12 months of audits, millions in ad spend, 5 takeaways

· Articles

12 months of audits, millions in ad spend, 5 takeaways

The game didn't get harder in 2025. The tactics that worked in 2021 just stopped working, and most accounts are still running them.

Phil Kiel
@PhilKiel · Dec 22 2025

10 years running Meta ads. Here's what actually matters:

Creative quality beats everything. Your account structure can't save bad ads. Your targeting can't save bad ads. Your bid strategy can't save bad ads.

If your creative doesn't stop the scroll, nothing else matters.

Account structure is the second variable. Most accounts are sabotaging creative performance with fragmented campaigns, over-segmented ad sets, and too many ads competing for limited budget.

Exclusion lists matter more than targeting. Stop trying to tell Meta who to target. Start telling it who NOT to target. Purchasers, email list, high-intent site visitors, exclude them all and let the algorithm find everyone else.

New visitor percentage is your leading indicator. If it's dropping, Meta is struggling to find efficiency with new audiences. 70% is the sweet spot.

Testing velocity beats testing volume. Brands that launch 8-12 new concepts per month outperform brands that launch 40 variations of the same concept. Conceptual diversity > asset volume.

The rest is noise. Platform updates, new features, beta access, none of it matters if your creative is weak and your structure is broken.

Fix those two things first. Then optimize the details.

↗ View on X

After auditing hundreds of Meta accounts over the past year, the pattern is clear: the brands scaling past $100k/month in 2026 are doing fundamentally different things than the brands stuck at $30k wondering why performance is inconsistent.

This isn't about new features or beta access. It's about understanding how Meta's algorithm actually works in 2025 and building your account structure and creative strategy accordingly.

Here are the five focus areas that will determine who wins and who gets left behind in 2026.

Focus Area 1: Campaign Consolidation Over Fragmentation

The Problem:

We're regularly auditing accounts spending $30-50k/month split across 20-50 campaigns. Each campaign gets $50-200/day. Most are stuck in permanent learning phase. Performance is inconsistent. The founder spends hours every week trying to figure out which campaigns are working and which aren't.

Three campaigns improve. Three get worse. Next day it reverses. You're constantly seesawing.

This is the hardest way to run an account.

Phil Kiel
@PhilKiel · Jan 08 2026

For every drop shipper day trading a million ad sets there's a Chad with scaling a single budget.

Phil Kiel
@PhilKiel · Jan 07 2026

Last year we did $20 mil top line with this setup and we're not planning on changing it:

ASC (v1) Single Ad Set - we launch new ads here, 80% of daily spend

CBO (legacy campaign) ROAS bid, 4x ad sets

At a time when the brands competitors are struggling, this brand 2x YoY, with actually less campaigns and ad sets.

The focus area for this account:

- new creative

- new gifting and partnerships

- testing upper funnel

- tracking net new reach (v1 ASC doesn't have full exclusions)

↗ View on X
↗ View on X

Why This Happens:

When performance dips in a single campaign, the natural instinct is to duplicate it at a lower budget to "control" CPA. The problem? You're now forcing Meta to run separate auctions for the same ads against the same audiences. Each campaign competes with itself for delivery.

Result: fragmented data, slower learning, inconsistent performance across all campaigns.

The 2026 Standard:

3-5 campaigns maximum for most accounts:

  • One prospecting CBO (all cold traffic creative testing)
  • One Advantage+ (proven winners, scaled delivery)
  • One retargeting campaign (if purchase volume supports it)

That's the structure. Everything else is overthinking.

Why This Works:

Meta's algorithm optimizes at the campaign level. When you consolidate budget, you're giving the algorithm more signal in fewer places. The campaign accumulates conversion data faster. The learning phase actually completes. Performance stabilizes.

The brands spending $40k/month across 3 campaigns consistently outperform brands spending $40k/month across 20 campaigns. And they're working less to manage it.

The Objection We Hear:

"But when I scale one campaign, CPA spikes for 3-5 days."

Yes. It does. That's the algorithm expanding reach beyond your warmest audiences. The solution isn't to fragment into 20 small campaigns to avoid the dip. It's to accept the dip as part of scaling and build a routine that removes decision-making fatigue.

Monday: increase budget 20% Wednesday: launch new creative Friday: pause underperforming ads

Boring. Systematic. Predictable. Stop jet skiing. Start cruise shipping.

Focus Area 2: Creative Diversity Over Creative Volume

The 318 Ads Problem:

Just audited an account burning $40k/week where 318 ads were active. Only 10 were getting meaningful spend. 55% of budget flowed to those 10 ads. The other 308 received pennies.

Months of creative effort invisible to the algorithm.

The founder saw this data and thought: "We need better creative."

Wrong diagnosis.

The Real Problem: Creative Sameness

When your ads are too similar, Meta's Andromeda retrieval system treats them as the same thing. It picks a winner early and starves everything else of delivery.

You're not testing 318 concepts. You're testing minor variations of the same concept 318 times.

Phil Kiel
@PhilKiel · Oct 17 2025

A brand came to me yesterday with $140 CPMs.

"Our CTR is 2%. Conversion rate is 4%. But we're getting killed on cost."

I knew the issue immediately.

Every ad looked the same to Meta.

Same product shot. Same background. Same angle.

Meta served them all to the same 10,000 people. All increasing: Frequency, CPMr, CPC.

The Cost To Advertise isn't about your bid.

It's about creative diversity.

↗ View on X

What "Creative Diversity" Actually Means:

Different concepts = different:

  • Emotional drivers (fear vs aspiration vs belonging vs status)
  • Objections handled (price vs efficacy vs trust vs convenience)
  • Formats (UGC talking head vs product demo vs static vs carousel vs founder story)
  • Audiences addressed through the creative (busy parents vs fitness enthusiasts vs budget-conscious vs premium buyers)
  • Proof types (customer reviews vs before/after vs expert validation vs testing data vs social proof)

Stop making 40 variations of the same talking head with different b-roll. Start testing conceptually distinct angles that unlock different audience segments.

Phil Kiel
@PhilKiel · Nov 09 2025

I put together 10+ creative inspiration boards with 200+ ads across every niche you can imagine.

If you run paid social, this thread will keep your next 6 months of creative testing stocked 👇

↗ View on X

The Creative Volume Trap:

At $40k/month spend, you need 8-12 new concepts per month to maintain performance. That's 24-48 total assets when you account for 3-4 variations per concept.

Most brands at this spend level are testing 3 concepts per month and wondering why performance is inconsistent.

They find one winner, scale it to $1,200/day, watch it crater in week 3, then scramble to find the next one. No pipeline. No redundancy. No systematic approach.

The 2026 Reality:

Creative production capacity is the actual bottleneck to scale—not budget, not account structure, not targeting sophistication.

If you're spending $40k/month and only producing 3 concepts monthly, you're either:

  1. Overworking your winning ads into the ground
  2. Underdeploying your budget
  3. Both

The accounts that scale past $100k? They built creative infrastructure first, then added budget.

Focus Area 3: Leading Indicators Over Lagging Metrics

The Metrics That Actually Predict Performance:

Most brands optimize for ROAS and CPA. Those are lagging indicators—they tell you what already happened.

The brands winning in 2026 are watching leading indicators that predict what's about to happen:

Phil Kiel
@PhilKiel · Nov 27 2025

If your account feels stuck, it’s probably these 4 issues:

Rising CPMr

Declining reach

Frequency blowouts

Low creative win rate

👉 Bookmark this as an audit checklist.

DM me if you want the full audit process we use to scale brands.

↗ View on X

New Visitor Percentage

Should sit around 70%.

If it's above 75%, Meta is struggling to find efficiency and burning budget on cold prospecting only. You're constantly reaching strangers at the coldest possible conversion temperature.

If it's below 60%, you're over-indexed on retargeting and growth will stall. You're recycling the same warm audience without expanding reach.

70% is the sweet spot—balanced cold/warm mix that drives both efficiency and growth.

Net-New Reach Percentage

Your leading indicator for incremental demand. This shows what percentage of people seeing your ads this week have never seen them before.

When net-new reach drops while frequency rises, you're saturating your audience. CPA will follow within days.

We just analyzed an account where:

  • Year-over-year spend down 35%
  • ROAS flat
  • Total reach down from 34M to 19.6M
  • Net new reach for the year: 29%

That means 71% of people seeing their ads had already seen them before.

The founder thought lower spend + same ROAS = efficiency win.

Actually: reach saturation. Growth ceiling hit. Can't scale new customer acquisition when 70% of ad delivery goes to people already in your funnel.

Frequency Above 5.5-6 for 7+ Days

This is your early warning system for creative fatigue. When frequency stays elevated for a week and CPA starts trending up, it's time to rotate concepts immediately.

You're beating the same people over their heads. The algorithm can't find new people to show your ads to. Either your creative has saturated or your exclusions are too tight.

The Decision Framework:

Don't judge ads on early ROAS. Judge them on whether they hit your CPA target at meaningful spend thresholds:

  • $500 minimum for sub-$100 AOV products
  • $1,000 minimum for $100-300 AOV products
  • $2,000+ minimum for higher AOV

Below these thresholds, you're looking at noise, not signal.

When an ad scales from 4x ROAS at $500 spend to 2.5x ROAS at $5,000 spend, that's not failure. That's the algorithm expanding reach beyond your warmest audiences. If your target is 2x, the ad is winning.

ROAS compression at scale is normal algorithm behavior. Stop killing winners prematurely.

Focus Area 4: Exclusion Lists Over Targeting Precision

The Attribution Trap:

Customer sees your Meta ad → searches your brand on Google → buys through Google Shopping.

Meta gets nothing. Google gets the conversion. Your reporting says "Google is profitable, Meta isn't."

So you shift budget to Google. Performance tanks. Meta was driving the demand that Google was capturing.

The Bigger Issue: You're Retargeting Your Own Organic Audience

Phil Kiel
@PhilKiel · Nov 05 2025

Last week I made an offer to audit Meta ad accounts if anyone was struggling during Q4.

Here's the issues that are causing the most trouble and how to solve them... 👇

⚠️ Most accounts were bleeding money because they were fragmented to death.

The worst account had 10+ campaigns for a $35K monthly budget.

The fix? One campaign. Maximum two.

⚠️ The silent performance killer? Your new visitor percentage.

Every single account below 70% is showing clear signs of decay.

Meta is serving your ads to the same warm audiences over and over again, driving up Frequency and CPMr and tanking growth.

"But I need separate campaigns to test different ideas!"

No. You don't.

55% of spend in one account was going to just 10 ads... out of 318 live.

The other 308 ads? Just burning money on management time.

⚠️ Website visitor exclusions missing from EVERY account I audited.

This isn't 2022. You can't just target broad audiences without excluding past visitors. Awareness growth is a key problem this year.

Add these exclusions at the ad set level:

All website visitors (180 days)

Purchasers

Email list

Social engagers

⚠️ Low cost caps and bid caps are throttling your new creative.

One account had ad sets with bid caps of $5 that weren't spending... yet the CPA was $35.

What happens if you increase it? Or test lowest cost instead?

⚠️ View-through attribution is making you feel better than you are.

I saw accounts where removing 1-day view brought ROAS down 50%.

If you're getting sales from multiple channels, Meta is taking credit for sales it didn't drive.

⚠️ "Worldwide excluding US" targeting is lazy strategy.

One account could immediately boost performance by targeting just 4-5 top-converting countries instead.

Tighter targeting = better delivery = more consistent results.

⚠️ The boring fix nobody wants to hear: naming conventions.

Try analyzing which creative angles work when your ads are named:

"Copy of Copy_V3_NEW_Final2"

Proper naming unlocks patterns you can't see otherwise.

✅ TLDR: Meta success in 2025 comes down to:

Account consolidation (1-2 campaigns for most budgets is enough)

Proper exclusions (target 70%+ new visitors)

10-15 quality ads (not 300+ variations)

Test bidding type if your current isn't fit for purpose

Click-only attribution

Focused geo-targeting

If you're struggling and want me to record a quick 5 minute Loom, drop me a DM.

↗ View on X

Just reviewed an account with 6-8 million organic impressions per month. Strong comedy-focused social content. People follow for entertainment, not purchase intent.

Their Meta ads had zero social engager exclusions. Result: paid ads were retargeting comedy sketch fans (low intent) instead of true purchase-intent traffic.

They were burning 30-40% of paid budget on people who already see their content organically and have shown zero buying signals.

The 2026 Standard for Exclusions:

Minimum exclusions for cold prospecting campaigns:

  • Purchasers (180 days)
  • Email list
  • High-intent site visitors (cart viewers, product page 90+ seconds)
  • Facebook engagers (180 days)
  • Instagram engagers (180 days)
  • Website visitors (180 days) *if your organic content is highly social

Why This Matters More Than Targeting:

Stop trying to tell Meta who to target. Start telling it who NOT to target.

Meta's algorithm in 2025 is sophisticated enough to find your buyers when given broad targeting parameters. But it's not sophisticated enough to know that the person who laughed at your Instagram comedy sketch yesterday isn't the same person who wants to buy your product today.

That's your job. Exclusions are how you teach the algorithm the difference.

The Test Structure:

Campaign 1: 80% of budget, current exclusions (purchases only) Campaign 2: 20% of budget, full exclusions (purchases + website visitors + social engagers)

Run this for 14 days. Compare new visitor %, conversion rates, and new customer CPA.

Brands with strong organic social and weak paid performance? This test almost always reveals that organic audiences are sabotaging paid efficiency.

Focus Area 5: Systematic Creative Production Over Ad Hoc Creation

The "5 Ads Per Week" Reality:

We get on calls with founders who say: "My goal is 5 new ads per week."

Actual output: 1 ad in the last month.

This isn't a discipline problem. It's a structure problem.

The founder is running the entire business. Product development, customer service, operations, fundraising. At 5pm on Thursday they remember "I need to make ads today."

They open Canva. They stare at a blank canvas. They close Canva. They'll do it tomorrow. Tomorrow never comes.

The Creative System That Actually Works:

Phil Kiel
@PhilKiel · Oct 12 2025

We added 4 elements to our creative briefs and our hit rate jumped 23%.

Now they're non-negotiable.

Here's what changed (and why it matters) 👇

↗ View on X

Step 1: Pick 2-3 Core Pillars

Not 20 angles from ChatGPT. Not "let's test everything." 2-3 pillars that represent your strongest value propositions.

Phil Kiel
@PhilKiel · Oct 09 2025

A good brief isn’t a script, it’s a sandbox.

We define the concept + key message, some creative treatments, then let designers create multiple interpretations.

Same strategy, different executions.

That’s how you unlock variety that still performs (and scales).

↗ View on X

For a male personal care brand:

  1. Performance/results
  2. Convenience/routine
  3. Confidence/identity

Every ad speaks to one of those three things. No more angle brainstorming. No more analysis paralysis.

Step 2: Build The Weekly Routine

Monday: increase budget 20% Wednesday: launch 2 new ads (1-2 concepts from your pillars) Friday: pause underperforming ads based on spend thresholds

Not "whenever I have time." Not "when performance dips." Not "when I feel inspired."

Monday, Wednesday, Friday. Every week. Remove decision fatigue.

Step 3: Quality Over Quantity

2 conceptually different ads per week beats 10 minor variations.

Different emotional driver. Different objection. Different format. Different proof type.

Give Andromeda actual diversity to work with, not 10 versions of the same thing.

Phil Kiel
@PhilKiel · Oct 28 2025

Creative format we're seeing work across a number of accounts:

B-Roll background video with text overlay, 7-12 seconds long, one or two shots max.

Will share examples this week.

↗ View on X

Step 4: Leverage What You Already Have

Brands with strong organic content already have the creative engine. They just aren't amplifying it through paid.

Take your top-performing organic content from the past week. Drop it into your testing campaign. Let it get spend. Kill it if it doesn't perform.

This isn't about "boosting posts." It's about systematically testing whether your highest-engagement organic content can drive conversions when backed by budget.

Phil Kiel
@PhilKiel · Nov 26 2025

First time seeing this Checkout style BFCM ad.

It’s not even a mockup - just go to your checkout, screenshot and let it rip.

With/without overlays.

↗ View on X

Partnership ads with creators you already work with give you delivery preference without additional creator cost. Meta's algorithm prioritizes partnership content in the auction.

Examples of what's working right now:

Phil Kiel
@PhilKiel · Dec 29 2025

New year New You is around the corner, here’s 4 supplement statics I’ve saved this week 📍

Start with the headline

Add USP and value prop call outs

Feature your offer with an accent colour

For me the Huel ad is the stand out as it uses a relatable at-home shot of the product

↗ View on X
Phil Kiel
@PhilKiel · Dec 30 2025

Static ads from the creative team 📍

↗ View on X
Phil Kiel
@PhilKiel · Jan 03 2026

New statics from the creative team 📍

↗ View on X

The Bottleneck Is Infrastructure, Not Ideas:

Most brands aren't failing because they lack creative ideas. They're failing because they haven't built the infrastructure to consistently produce and test those ideas.

At $30k/month, you need 6-8 new concepts monthly. At $50k/month, you need 10-12. At $100k/month, you need 15-20.

If you're the sole creative producer and you're struggling at $30k/month, you won't make it to $60k/month without either:

  1. Hiring creative help
  2. Building a sustainable 2-ads-per-week system
  3. Both

The accounts spending $100k+? They solved creative production before they hit $50k/month. They didn't try to scale budget first and figure out creative later.

Tracking What Actually Works:

Phil Kiel
@PhilKiel · Dec 20 2025

New app in development: Creative Win Rate Tracker

For brands: measuring win rate per format, angle, creator, month (date ad launched)

For the agency: win rate per ad account, designer, creative strategist and total win rate for the agency

↗ View on X
Phil Kiel
@PhilKiel · Nov 27 2025

A useful update to the app, I've added highlights showing when budget adjustments were made, helps to show what impact the change had.

↗ View on X

The Optimization Hierarchy: Where To Actually Spend Your Time

Most accounts are optimizing in the wrong order. They're blowing up entire campaigns when one ad set underperforms. They're leaving underperforming ads running for weeks because "maybe they'll improve."

The 2026 Hierarchy:

Pause campaigns: RarelyOnly if the entire structure is fundamentally broken. If you're pausing campaigns weekly, you're creating chaos, not optimizing.

Pause ad sets: More oftenWhen specific concepts fail at scale. If an ad set gets $500+ spend and converts at 3x your target CPA, kill it. The concept doesn't work.

Pause ads: Most oftenWhen individual creative executions fail. This is where most of your optimization time should go. Kill bad ads. Keep good ad sets running.

The Pattern We See:

Founder launches campaign. Runs for 5-7 days. Performance isn't perfect. Pauses entire campaign. Launches new structure. Repeat.

Result: 31 campaigns in 3 months. Account ROAS stuck at 0.6x.

Meanwhile, one of those paused campaigns was doing 0.95x ROAS (above average) with three ad sets:

  • Ad set 1: 1.2x ROAS
  • Ad set 2: 1.4x ROAS
  • Ad set 3: 0.3x ROAS

The fix? Pause ad set 3. Let the other two keep running. The campaign structure was working—one element inside it wasn't.

You can't find a winning structure if you never give a structure enough time to optimize. Pick one. Stick with it for 30+ days. Make changes at the ad and ad set level, not the campaign level.

Action Plan: Your 2026 Roadmap

Week 1: Audit Current State

  1. Count your active campaigns. If it's more than 5, consolidation is your top priority.
  2. Review creative diversity in your top-spending campaign. Are you testing concepts or variations?
  3. Check new visitor % and frequency. Are you in the 70% / <5.5 sweet spot?
  4. List your exclusions. Are you excluding social engagers and site visitors?

Week 2-4: Structural Consolidation

  1. Build your consolidated structure (1-3 campaigns)
  2. Migrate ads into new structure gradually (don't pause everything at once)
  3. Set your Monday/Wednesday/Friday routine
  4. Extend ad evaluation thresholds to appropriate spend levels

Month 2: Creative System Build

  1. Define your 2-3 core pillars
  2. Audit existing organic content for paid amplification opportunities
  3. Set up partnership ads with existing creators
  4. Build your concept pipeline (2 new concepts per week minimum)

Month 3: Leading Indicators Focus

  1. Add new visitor % and net-new reach to your weekly reporting
  2. Set up proper holdout testing for incrementality measurement
  3. Track frequency as early warning system
  4. Optimize based on leading indicators, not just lagging ROAS

The Fundamental Shift: From Tactics to Marketing

Meta's increasing control over advertising mechanics means we're moving from a media buying tactical game to a marketing game.

Growth hacks are fading. The era of "split test everything" and "find the perfect audience" is over.

Creative quality, message clarity, and systematic production are now the primary levers you control.

This is good news for real marketers. Bad news for agencies selling "proprietary targeting strategies" and "secret campaign structures."

The brands that win in 2026 will be those that:

  • Build creative infrastructure before scaling budget
  • Optimize for leading indicators that predict performance
  • Consolidate account structure to feed the algorithm better data
  • Understand that creative diversity unlocks algorithmic efficiency
  • Build systematic routines that remove decision fatigue

The brands that lose will be those still:

  • Running 20+ campaigns hoping to find the magic structure
  • Killing ads at $500 spend because early ROAS dropped
  • Testing 40 variations of the same concept
  • Excluding only purchasers while retargeting their entire organic audience
  • Trying to day-trade their way to scale

TL;DR: The 2026 Playbook

Phil Kiel
@PhilKiel · Jan 04 2026

Simple Scales.

↗ View on X

Campaign Structure:3-5 campaigns maximum. Ad set per concept. 10-15 ads per ad set. Consolidation beats fragmentation.

Creative Strategy:Conceptual diversity over asset volume. 8-12 new concepts monthly at $40k spend. Format variety prevents saturation.

Key Metrics:New visitor % at 70%. Net-new reach as leading indicator. Frequency below 5.5-6. Judge ads at spend thresholds, not early ROAS.

Exclusions:Purchasers + email + website visitors + social engagers. Stop telling Meta who to target. Start telling it who NOT to target.

Creative Production:2-3 core pillars. Weekly routine (Monday/Wednesday/Friday). Quality over quantity. Leverage organic content.

Optimization Hierarchy:Pause ads most often. Pause ad sets sometimes. Pause campaigns rarely. Stop blowing up structures when one element fails.

The gap between winners and losers is widening. The brands that adapt to these fundamentals in Q1 2026 will have a massive advantage over those still running 2021 playbooks.

The time to adapt is now.

Phil

Back to all articles