At Nord Media, we focus on profitable growth at scale, we help forecast revenue, and contribution margin goals, then build the marketing plans to execute those goals.

This means full clarity on testing, scaling and improving profitability at every aspect of your funnel.

If you want a growth partner aligned with your real business goals, let’s chat

My feed is full of Black Friday account builds right now. Some operators split their spend across a stack of campaigns, one for each stage of the funnel, while others load 100 ads into just three ad sets and let Meta sort it out.

None of those builds starts where I’d start.

Before you copy anything you see online for a last minute push - how many ad sets can your account afford?

Each ad set needs a steady stream of conversions before Meta learns who to show it to, and your account only produces limited sales a week.

Spread it across too many ad sets and some of them never get enough to finish learning, so think of your weekly sales as a budget you spend on ad sets. 

Here’s how to work out yours, before you copy anyone’s build.

I still put my newsletter ahead of every social channel I use.

I know it’s a slow move, but I refuse to build on ground someone else can move. Each feed decides who sees my posts, and the algorithm changes without notice.

My list doesn't work like that.

When I hit send, it goes straight to the people who asked for it. Nobody sits in the middle turning the reach down. So I keep writing it, and my readers get my best thinking.

beehiiv newsletters grow 2.75x faster than average, because growth tools are integrated into the platform from day one. A recommendation network, a referral program, automations, analytics, magic links, and more, are all tools that can be used to grow your list.

In practice, the tools that help readers find and share your newsletter sit where you write it.

If you're starting a newsletter or giving yours a proper home, Start for free or use KODY30 to save 30% when upgrading to a paid plan 

Meta’s learning phase ends when an ad set reaches about 50 optimization events within 7 days.

Flip that around and you get the formula:

Ad sets you can feed = weekly conversions ÷ 50

Two examples (illustrative):

• 175 purchases a week ÷ 50 = 3.5

This account can support three ad sets. Each one gets about 58 purchases a week, which is enough to finish learning.

• 90 purchases a week ÷ 50 = 1.8

This account can support one ad set comfortably. Split it into two and each gets about 45 purchases, so neither reaches 50.

Your ad budget doesn’t decide how many ad sets you can run, and neither does the setup you saw in someone else’s post. Your weekly sales do.

A common reaction when an ad set is labeled Learning Limited is to launch a fresh set to make up the difference.

That spreads the same sales even thinner.

The better move is to merge the struggling ad set with the one most similar to it, and give the merged version the budget of both.

To find your number, open Ads Manager, add up last week’s conversions for the result you optimize for, and divide by 50.

Meta’s help center makes the case itself. Run too many ad sets at once and each one gets fewer chances to learn, so it stays in the learning phase longer and you spend more before results settle.

In the accounts I work on, the damage shows up in three places.

Starved ad sets. With only a thin slice of conversions each, several of them never exit learning, and their costs swing day to day.

Bidding against yourself. Two of your ad sets go after the same buyer, they compete in the same auction, and your own spend pushes up your price.

Reports stop telling you anything. With results spread over a dozen small ad sets, no single one has enough data to tell you anything. You end up making budget calls on noise.

One agency shared an apparel brand that went from 16 ad sets across five campaigns down to two Advantage+ campaigns, Meta’s automated campaign type.

Over the next six weeks, its cost per purchase dropped 23% with the same ads running.

That’s one account, so treat it as a reason to test, with no promise of the same result.

You may already know Meta favors fewer, more distinct creatives over piles of lookalike ads. Ad sets follow the same rule.

Sort your ad sets by results for the last 7 days and flag each one that came in under 50.

Here’s how to rebuild your ad sets in four steps.

1. Start from the number you just calculated. That’s how many ad sets you get.

2. Find ad sets going after the same result at the same price point, and merge them. Three broad ad sets pushing for one product at one price are really one ad set split three ways.

3. Keep an ad set separate only when something real sets it apart. That could be a different goal, a very different price point (a $40 single product next to a $300 bundle), or a setting you need to protect, like a cap on how much spend goes to existing customers.

4. Move your creative range into the merged ad sets. Meta learns which ads work by comparing them, so an ad set running three near-identical ads gives it nothing to compare.

Only merge ad sets that are true duplicates. Don’t swing too far the other way and put all your products and goals into one ad set. 

Products with different prices or goals still need their own ad sets, as in step 3.

One honest caveat: merging won’t fix weak ads, the clean structure just exposes them faster.

Merging counts as a major edit, so the new ad set starts learning from zero, and it needs about a week to settle.

Black Friday is November 27, so your last big change should happen around November 17 to 19 AT THE LATEST.

Saturday: add up last week’s conversions, divide by 50 and mark the duplicates.

Sunday or Monday: merge them, move the budget over and add the ads.

If small tweaks come up, save them for one session instead of making them as you go.

Feed the ad sets your conversions can afford. Cut the rest before Monday.

If you want a growth partner aligned with your real business goals, let’s chat.

Want to learn from me or connect on other platforms? 👇
Twitter • LinkedIn • Instagram • Threads