Most Facebook Ads agencies do their best. Most also have an awkward conflict of interest: telling you the truth about your account often shortens the engagement. Saying "your funnel is the problem, not our ads" is a sentence that ends a retainer.
So they do not say it. They keep optimizing things that are mostly fine while ignoring the leaks that actually matter.
This post is the 11 things they have on their internal Slack but do not share on the QBR.
#1. Your audience size does not matter.
Most account managers report on audience size like it means something. It does not. A 14-million-person audience is not 14 times better than a 1-million-person audience. The pixel learns to find the right buyers regardless of audience definition. What actually matters is creative quality and offer-market fit.
Brands obsess over targeting. Meta does not. Meta optimizes against your conversion event. The targeting is mostly cosmetic for accounts above $4,000 a month in spend.
#2. Broad targeting works in 2026.
The "broad targeting" debate ended in 2024. In 2026, brands at $14,000+ a month spend almost universally outperform on broad targeting versus interest stacks. The reasons are well-documented and have to do with how Advantage+ campaigns find buyers.
Your agency may still recommend "lookalike + interest" stacks because that was best practice in 2021. Best practice in 2026 is broad on conversion events. If your agency is not running broad on at least 47 percent of cold spend, ask why.
#3. Creative refresh schedules matter more than creative quality.
Counterintuitive but true. A "good enough" creative refreshed every 11 days outperforms a "great" creative left in market for 47 days. Audience fatigue compounds faster than most agencies model.
A working refresh cadence: 3 to 5 new creative variants per week, swapped in on Monday, paused on Friday if they fail to hit baseline CTR by Wednesday. Most agencies refresh "monthly" because their internal capacity does not support weekly. That is their problem, not your problem. You pay them to solve it.
#4. The 14-headline structure beats every other ad format.
Carousel ads. Video ads. Image ads. They have their place. The format that consistently outperforms across verticals is a single image plus a 14-line text variant where each line is a complete benefit statement.
The structure: 1 hook line, 11 benefit lines, 1 CTA line, 1 P.S. The reader scans the benefits, lands on one that resonates, then clicks. Conversion rates 1.7 to 3.4x video at one-fifth the production cost.
#5. CBO is not always better than ABO. It depends on spend.
Below $147 a day, ABO usually outperforms because Meta does not have enough signal to learn at the campaign level. Between $1,470 and $14,700 a day, CBO is almost always better. Above $14,700 a day, you should be running both with surgical separation.
If your agency uses one structure for all your campaigns regardless of budget, they are not optimizing. They are templating.
#6. Retargeting is overrated for most brands.
Retargeting captures attribution that prospecting earned. Most "retargeting wins" are actually prospecting wins reported in the wrong column. Brands at $4M ARR or below should run minimal retargeting (under 14 percent of total spend) and put the savings into prospecting creative.
The exception: cart-abandon retargeting in D2C. That works. The "engaged with your page in the last 90 days" retargeting that most agencies recommend is decorative.
#7. Your tracking is undercounting by 30 to 47 percent.
This is the one most agencies will never tell you. Since iOS 14.5, Meta's reported conversions undercount actual conversions by 27 to 47 percent in most accounts. Server-side tracking via the Conversions API closes the gap to roughly 7 to 14 percent.
Most accounts are not running CAPI. Or they are running it badly. The agency knows. They do not bring it up because fixing it requires technical work they do not want to scope. So your reported ROAS is artificially low, you panic, and they have an excuse to "test more creative."
Set up CAPI properly. Watch reported ROAS jump 1.4x with no other change. That is a free 1.4x sitting on the table.
#8. Pixel events most accounts never set up.
InitiateCheckout, AddPaymentInfo, ViewContent, Search, AddToWishlist. Most accounts have Purchase and Lead. They are missing the funnel-middle events that let Meta optimize correctly.
Without funnel-middle events, the pixel optimizes on whoever clicks "buy now" first, which biases toward existing customers and high-intent traffic that would have converted anyway. With funnel-middle events, the pixel learns to find net-new buyers.
This is a 2-hour technical setup. Worth approximately Rs 11,000 a month in efficiency on a Rs 4 lakh ad budget. Nobody scopes it.
#9. The forensic audit script.
A 17-minute audit any agency could run on any account. The output: a list of $4,800-a-month leaks. Top 3 leaks usually account for 87 percent of total waste.
The script: pull the last 30 days of campaign data. Sort by ROAS. Identify ads spending more than 7 percent of budget with ROAS below account average. Identify ad sets with frequency above 4.7 (audience fatigue). Identify creatives with CTR below 0.7 percent (creative fatigue). Pause the bottom 30 percent. Reallocate to the top 30 percent. Done.
Agencies do not run this script weekly because it makes their previous week look bad. The brands that run it themselves outperform the brands that wait for the QBR.
#10. The bidding strategy switch most accounts skip.
Manual CPC. Maximize Conversions. Target CPA. Target ROAS. They sound interchangeable. They are not.
A Manual CPC campaign with a $4.70 bid is not the same as a Maximize Conversions campaign that ends up at $4.70 average CPC. The first is constrained. The second is unconstrained. Outcomes differ by 1.7 to 3.4x in CAC.
The right strategy depends on volume, account history, and conversion-event reliability. Most accounts get it wrong because the agency picked one as default and never revisited.
#11. Why most agencies will not let you keep the audit script.
If you got the script, you would run it yourself. If you ran it yourself, you would not need the agency for "optimization." If you did not need the agency for optimization, you would only need them for creative. And creative-only retainers are 1/4 the price.
That is the conflict of interest. Not malicious. Just structural.
What you do with this.
Run the 17-minute audit script on your own account this week. Pull the campaign data. Sort by ROAS. Pause the bottom. Reallocate. See what happens.
If you want a second pair of eyes on it, the GetNos audit takes 30 minutes and walks you through the same process live on shared screen. We open your account. We run the script. We give you the 3 leaks costed in monthly rupees. You leave with the audit doc.
Free. Two clients per month is the cap on actual engagements. The audit itself is unlimited.
