Happy Sunday, Everyone!
I hope you’re all surviving the frantic dash to complete your 2025 holiday shopping & to-do lists…all while simultaneously trying to plan for 2026, execute Q5 and enjoy a little bit of rest/relaxation/family time. There’s no doubt about it: holidays are always a challenge.
And, if all that wasn’t enough, there’s a never-ending refrain of "Expert Takes" flying in faster than Santa on his sleigh. Everyone has a theory on what you need to change, what you need to buy and what you need to fix.
The reality? 99% of it is noise.
Most think pieces, “guides” and audits are a waste of time because they focus on things that are easy to articulate and sound compelling, but ultimately have little impact on the relevant outcome. They obsess over tactical levers - CTR, Quality Score, Core Web Vitals, color palettes, immaterial preferences instead of the more complex, difficult-to-grapple with strategic drivers behind those choices. And, at the end of it, you get a document that makes you feel as if you’ve made progress on your journey to a higher performing marketing organization, when all it’s really done is run you in circles.
I’ve said it before: 90% of what happens inside your ad account is a direct result of everything outside of it.
If you are auditing your paid media today, you need to stop looking for "hacks" and start looking for "leaks."
Here are the 10 things you need to do to effectively audit your paid media machine, plus a critical look at why your P&L matters more than the pixel.
1. Financials First
Before you touch a single campaign, review the math for the business. Every ad account - Google, Meta, TikTok, whatever - is an extension of and accelerant to the business. The world’s best ad account won’t save a non-viable business. A 4.0 ROAS is meaningless if the break-even point is 5.0.
The Audit: Calculate your Contribution Margin per SKU/Service-Level SQL. Most accounts bid based on a blended average ROAS target. This is a mistake - one that (almost always) results in platforms under-spending on your high-value, high-margin products/services and over-spending on low-value, high-volume products/services. From a top-level perspective, this results in an acceptable ROAS/CPA – but because the underlying distribution is skewed, the business finds itself in an inferior position.
The first step of any successful audit is to align your ROAS/CPA targets with the actual profit profile of the specific product/service being sold/offered, not a blended average.
2. Data Passback & Validity
The inevitable result of automation and machine learning playing a progressively larger role in traditional ad account operations (targeting, bidding, optimization) is your control lever shifting to the data layer. Put another way: whoever has the best data wins. Some days, that’s you. Some days, it’s not. The role of the audit is to maximize the probability that your data is both accurate and timely.
Put another way: does the data in your pixel align with the money hitting your bank account or the leads in your CRM? If the data being used to optimize your campaigns is inaccurate, don’t be surprised when results it produces are off-target.
What To Audit: Check your "deduplication" logic and your CAPI (Conversions API) setup. Meta loves to claim credit for sales that would have happened anyway (view-through conversions). Worse, many setups double-count browser and server events, which inflates performance numbers and (far too often) results in brands over-investing in Meta. If your Event Match Quality Score is below 6.0, you are (effectively) flying blind.
On Google, the same basic logic applies. The optimal setup is one that uses enhanced conversions AND offline conversions. For eCommerce sites, be sure to restate conversion values after re-orders or returns/refunds – this prevents the algorithm from optimizing based on non-real events / people who don’t keep your product.
3. Apply the "IRIS" Framework to Your Creative
A few weeks ago, a friend/colleague sent me an audit he had received from another agency, asking for my thoughts/feedback. One of the biggest red flags I saw? The creative review + recommendations were all subjective + brand-focused, vs. actually analyzing the creative performance by angle/hook, style and audience.
A proper account audit reviews each creative based on alignment/resonance with a specific target audience segment, the psychological/emotional state and challenge/desire of that segment, and the relative diversity (style, format, hook, etc.) of creatives.
The Audit: Review your top 5 and bottom 5 ads against the IRIS framework:
- I - Instinct: Does it trigger a visceral reaction in 3 seconds?
- R - Resonance: Does it align with the user’s current emotional state?
- I - Imprint: Is it memorable enough to recall 10 minutes later?
- S - Signals: Does it clearly flag who this is for?
If your ads are "pretty" but fail the Instinct test, don’t be surprised when your ads fade into your audience’s “sea of sameness” background, becoming little more than wallpaper in their feed. The reality is that you need ads that break scroll behaviors in order to stand out today.
4. Assess Campaign Structure In Terms of Power Laws
We live in a winners-take-most world. Your account structure should reflect that.
The Audit: Apply the Golden Rule of Simplification: default to consolidation unless the value gained by fracturing a given ad set / ad group / audience is greater than the cost of doing so (that cost = more fragmented data, slower learning/optimization).
Why? Because power laws state that the top 20% of your creatives, ad groups + audiences are responsible for 80% of your results.
I see far too many ad accounts where a single high-performing audience/ad set was broken into 5 to “test” something, only to dig deeper and find that they’re all being served remarkably similar (if not the same) ads. The only thing this accomplishes is fracturing your data and preventing your winners from scaling.
Everything done in your ad account should be in service of identifying the winning ad groups / audiences / creatives / keywords as quickly as possible, then funneling as much of your resources to those winners until they reach a point of diminishing returns.
The corollary to this: if you’re going to split an audience/ad group/ad set for some reason – you want to serve that segment a different message, you want to hit on a different pain point, you have a tailored post-click experience for them - then actually take advantage of it.
5. Don’t Fall For Simpson’s Paradox
If you’ve read some of my articles, you’ll remember one of my favorite sayings: averages hides insight. That’s even true of the statement itself – there are multiple ways this plays out inside ad accounts:
Good Top Line Numbers Hide An Unhealthy Account: “good” blended numbers often mask what’s driving the underlying performance. Far too often, a prospective client will point to above-target ROAS or below-goal CPA, only for deeper analysis to show the account is being propped up by remarketing, existing customers or branded search. The top-level metrics look healthy (and their dashboards/reports often go up-and-to-the-right), but there’s a bloody reality hiding under the rosy picture.
Bad Top-Line Numbers Arise From Underlying Shifts: another common issue is when top-level performance appears to improve/decline, even when performance within segments is stable. In this case, the issue isn’t demand capture; it’s a shift in the underlying mix. If spend/delivery shifts toward historically weaker segments (for example, Android vs. iOS), blended ROAS will fall even though true efficiency has not changed. Nothing “broke per se - your problem/challenge is redirecting more of your resources back to the higher performing segments.
The audit: Break performance down by category (branded vs. non-branded vs. competitor), device (iOS vs. Android), and customer type (new vs. returning). This is how you separate artificial stability from real performance—and how you avoid reacting to what is, in reality, a mathematical illusion.
I walk through this exact trap in detail in my breakdown of Marketing Paradoxes here.
6. The "Times Square" SERP Audit
One of the biggest issues I see in audits/reviews: “experts” citing random metrics in G Ads or GSC, then making confident proclamations about what to do next.
It’s bullshit.
The only way to understand the full picture is to see what your audience sees - which means actually searching the relevant keywords/phrases from both a logged-in and logged out browser. That is the only way you’ll be able to understand what Google thinks the intent of the search is (if you see a AIO, it’s likely more informational; if you see a shopping carousel or 4-pack of ads at the top, it’s more transactional; if you see a map pack, it’s local), what your competitors are doing (take the time to read each ad - do they all say the same thing? Does one stand out? Are some hitting on different pain points?), and what else is being served (is there an AIO? People Also Asked? Knowledge Panel?).
If you see that most SERPs only have 1 ad at the top and 4 at the bottom, your strategy must shift to prioritizing Absolute Top Impression Share (Abs. Top IS) - because (being honest) most people aren’t even going to see - let alone click on - those ads at the bottom.
The Audit: Search your top 20 keywords from both incognito and a logged-in browser. Screenshot each SERP. Focus on:
- What is the overall layout of the SERP?
- How is Google evaluating this SERP (informational, transactional, navigational)?
- What other brands/competitors/alternatives are advertising?
- What offer is each other advertiser promoting? If a competitor has launched a "50% OFF" headline right next to your full-price listing, your drop in CTR isn't a copy problem; it's an offer problem.
- What other features (i.e. knowledge panel, map pack, AIO) are on the SERP?
7. Audit Your SEO & Paid Synergy
Paid and organic search are not enemies or competitors; they’re teammates. They share the same signals, similar intents and (often) the same SERP real estate. Treating them as independent channels is one of the most persistent - and costly - errors in modern marketing.
From the perspective of the user (which is the only perspective that matters), there’s one SERP - which means you (as a business) need one strategy to earn a click from that SERP. To figure out how to do that, there are 2 major questions to ask:
The 1st audit question is defensive:
Are you paying for branded search terms where you already rank #1 organically and there are no competitors bidding? If so, you are not “protecting the brand.” You are buying traffic you already own. In that scenario, paid spend does not create incremental demand; it cannibalizes organic efficiency and inflates perceived performance.
This does not mean branded search is always wasteful. The best analogy for branded search is insurance - it’s not a product intended to grow your net worth, but rather to protect it while other channels (i.e. non-branded search, social, video, traditional) grow it. And, just as you don’t hold insurance policies on every single thing in your real life, neither should you hold insurance policies on every branded search.
Branded search must earn its keep. If competitors are present, if the SERP is crowded with map packs, shopping units or AI summaries, or if branded paid materially improves conversion rate or downstream behavior, then it may make financial sense.
The 2nd audit question is offensive:
Are you relying on a mid–P1 organic ranking to carry a high-intent, non-brand keyword where the top four positions are dominated by paid ads? If so, you are functionally invisible. In today’s SERP layout, “page 1” is no longer a meaningful concept; screen real estate is. If your organic listing sits below the fold, sandwiched between multiple other similar-looking blue links, it may as well be on page 10.
Most users never reach the 10th result. Fewer still click it. Expecting organic alone to carry demand in a paid-dominated SERP is akin to letting the French fight a war: surrender by another name.
The goal is not to eliminate overlap; it is to maximize intent coverage. Paid search should be deployed where organic visibility is structurally constrained or competitively suppressed. Organic should be leveraged where authority and rank create durable, compounding returns. Together, they should maximize incremental capture of demand, not fight over credit for the same click.
If your SEO and paid teams are not planning + working together, you don’t have a cohesive search strategy; you have 2 disconnected tactics, both of which you’re overpaying for.
8. The "One-in-One-Out" Channel Rule
The mere existence of a channel is not a strategy. Presence is not progress.
One of the most common issues I see when auditing accounts is brands spreading themselves thin across platforms they feel obligated to be on. Pinterest. TikTok. Bing. Snap. Each one (on its own) feels smart, defensible, even strategic. But collectively, they fracture attention, destroy focus, slow learning + decision-making and handicap growth.
The audit: Examine your channel mix through the lens of return on attention, not just return on ad spend. If a channel contributes less than 5% of revenue but consumes 20% of your team’s cognitive load (reporting, meetings, creative iterations, “just in case” optimization), then the channel is just a drag on returns.
In today’s world, doing more almost always means doing worse. Focus is a superpower. In an ecosystem dominated by power laws - where 80% of the rewards accrue to 20% (or less) of the advertisers, you don’t need to be everywhere - you need to dominate somewhere.
This does not mean “play it safe.” It means commit fully where you have structural advantage and eliminate everything else. If you want to test a new channel, stop something else. Pause the lowest-performing channel and reallocate the attention. Else, you’re not experimenting - you’re just adding complexity and kneecapping your ability to compete on other channels where you can win.
Simplification is not retreat; it is strategy. Stop playing games where you can’t win.
9. Check for "Tilt" (Emotional Decision Making)
Poker players go on tilt after a bad hand. Media buyers do the same after a bad day. The result is the same in both cases: emotionally driven decisions masquerading as strategy.
This is a psychological audit, not a performance one.
The audit: Open your change history. Look for signs of panic - a flurry of bid changes late on a Friday night. Massive budget changes immediately after a rough reporting call. Wholesale restructures triggered by a single down day.
These are not optimizations. This is tilt.
Managing paid media today is like steering an aircraft carrier – it’s big, powerful and slow. You can’t steer it the same way you’d steer a speedboat.
In the same vein, ad platforms don’t reward reactivity – they reward consistency, clean data/signals and controlled experimentation. Emotional interventions have the same impact on the ad account as dropping a port side anchor + making a sudden hard turn have on an aircraft carrier: things go flying. And not in a good way.
Rash changes inside an account disrupt algorithmic optimization, obscure performance insights and make it impossible to distinguish normal variance from actual issues.
The reality is that most accounts don’t underperform because the algorithm is broken; they underperform because the operator lacks discipline. Short-term volatility is inevitable - but responding to it impulsively is optional. If your change history tells a story of constant intervention, the issue isn’t Meta, Google or TikTok - it’s the person running them.
Great media buyers know when not to act. They separate noise from signal, protect the system from their own anxiety and let probability play out over a meaningful sample size.
If you see tilt in your change history, you don’t need a new platform. You need better process and stronger nerves.
10. The "Red Team" Interrogation
Finally, audit your strategy by trying to break it. If you can’t articulate how your account fails, then you don’t understand why it works in the first place.
The audit: Ask a simple, uncomfortable question: If this account goes completely sideways next month, what caused it?
Resist the urge to give a vague answer. Look for the single point of failure: the assumption everything else depends on. The keystone holding up the entire structure.
It might be one Meta campaign or audience that has historically outperformed. It could be a single creative that’s a year old, but somehow still delivering 60%+ of the account’s leads. Maybe it’s a single keyword, a specific video on YouTube or a single traffic source (like organic or referrals) driving a disproportionate share of the outcomes.
Stability built on concentration is fragility. Just as over-diversification can lead to across-the-board mediocrity, over-concentration in a single channel/tactic/creative can lead to collapse when (not if) something breaks.
Most strategies don’t fail because they’re fundamentally flawed or tragically misinformed (though I’ve seen my fair share of those). They fail because they’re brittle. They work - until something changes, competition increases or the one thing propping them up stops working.
The fundamental job of any marketer isn’t to win today; it’s to build a system that is anti-fragile - that gets better/stronger when things go wrong. Doing that requires you to identify dependencies, diversify intelligently, and build redundancy.
Markets punish blind spots. Find yours before the market does.
Where Tools Like Optmyzr Make This Possible
You cannot focus on "Financial Physics" (Point #1) or apply the "IRIS" framework (Point #3) if you are buried in the weeds of manual bid adjustments and cross-platform reporting.
The 10 audit points above require high-level strategic thinking. But high-level thinking is impossible if your day is consumed by low-level manual labor.
Optmyzr is the difference between an auditor who spots the problems and an operator who actually fixes them.
- Audit on Autopilot: Use the Rule Engine to automatically flag "Simpson's Paradox" anomalies, broken URLs, or data drop-offs before they destroy your week.
- Prevent "Tilt": Set guardrails that stop you (or your team) from making emotional, panic-driven changes when performance dips. Let the system enforce discipline.
- Enforce Structure: Use Blueprints to ensure your account adheres to the "Power Law" structure we discussed in Point #4, preventing the fragmentation that kills performance.
Stop drowning in the "middle" of your ad account. Let the machines handle the hygiene so you can focus on the strategy.
|
The Bottom Line
Marketers today spend an inordinate amount of time trying to reverse-engineer the "black box" that is platform algorithms. It’s fun. It’s interesting. It’s virtually guaranteed to get engagement and comments.
But the real opportunity isn't in pulling back the curtain on the ever-changing, constantly-evolving machine - it's in fixing the foundation of your business so you can win no matter what the machine does.
Jason Kelce - the legendary Philadelphia Eagles Center - said it best: “If you don’t want the refs to decide the game, play the game so well that the refs can’t decide the game.” The same is true of marketing - if you don’t want Zuck’s mood to control your destiny, build a business that either (a) can survive Meta’s vacillations or (b) doesn’t care about them at all.
Cheers,
Sam
Loving The Digital Download?
Share this Newsletter with a friend by visiting my public feed.
Follow Me on my Socials