How I Use Flexible Ads for Accounts Under 5K EUR

Meta quietly moved Flexible format out of the Ad Set level back in March 2026.

It's not gone. It's just hides under "Format" in Ads setup level.

If you went looking for it and couldn't find it, that's why.

Here's how I actually use it on small accounts.

I run a lot of accounts under €5000 a month.

On those budgets, you don't get to run 10 ad sets and let the data sort itself out.

You get one shot to feed Meta good material and one shot to get the structure right.

Flexible ads are built for exactly that.

You feed Meta up to 10 creatives (photos/videos), 5 primary texts, and 5 headlines.

Andromeda doesn't run them as separate ads.

It builds a different combination per person, based on what's converting for people like them.

One ad, dozens of versions, all live at once.

But I don't load all ten creatives in on day one.

That's the mistake I see most often.

Here's the sequence I use:

Small account, 2-3 weeks, one consolidated ad set, all genuinely different concepts, not five recolors of the same ad.

Five distinct ideas beat fifteen variations of one idea every time on Meta right now.

Andromeda clusters near-identical creatives into a single retrieval ticket anyway, so polishing one concept fifteen ways buys you nothing.

Not the ones that "got likes."

The ones that moved cost per result, moved ROAS, whatever the account's real metric is.

Pair them with the 3-5 headline and primary text angles that also held up.

Let the system build the per-person combo from material you already know works.

Why not just load all 10 creatives straight into a flexible ad on day one and skip the testing step?

Because on a small budget, Meta doesn't have the spend to learn fast enough across that many combinations before your account stalls in the learning phase.

You end up paying full price to find out what a 3-week isolated test would've told you for less.

Now the part nobody tells you: the drawback.

Once those creatives go into a flexible ad, you lose visibility.

Meta's per-image and per-video breakdown does not work cleanly on flexible/Advantage+ creative delivery.

There's a "breakdown by dynamic creative element" option buried in reporting, and it's flaky, sometimes it populates, sometimes it doesn't.

So the ad as a whole might be performing great.

You just can't always tell which creative, which primary text or headline are carrying it.

That's the trade I make on purpose: better delivery and better matching per person, in exchange for losing the granular "which exact ad is winning" view.

Also, I know from the previous tests that the ad creatives in the flexible ads are winners.

On a small account, I'll take that trade every time, because the alternative is running 10 separate ads to keep clean breakdown data.

This splits an already-thin budget ten ways and nothing gets enough spend to learn.

The fix isn't perfect, but it works: I keep the standalone creative test (Step 1) as my source of truth for "which concept works."

The flexible ad after that is for delivery, not diagnosis.

I don't try to make one ad format do both jobs.

That's the whole system.

Test isolated, then consolidate the winners, then accept you're flying a little blind on attribution once they're combined.

TLDR