
- Build variety into every batch. Different angles, different formats (UGC, static, video), different narrators. This also helps you avoid the Entity ID clustering problem described earlier.
A Simple Example
Polished version: A studio-shot video with a professional model, scripted voiceover, dramatic lighting, and a tight sales message ending in “Shop Now.”
Authentic version: A customer filming themselves unboxing the product on their phone, talking through their honest first impression, filmed in their living room.
Both can be well-made. But the second one looks like something a person would naturally send a friend, not something built specifically to sell. That distinction is what the Authenticity Score concept is trying to capture.
Mistakes to Avoid
- Treating “authentic-looking” as a style you can fake with better acting, since audiences and platforms both catch performative authenticity quickly
- Running only one or two “real” ads while the rest of your account stays fully scripted
- Ignoring landing page experience, since even a strong ad won’t help you if it leads to a page that breaks the trust it built
- Waiting too long to refresh, given how much shorter fatigue windows have become
- Assuming raw, unpolished automatically means better. The goal is genuine, not sloppy
Build Authenticity Into Your Weekly Routine
A simple weekly workflow keeps this manageable:
- Monday: Review last week’s frequency, CTR, and fatigue flags in Ads Manager
- Tuesday to Wednesday: Source or shoot 2 to 3 new authentic-style concepts (UGC, testimonials, behind-the-scenes)
- Thursday: Launch new variations alongside proven performers, watching for Entity ID clustering
- Friday: Check engagement signals (saves, comments, watch time) on newer ads versus polished ones
This keeps a steady pipeline of authentic content flowing, rather than scrambling for it only after performance drops.
Conclusion
Meta’s Andromeda update didn’t ban polished advertising. It changed what gets rewarded at the retrieval stage, before the auction even starts, and right now, that favors creative that feels real over creative that feels produced. The idea of an Authenticity Score isn’t an official Meta metric, but it’s a useful way to think about this shift: ads that read as genuine tend to hold attention, avoid quality penalties, and dodge the fatigue that catches polished, low-variety creative faster than before. Instead of chasing a flawless studio look, focus on building trust. Real people, honest messaging, and creative variety will do more for your results than another polished shoot.
FAQs
1. Is Authenticity Score an official Meta metric?
No. It’s not a term Meta has published. It’s informal shorthand marketers use to describe the pattern where authentic-feeling ads outperform polished ones under Andromeda.
2. Can I check my Authenticity Score in Ads Manager?
There’s no dashboard labeled this way. Instead, watch related signals Meta does report, such as frequency, CTR trends, and “Creative Fatigue” or “Creative Limited” flags.
3. Does a low Authenticity Score mean my ad will stop delivering entirely?
Not necessarily. It means the ad may struggle to stand out at the retrieval stage or may hit fatigue faster, which can raise costs and limit reach.
4. How often should I refresh creative to protect my Authenticity Score?
Given the compressed 2 to 4 week fatigue window under Andromeda, refreshing every two to three weeks is a reasonable baseline for active campaigns.
5. Will a higher production budget improve my Authenticity Score?
Not directly. Production value isn’t the issue; how genuine and relatable the ad feels is. A low-budget, honest customer video can outperform an expensive studio shoot.
