Your Meta ads aren't broken.
The platform is. Here's how to work around it.
đ Thursday, 11 Jun 2026
Hey â It's Rhythm. Thrilled to have you back in Ecom Circle, where I spend my time scaling ecom brands & finding best growth strategies to share with you.
The last few months on Meta have been unpredictable, and if youâve been feeling that, youâre not alone. Budgets that used to be reliable started swinging. Audiences that used to convert started going cold. CPMs moved in directions that didnât track with anything you were doing differently. A lot of people I talk to are quietly wondering if theyâre missing something or if the platform just broke on everyone at the same time.
Spoiler: it kind of did.
So letâs skip the post-mortem on what broke. You lived it. What I want to talk about is whatâs actually working right now, from accounts Iâm actively running and from people in the trenches alongside me.
Pull up a chair.
You donât need to be an AI expert to create and analyze social content at scale. Everything your team needs to go from âwe should use AIâ to actually using it.
This free toolkit has two on-demand webinars and two prompt guides built specifically for social media managers who need to make AI work now, not someday.
The first thing Iâd tell you to do is cut Audience Network.
Just turn it off. I know Meta defaults it on and it looks like free reach. Itâs not. Turning it off has meaningfully moved the needle for me across accounts. The placements outside of Metaâs core surfaces are doing work you canât really see, and not the good kind. Once you cut it, your data gets cleaner and your CPMs start making more sense. Thereâs a tweet I put out walking through when I leave it on vs. off (rare scenarios), but the default should be off.
Second: stop running more than 4-5 sales objective campaigns.
I know it feels like more campaigns equals more shots on goal. What actually happens is your incremental reach starts cannibalizing itself. Four or five consolidated campaigns, run well, will almost always outperform eight scattered ones. The algorithm needs room to breathe and spend, and when you fragment it across too many campaigns, it just competes against itself.
On creative and landing pages, hereâs something most people skip.
When you find a winning creative, the instinct is to scale it everywhere. The smarter move is to take that same creative and run it to a different landing page to avoid overlap. One ad set per landing page. Run flex ads or post IDs. This keeps your top-of-funnel from stepping on your retargeting and gives you cleaner data on whatâs actually converting.
And while weâre on landing pages: if all your traffic is going to PDPs or collection pages, youâre going to see high CPMrs and struggling traffic. You need a mix. Listicles, advertorials, editorial-style LPs. A split traffic strategy isnât optional anymore. Itâs how the platform works now.
For attribution, the picture is more nuanced than most people admit.
Incremental Attribution is worth testing, but itâs not for everyone. In my Haus tests, itâs proven genuinely incremental for brands with a real social presence, a strong retention audience, or a longer consideration cycle. If your brand doesnât have that foundation yet, donât expect IA to fix a spend efficiency problem. Know what youâre diagnosing before you reach for the solution.
Daily breakdowns changed how I think about spend.
Iâve been building out a daily breakdown in Manus that tracks the breakdown effect of ad spend across each ad, using CPMr as the main lens. It tells you how each ad is playing in the funnel on any given day, not just what the aggregate numbers say at the end of the week. This is where you catch problems early and figure out when to push and when to pull back.
Partnership ads and whitelisted DPAs are two things Iâd add right now.
Partnership ads at each stage of the funnel, mixed into your main scaling campaign, adds social proof in a way standard ads just canât replicate. And whitelisted DPAs running from creator accounts? I picked this up from the team at Select and itâs performing well out of the gate. If you havenât tried it yet, add it to your test list this week.
Two more things that round out the playbook.
Allocate about 5-10% of your total ad spend to sales-based campaigns using daily unique reach, running your top performers in a sequence. This builds net-new reach consistently without burning your main campaigns. Think of it as a reach tax that pays compounding dividends.
And pair bid caps with a lowest cost testing framework simultaneously. Running both gives you two reads on the same creative: what it does when you let Meta spend freely, and what it does when you put a ceiling on it. When bid cap spend lines up with your lowest cost results, thatâs your signal to scale.
None of this is magic. Itâs iteration on top of iteration, paying close attention to what the data is telling you, and having the discipline to simplify when the platform is pushing you to complicate things.
Thatâs whatâs working right now. Check back in six months, because itâll be different again.