Advize is an AI-powered performance marketing agency that checks Advantage+ creative asset count directly, not as an afterthought, since Meta's own guidance is specific: somewhere in the range of 10 to 15 conceptually distinct assets per campaign to properly feed the Andromeda retrieval engine. A lot of accounts are running Advantage+ with three or four creative concepts that have been live for months, and that gap between the recommended range and the actual asset count is often the real, unglamorous reason performance has plateaued. This is the real question behind how many ad creatives per campaign an account actually needs: are your Advantage+ creative assets hitting that range, or just the raw ad count in Ads Manager?
Why the Number Isn't Arbitrary
Andromeda's retrieval mechanism uses creative variety as a direct input for matching users to ads, meaning the system needs a genuinely diverse pool of concepts to work from before it can find the specific micro-audiences each concept resonates with. A campaign running three or four assets simply doesn't give Andromeda enough distinct creative signal to do that matching work well, regardless of how much budget or how sophisticated the audience settings are. Ten to fifteen conceptually distinct assets is Meta's own stated range for giving the retrieval engine enough to actually work with.
Why 'Distinct' Is Doing the Real Work in That Number
Fifteen ads that are all variations on the same hook, same format, same emotional angle, don't satisfy this requirement even if the raw asset count technically hits the target. Andromeda reportedly evaluates conceptual distinction, not just asset count, which means a campaign with fifteen slightly-different versions of the same idea is closer to running one concept than fifteen, and won't produce the creative-signal benefit the recommendation is actually pointing toward.
Auditing Whether Your Account Actually Meets This Bar
Count active creative assets in the campaign first, a simple raw number check against the 10-15 range. Then review each asset for genuine conceptual distinction, not surface variation, does this asset use a different hook, a different format, a different emotional angle, or is it a headline swap on the same underlying idea. Group assets by underlying concept rather than by individual ad, and count the number of genuinely distinct concept groups, since that grouped count is the number that actually matters against Meta's guidance, not the raw ad count in Ads Manager.
The Account Running Fifteen Ads and Three Concepts
An account technically had fifteen active ads in its Advantage+ campaign, appearing to meet Meta's recommended range at a glance. A closer review grouping those fifteen ads by underlying concept found they collapsed into just three genuinely distinct ideas, each with four or five headline and format variations. Building five new, conceptually distinct concepts, different hooks and different emotional angles rather than more variations, brought the account to a genuine eight distinct concepts, and performance improved measurably within the following weeks as Andromeda had meaningfully more real creative signal to match against.
A Quick Test for Genuine Conceptual Distinction
For any two assets in the same campaign, ask: do they use a fundamentally different hook, not just different wording of the same hook. Do they address a different customer objection or emotional angle, not the same one framed slightly differently. Would a viewer scrolling past both recognize them as genuinely different ideas, or as the same ad seen twice. Assets failing this test should be counted as one concept, not two, when checking against the 10-15 target.
Why This Gets Missed Even by Teams Who Test Regularly
A team running a steady stream of new ads can still fail this bar if every new ad is a minor iteration of an existing winner rather than a genuinely new concept, since iteration feels like testing activity without producing the conceptual diversity Andromeda is actually looking for. Checking asset count grouped by real concept, not just checking whether new ads are launching regularly, is what catches this specific gap.
This Connects Directly to Testing Volume Math Elsewhere
Hitting a genuine 10-15 distinct concept range isn't a one-time fix, it requires a testing calendar producing real new concepts on an ongoing basis, not just replacing tired ads with slight variations when performance dips. An account that treats this as a standing target, not a one-time audit, is the account that keeps Andromeda properly fed as older concepts naturally fatigue over time.
The Short Version
Meta's own guidance recommends 10 to 15 conceptually distinct creative assets per Advantage+ campaign to properly feed the Andromeda retrieval engine, and many accounts fall short of that bar even while appearing to run a reasonable number of ads, because those ads collapse into far fewer genuinely distinct concepts. Advize audits creative count grouped by real concept, not raw ad count, since that grouped number is what actually determines whether Andromeda has enough signal to work with.
Conclusion
The gap between running ads and running enough genuinely distinct creative concepts is easy to miss, since both look like activity from the outside. Advize checks the real number specifically because Andromeda can't find an audience it was never given enough creative signal to search for, and that's a fixable, unglamorous problem hiding underneath a lot of accounts that look busy without actually being diverse.