←
Ben Radack 🏝️13,905 posts
Ben Radack 🏝️ avatar

Ben Radack 🏝️

@benradack

Meta ads media buying + creative for DTC brands. Small senior team, zero big agency energy. Book a free strategy call with me below.

◷ Joined September 2020▧ 990 Photos and videos
Ben Radack 🏝️ avatar
Ben Radack 🏝️@benradack·28 Dec 2025

I took over this account 2 days ago and dropped their CPA by 74%. All with a SUPER simple account structure & testing strategy Creative Testing - ABO → Went through their ad library (hundreds of ads) → Batch tested a ton of them across 5 different ad sets → Yes, Meta skipped a bunch, but it helped me find winners fast Scaling Campaign - Cost Cap CBO → Ad set 1: All historical winners from the account → Ad set 2: All proven winners from December testing - min spend limit That's it. What was broken: Their past team was hardly testing anything. All the spend was going to 1 ad that was extremely fatigued. The account had zero velocity. The fix: Simplify the account structure + increase testing velocity significantly. Sometimes the best strategy is just testing more and organizing better. You don't need 10 campaigns and complex structures. You need clean separation between testing and scaling, and you need to actually test. 2 days in, CPA dropped 74%.

Post attachment
Ben Radack 🏝️ avatar
⟳ Ben Radack 🏝️ retweeted
Ben Radack 🏝️@benradack·17h

Breaking down my strategy for stopping the bleeding 🩸 on a brand spending $500k/mo and profiting $0 ㅤ 1. Consolidation ㅤ Cut what's weighing the account down → Sort every campaign by spend → Shut off campaigns/adsets that are pulling spend from top performers ㅤ Move what works into one CBO → 1 ad set for old winners → 1 ad set per new batch of winners → Min spend on new ad sets, max so they don't run away ㅤ Bring back old winners → Date range: all time, sort by spend → Relaunch the old winners in the new CBO ㅤ 2. Large batch tests → 20-30 ads per ad set → Grouped by creator or concept → Budget at least 4x your avg CPA ㅤ Carve out the winners → In ~2 days Meta pushes spend to a few ads → Rename them "ad name - Winner" and move them to scaling → Turn them off in the batch, Meta picks the next best ㅤ What's left goes to the zombie campaign.

Ben Radack 🏝️ avatar
Ben Radack 🏝️@benradack·8h

This Black Friday, keep a close watch on your incremental attribution column ㅤ If your evergreen campaigns are set to include 1 day view or engaged view, Meta will take credit for purchases it didn't drive. ㅤ 1. Why it matters during BF ㅤ → Your emails drive traffic to your site → Someone sees your ad, then buys from your email → Meta counts it as an ad purchase, and you end up scaling into nothing ㅤ 2. What the incremental column shows you ㅤ → Only the purchases your ads actually caused → So you can see which campaigns are really performing and which are just taking credit ㅤ 3. How to turn it on ㅤ → Columns > Compare attribution settings > Incremental attribution → It sits right next to your normal results, no need to change your campaign settings

Post attachment
Ben Radack 🏝️ avatar
Ben Radack 🏝️@benradack·10h

Breaking down my Zombie campaign (for failed tests) 🧟‍♂️ ㅤ CBO with a low cost cap. Start it at 50% of your avg CPA, then raise it until it spends and stays profitable. ㅤ 1. Zombie winners ad set ㅤ → Failed tests that came back to life in the zombie campaign → This works like a scaling ad set for them ㅤ 2. New ad sets for recently failed tests ㅤ → One ad set per batch of recently failed tests → Min spend limit to force a bit of spend into them → The ones that start converting move into the zombie winners ad set ㅤ Great way to bring dead ads back to life

Post attachment
Sean Frank avatar
⟳ Ben Radack 🏝️ retweeted
Sean Frank@Seanfrank·Nov 12

clipping wont save you. trybe wont save you. UGC wont save you. naming conventions or adset structure wont save you. aftersell or rivo or northbeam or ai wont save you. you need to do all of it you need to be testing. Its attention, urgency, offer, value. thats the formula.

Phil Kiel avatar
⟳ Ben Radack 🏝️ retweeted
Phil Kiel@PhilKiel·Nov 17

More creative means more media buying. Your activity at 50 ads per month won’t cut the mustard at 500+ ads per month.

Everyone says "launch more creative." So we looked at what happened accounts increased their creative output, the winners and the losers. Some accounts nearly doubled their new ads and got 40 to 70% more winners. Others doubled their creative output and got fewer winners than before. The difference was what happened AFTER they launched the ads. The accounts that won made 1.3 to 3.6 budget decisions per new ad. Step up, hold, step down or kill, every day, against written rules. The ones that didn't made 0.6. Most new ads were never touched again, so Meta picked a few and the rest never got a real test. If you’ve recently increased how many ads you’re launching, have you changed your process to account for it?

Ben Radack 🏝️ avatar
Ben Radack 🏝️@benradack·Nov 17

You NEED to pay attention to your "Audience Segment Breakdowns" Otherwise you'll think ads like this are "crushing" When in reality they just did some good retargeting Here's how you can check: ㅤ 1. Set up your segments ㅤ → All Tools > Advertising Settings > Audience Segments → Existing customers: your customer email list + purchasers → Engaged audience: 180 day website visitors ㅤ 2. Check your breakdowns ㅤ → Breakdown > Demographics > Audience segments → Splits spend, purchases and CPA into new, engaged and existing ㅤ 3. What to look for ㅤ → Mostly new customers + good CPA: that's a real winner, scale it → Mostly engaged or existing: it's closing people who already knew you, and it likely won't hold up at more spend

Post attachment
Ben Radack 🏝️ avatar
Ben Radack 🏝️@benradack·Nov 17

Quick tip: run your top organic posts as ads ㅤ → Pull your 10 highest engagement posts from the last 90 days → New campaign, run them as ads using their post IDs Make sure you run every post that has engagement Even the posts that weren't meant to be ads perform well as ads

Ben Radack 🏝️ avatar
Ben Radack 🏝️@benradack·Nov 17

How I structure creative tests when I actually want to learn something ㅤ 1. Change one thing → Pick one variable: angle, hook, format, creator or landing page → Keep everything else the same ㅤ 2. Give it a fair test → Its own ad set with a set budget → Let it run 4-7 days before deciding → Judge on cost per add to cart first, then CPA and ROAS once there's enough spend and purchases behind them ㅤ 3. Example: testing an angle for a supplement brand → Same format (static), same landing page, same offer → Ad 1: "Tastes better than your greens powder" → Ad 2: "Finally a vitamin you'll actually remember to take" → Ad 3: "Recommended by 500+ doctors" ㅤ Three different reasons to buy, same everything else. When one wins, I know it was the angle. ㅤ 4. Build on the winner → Take the winning angle and test it across new formats, hooks and creators ㅤ If you change the hook, thumbnail, copy and landing page all at once and the ad works, you can't repeat it because you don't know what drove the result.

Ben Radack 🏝️ avatar
Ben Radack 🏝️@benradack·Nov 17

ABO is far from dead. It's still a great way to test creative. ㅤ Let me explain why... ㅤ 1. Meta misses winners ㅤ → Meta is great at choosing which ads to spend on → But it often skips right over winners. I seent it!! ㅤ 2. Small budget winners add up ㅤ → Imagine one big ABO campaign with 30+ ad sets at $200-$500/day each → Some ads do well at that level and fall apart at high spend in scaling → So they all add up. 30 ad sets performing at small spends make the campaign perform as a whole → Why? Each angle hits its own pocket audience. Too small to scale, but small enough to work at a restricted budget ㅤ 3. You actually learn something ㅤ → ABO lets you test new personas and new angles and see how each one performs → Dump them all into one ad set and let Meta pick, and you'll never know if a different persona would've gotten you better results

ABO is dead Do not force spend on every ad Meta is already better than you at picking winners Your job is giving it clean data Because Meta misses 40% of sales if you don't Just make more, better ads sir

Ben Radack 🏝️ avatar
⟳ Ben Radack 🏝️ retweeted
Ben Radack 🏝️@benradack·Nov 16

These strategies will save you 💰💰💰 during Black Friday - SAVE this one! ㅤ 1️⃣ BF Campaigns = 7 day click attribution → Your BF emails and texts drive a ton of purchases → With 1 day view, Meta takes credit for them and your cost caps overspend chasing people who were going to buy anyway ㅤ 2️⃣ Split your audiences in your Black Friday CBO → TOF: exclude 180 day website visitors and all time purchasers → Retargeting engaged: 180 day website visitors, max spend at 20% → Existing customers: all time purchasers, exclude the last 30 days, max spend at 10% → Keeps most of the budget on new customers ㅤ 3️⃣ Track Segment Spending → Check audience segment breakdowns to track where Meta is spending → This helps you identify and pull back on campaigns that overspend in existing or engaged ㅤ 4️⃣ Check 7DC Attribution Settings → Go to Columns > Compare Attribution Settings > Select 7 Day Click → This will allow you to see which campaigns are not performing as well as others. (Which ones are getting clicks) ㅤ 5️⃣ Keep evergreen ads running → Evergreen winners stay live and testing keeps going → Cost caps pull back spend on their own when demand drops after BF

Post attachment
Ben Radack 🏝️ avatar
Ben Radack 🏝️@benradack·Nov 16

These are the 3 Meta ads metrics I track beyond just ROAS ㅤ 1. Cost per add to cart → The clearest signal of which ads are bringing people into the funnel → Low cost per ATC + low CPC = the ads doing the real top of funnel work → These are your best ads, even if the ROAS sucks ㅤ 2. Add to cart rate → Tells you how your offer and landing page are holding up with the audience each ad pulls in ㅤ 3. Checkout to purchase rate → Helps me spot when an ad is worth keeping on → People are moving through the funnel but haven't converted yet, so be patient, don't pull the plug → A dropping rate can also mean something's wrong with your checkout or shipping offer ㅤ Together these 3 will help you identify winners & losers way before your ROAS does

Ben Radack 🏝️ avatar
Ben Radack 🏝️@benradack·Nov 16

CTR is an irrelevant metric... ㅤ Getting people to click matters. The CTR metric itself doesn't. ㅤ What matters is: → How many clicks you're actually getting → Whether those people add to cart and buy ㅤ This campaign averages around a .4% CTR. Some would call that bad. ㅤ But with their CPMs, and beyond that, the audience that buys their products. Its actually pretty damn good ㅤ Every brand is different. Some will never get above a 0.5% CTR, and some average a 3. ㅤ So judge your ads on sales metrics instead (cost per add to cart, cost per purchase, ROAS etc) Don't make decisions based on CTRs

Post attachment
397Y2@397Y2Nov 11

CTR is irrelevant and CTR (all) is even more irrelevant

Ben Radack 🏝️ avatar
⟳ Ben Radack 🏝️ retweeted
Ben Radack 🏝️@benradack·Nov 15

Your scaling campaign will get top heavy if you only have one ad set 🏋️ ㅤ One ad takes all the spend and your new ads get nothing... ㅤ Here's the fix. Split your scaling CBO into separate ad sets 👇 ㅤ 🏆 Historical winners → 1 ad set for your proven ads ㅤ 🍽️ Feeder ad sets → A new ad set for every batch of proven winners from your testing campaign → Min spend limits on each to force distribution → The newest ones also get a max spend limit so a fresh winner can't run away with the budget before it's proven at scale ㅤ 📈 As they perform → Raise the min spend or remove the max → The account builds depth over time instead of leaning on whatever Meta liked first ㅤ 🛡️ Why it works → Meta can't dump 80% of spend into one old high engagement ad → When one ad dies, it's only a slice of your budget, not the whole thing → Other winners are already warmed up at real spend, ready to pick up the slack ㅤ Always be feeding the next winners in while the current ones are still working.

Post attachment
Jordan Menard avatar
⟳ Ben Radack 🏝️ retweeted
Jordan Menard@jordanmenard·Nov 15

I have been saying this for as long as this dumb saying has been floating around: Saying media buying is dead is self snitching. You’re admitting you’ve never actually skaled. That’s all it is.

Anyone claiming AI has killed media buying is basically admitting they never understood what media buying actually is.

Ben Radack 🏝️ avatar
⟳ Ben Radack 🏝️ retweeted
Ben Radack 🏝️@benradack·Nov 13

One of the easiest ways to turn around a failing ad account is consolidating down to one CBO campaign ㅤ If most of your creative tests fail, or you simply can't afford to test in ABO, one campaign makes things a lot more efficient. ㅤ 1️⃣ Why CBO works ㅤ → Meta distributes the budget to the ads it thinks are best → Losing ads barely spend, so you stop paying for every failed test → All your data sits in one place instead of spread across campaigns ㅤ 2️⃣ You still get to guide it ㅤ → Min spend limits on new creative so Meta actually spends on it → Max spend limits so a new ad can't run away with the budget → Everything in between is Meta's call ㅤ 3️⃣ How I structure it (3-4 ad sets) ㅤ → 1 for your video winners ㅤ → 1 for your new creative tests (min so they spend, & max limit so they don't overspend) ㅤ → 1 for your winning statics, which mostly spend on engaged (max limit so they don't overspend) ㅤ Helps you get things efficient and still keep your hand on the wheel

Post attachment