Hey everyone,
Welcome back for another bite to chew on.
For years, the Meta playbook was about outsmarting the machine. We used complex campaign structures, manual bid caps, and secret audience hacks to find our customers.
Those days are dead.
We recently sat down for a webinar with the team at DigiCom, true practitioners who’re handling hundreds of thousands of dollars across multiple accounts every single day.
They confirmed what we’ve been feeling: the smart work has officially moved under the hood.
With the rollout of Andromeda in 2025, the Meta algorithm is now more powerful than any media buyer, but it’s also more sensitive.
If you feed it garbage data, it gives you garbage results.
If you give it a cluttered account structure, it gets confused.
The reality? You don’t win post-Andromeda by out-muscling the platform.
As the DigiCom guys put it, you win by becoming the ultimate service provider to the AI, feeding it cleaner signals, disciplined structures, and creatives that do the heavy lifting for you.
Today, we’re breaking down the DigiCom framework for shifting from hacking to feeding the machine.
On the menu:
Creative is the new targeting
Structure is about feeding the algorithm
Creative Hygiene + Signal Discipline = Profitabilityome back for another bite to chew on.
Creative Is the New Targeting
In the old world of Meta ads, you could launch a single creative, let it simmer for a few days, and wait for the algorithm to slowly find your audience. Those days are dead.
Post-Andromeda, Meta has become faster, less patient, and more ruthless.
If an ad doesn't spark immediate engagement or show a conversion signal, the algorithm deprioritizes it instantly.
No more wait and see. Now it’s "show me something now or I'm moving on."
Here is how to restructure your testing campaigns to survive the new high-speed reality:
Stop the Drip-Feed
Trickling out one new ad at a time is a recipe for failure. Most ads die before they collect enough signal to prove themselves. Instead of the old methodical approach, you need to batch your creatives.
The Power of the Batch
Drop 10 to 15 different creatives into a single batch.
By launching a mix of videos, statics, testimonials, and founder videos all at once, you arm the algorithm with a massive amount of data.
This gives the CBO the flexibility to immediately prioritize the ads showing the strongest initial signals.
Diversify the Persona, Not the Variable
It used to be best practice to isolate one variable, like testing three different headlines on the same image. Now, Meta often sees those as identical ads.
To win, you must introduce enough variation (different messaging angles, problem/solution hooks, and promotional styles) so Meta doesn't lump them all together as a single underperforming asset.
Let Creative Drive the Targeting
The goal is to simplify your account structure as much as possible. So start using creative to find different pockets of users.
When you provide a diverse batch of concepts, you're allowing the algorithm to do its job and find your customers for you.
The kicker: If your testing process is slow, your growth will be too. You'd better start testing concepts instead of headlines. Give Meta 15 reasons to find you a customer, rather than one reason to kill your ad set.
Structure Is About Feeding the Algorithm
The biggest mistake operators make is to build complex structures designed to force Meta to do what they want, rather than what the algorithm needs.
In the post-Andromeda world, structure is about shaping the environment.
If you try to fight the algorithm’s natural momentum, you will lose. If you feed it correctly, it will scale.
Here is the DigiCom 3-part simplified structure for building single-purpose environments:
1. Testing Campaigns
In the Andromeda era, testing is about feeding the machine all the data, so you have a clear path to win. A few things to make this successful:
Get the cleanest new customer data you can by setting the proper exclusions in place. Make sure you’re excluding your retargeting audience.
Move your testing into a CBO structure (Campaign Budget Optimization).
Launch 10–15 diverse creative concepts per batch, targeting different personas, content types, formats, etc.
Use minimum ad set spend limits. This is the only way to "force" enough spend to get valid learnings on new batches without letting a single high-performing legacy ad hog the entire budget.
Only graduate the creatives that are more efficient than your break-even point.
2. Scaling Campaigns
Once a creative graduates from the testing lab, it moves here. This environment is built for flexibility.
Run Broad audiences in tandem with Interest Targeting. The DigiCom team noted that while Broad is the goal, reintroducing interests helps Andromeda find specific pockets of users faster at scale.
Use 7-day click attribution and remove the 1-day view signal. This attribution setting ensures the algorithm is hunting for intent.
A prospect here must be net-new, which means zero prior interaction with your brand (no site visits, no carts, no previous purchases).
3. Retargeting Campaigns
Retargeting still works, but Andromeda can easily overspend here and fake your ROAS.
Use Advantage+ Retargeting, but set a maximum spend cap (usually 20-30% of your total budget).
This ensures you are harvesting existing intent without "cannibalizing" your prospecting budget or over-paying for customers who are already in your ecosystem.
But First, Warm Up the Engine
Do not wake up tomorrow and kill all your existing campaigns.
The DigiCom team is not fond of “best practices” because every case is different, so even though that’s the setup they see working best, it might not be the best for your brand.
So, run the new CBO framework side-by-side with your old setup. Let the new structure prove it can outperform your current metrics before you shift the full budget.
In a nutshell: Simplify the account into these three buckets. Use minimums to test, interests to scale, and caps to retarget. Then, monitor performance daily and let the algorithm do its thing.
Creative Hygiene + Signal Discipline = Profitability
The difference between a nine-figure account and an underperforming one is its creative hygiene. In simple words, you have to be super careful with your account management.
During the webinar, the DigiCom team did an actual audit of Joyride, and the gaps were glaring:
They had a 3.92 ROAS creative (nearly 3x better than the account average) sitting at a measly $30 budget.
Meanwhile, they had burned $10,000 on bottom-performing designs that drove only 14% of total conversions.
If you want to stop leaving money on the table, you have to audit your account against these three profit leaks.
Are You Starving Your Winners?
If a creative is working, you need to move on it daily.
You shouldn't be asking if it might work; you should be forcing spend until you find the point of diminishing returns.
In the Joyride account, the average CPA was $28, but their winner was hitting a $17 CPA.
Rule #1: If you find a creative winner, scale it immediately. Don’t let it sit at a low budget while you overthink it.
Are You Overfunding Your Losers?
Closing the doors behind things that don't work is just as important as finding things that do.
Joyride spent $10,000 on ads that were below average across the board.
Every dollar spent on a loser is a dollar stolen from a winner.
Rule #2: Ruthlessly cut the bottom performers. Take the learnings, walk a different path, and never look back.
If the hook rate or add-to-cart CVR is consistently lower than your average, kill the ad.
Are You Playing Data Blind?
Strict naming conventions are mandatory.
Without tagging every attribute (the persona, the hook style, the messaging angle), you can’t export and pivot the data.
You’ll never know why the Chihuahua video worked, only that it did.
Rule #3: Tag every element. This allows you to pull the thread on successful attributes and lift the floor of your creative performance by focusing only on the elements that drive signal.
The Part We Couldn't Fit in the Newsletter
Everything above came straight from our webinar with DigiCom. But we also sat down with them for a full podcast episode, and trust us, a lot didn't make the newsletter.
Like the brand that ran a quiz funnel against a standard landing page, and watched the winner flip halfway through the week because a single ad shifted spend.
Or the breakdown of how to diagnose an add-to-cart problem: is it the ad, or is it your page? (The answer changes everything.)
And yes, the intentionally misspelled ad that became a top performer.
If you enjoyed this newsletter, you’ll love this conversation. Here we discuss everything performance marketing from one operator to another
Sum It Up
The rollout of Andromeda has officially ended the era of the media buying guru. Now you win by becoming the best tester and a service provider to the AI.
If your data is muddy and your testing is slow, you are not playing with an advantage against your competitors.
The DigiCom recs:
Feed the beast with batches: Launch 10–15 diverse creative concepts at once to give Meta the data points it needs to find your winning pockets immediately.
Structure for signal, not control: For testing campaigns, move your budget into CBO to let the algorithm chase the momentum, but protect your downside with ad set spend limits.
Ruthless creative hygiene: Scale your winners daily and kill your losers before they steal your margin. Don’t forget to use strict naming conventions for maximum data gathering.
And above all, don’t forget this all changes from account to account, so test, test, test, and scale what works for you!
Let us know how we did...
All the best,
Ron & Ash