Are all campaigns named with one consistent convention (e.g., {Account}_{Channel}_{Funnel}_{Audience}_{Date}) so reporting and search stay clean? Without a convention, filtering, pivoting, and handoff between team members become a guessing game. Is there exactly one canonical conversion source per platform — Google Ads native vs GA4 import — with no double-counting? Counting the same purchase twice destroys Smart Bidding signals and ROAS reporting accuracy. Is GA4 installed via GTM (not gtag direct), with enhanced measurement on and DebugView verified for every key event? GTM-deployed GA4 lets you fix tracking without code pushes — and verifies events before going live. Is the Meta Pixel firing on every key event AND is the Conversions API (CAPI) firing those same events server-side? Browser-only tracking now loses 20–40% of conversions to ad blockers, ITP, and iOS. Are Google Ads conversions imported from GA4 OR set up natively in Google Ads — but never both for the same event? Importing both ways double-counts and breaks Smart Bidding. Are Google Ads, GA4, Merchant Center, YouTube, and Search Console all linked under the same MCC/property? Unlinked accounts mean missing audience options, lost attribution, and broken remarketing pools. Is there a global negative placement / exclusion list applied across all video, display, and PMax campaigns? Without one, you'll find spend going to apps and sites you would never approve manually. Are internal/office IPs filtered out of GA4, and excluded from Meta retargeting audiences? Internal traffic pollutes audiences, inflates session counts, and biases optimization. Has the change-history log been reviewed in the last 30 days for auto-applied recommendations or unintended optimization-score edits? Google quietly auto-applies recommendations that can blow up performance overnight. Is billing healthy — primary card not expiring, backup payment method on file, no failed-charge pauses in the last 90 days? A 24-hour billing pause kills learning and resets Meta ad-set learning phases.
Have broad match keywords been eliminated on every campaign that isn't feeding Smart Bidding strong value signals? Broad match without value-based bidding is the fastest way to burn budget on irrelevant queries. Have you moved away from SKAGs unless legally required, and grouped keywords by intent into themed ad groups (5–20 kw each)? Modern match types make SKAGs counterproductive — themed ad groups give Smart Bidding more signal. Is there a shared negative keyword list at account level with 200+ entries, including misspellings of competitors? Most accounts under-invest in negatives and end up paying for unqualified clicks for years. Have search terms been reviewed in the last 7 days and irrelevant queries added to negatives? Without weekly review, broad/phrase match drifts into long-tail garbage. Do all RSAs have Ad Strength 'Good' or 'Excellent' with 8+ unique headlines and 3+ unique descriptions per ad? Low Ad Strength suppresses delivery and limits Smart Bidding's ability to optimize. Are headlines pinned ONLY when there's a legal or brand requirement — otherwise unpinned for performance? Excessive pinning kneecaps RSA learning and tanks Ad Strength. Are sitelinks (min 6), callouts (min 8), and structured snippets configured at campaign level with mobile-specific variants? Assets routinely deliver a 10–20% CTR lift — and they're free real estate on the SERP. Does the bidding strategy fit the goal — tCPA for lead-gen with 30+ conv/30d, tROAS for ecom, Max Clicks only during testing? The wrong bidding strategy outperforms only by accident. Are you feeding Smart Bidding offline conversion uploads — qualified leads, MQLs, revenue — not just form fills? Bidding on form fills optimizes for form-fillers, not customers. Are geo-targeting and ad-schedule bid modifiers set from actual performance data, not platform defaults? Default 'Presence or interest' targeting is the #1 source of out-of-area wasted spend. Are device bid modifiers (mobile/desktop/tablet) reviewed monthly and adjusted based on conversion data? Mobile and desktop convert very differently — a flat bid leaves money on the table. Are first-party audiences (customer match, site visitors, YouTube viewers) layered onto Search at Observation with bid adjustments live? Audience layering on Search is a quick win most accounts skip.
Are asset groups segmented by theme/category (not one mega-group) with 5+ headlines, 5 descriptions, 5 logos, 5 images, 1 video each? Mega-groups deny PMax the signal it needs to allocate spend; thin asset groups get throttled. Are audience signals populated with first-party customer lists and high-intent search themes — not just demographics? First-party signals shrink the learning period and lift early ROAS. Are brand keywords excluded from PMax via the Google Ads brand exclusion list (Account Settings > Brand Lists)? PMax cannibalizes branded search by default and inflates blended ROAS. Are search themes configured (max 25 per asset group) where appropriate, focused on incremental queries? Search themes guide PMax toward the queries you want without restricting it. Is the Merchant Center feed 100% approved — zero disapprovals and zero 'limited' items in priority markets? Disapprovals shrink your inventory in shopping auctions silently. Are GMC issue reports clean — no missing GTIN, no policy issues, no shipping/tax misconfigurations? Shipping/tax misconfig kills CTR in the SERP before the click ever happens. Is a supplemental feed in use for title/description enrichment, custom labels, and seasonal optimization? Title enrichment is the single highest-ROI shopping feed move you can make. Is item-level ROAS data being pulled (Google Ads Editor / scripts) and underperforming SKUs moved to a separate listing group? PMax overspends on a few SKUs by default — isolation gives you control. Are listing groups subdivided by performance tier (top sellers, profit margin, new launches) — not a flat structure? A flat structure forces PMax to treat your $5 trinket like your $300 hero product. Is brand traffic isolated from non-brand — via a Standard Shopping campaign with brand priorities, or PMax brand exclusions on non-brand? Without isolation, you can't tell incremental from organic-recapture spend.
Is the Meta Pixel firing on PageView, ViewContent, AddToCart, InitiateCheckout, Purchase, and Lead with correct parameters? Missing events kill optimization options and audience-building precision. Is the Conversions API (CAPI) installed server-side via Shopify native, Stape, or custom — firing all the same events as the Pixel? Without CAPI, you lose 20–40% of conversions to ad blockers and browser privacy. Is deduplication correct — matching event_id passed from browser and server, dedupe rate >70% in Events Manager? Bad dedupe inflates conversions and lies to Advantage+ optimization. Is the Event Match Quality (EMQ) score ≥6.0 on Purchase and ≥7.0 on Lead, with email, phone, fbp, fbc, IP, and UA passed? EMQ <6.0 means Meta is largely guessing which user converted. Is domain verification complete and the primary purchase domain owned in Business Manager? Unverified domains lose AEM event priorities and can be hijacked by an old agency or partner. Are the 8 Aggregated Event Measurement (AEM) priorities configured with Purchase ranked #1 on the primary domain? Misconfigured AEM caps your iOS attribution at lower-value events. Is iOS 14+ traffic handled correctly — first-party data prioritized, browser-only attribution accepted as floor? Pretending iOS hasn't happened is the most expensive mistake in this audit. Is the Pixel-to-CAPI dataset quality 'Good' or higher in Events Manager, with no recent quality drops? Quality drops precede performance drops by 1–2 weeks — early warning system. Is the Conversions API Gateway (CAPI-G) deployed, or is Stape / server-side GTM in place where first-party data is constrained? A gateway catches the events client-side tracking misses. Is Business Manager verified, and is domain/page/asset access scoped to the current team only — no ex-agencies, no expired logins? Old agency access is one of the most common sources of catastrophic account loss.
Are prospecting budgets set at the campaign level (CBO) unless you're deliberately forcing learning into specific ad sets (ABO)? CBO scales better at most spend tiers; ABO is a tactical tool, not a default. Is Advantage+ Shopping Campaign (ASC) running for ecom — and is manual prospecting only present where it beats ASC on incremental CAC? ASC outperforms hand-built prospecting for 70%+ of ecom accounts above $50k/mo. Have audiences been consolidated (fewer, broader) to escape the learning trap (≥50 conv/ad set/7d)? Slicing audiences thinly is the #1 reason ad sets stuck in learning forever. Is there a dedicated low-budget creative-testing campaign (broad audience, 3:1:1 structure) running continuously? Without a testing campaign, you're scaling on hope, not data. Are retargeting layers segmented by funnel stage — Viewers, ATC-no-purchase, IC-no-purchase, Past Purchaser — with separate offers and exclusions? One retargeting bucket sends the same message to people 14 days apart in intent. Are exclusions correctly stacked — past purchasers excluded from prospecting, current customers from acquisition, employees from all? Without stacked exclusions, you pay to re-acquire people you already own. Are 1% LALs built from high-quality seeds (top 25% LTV customers or 1k+ recent purchasers) — not 'all website visitors'? Garbage seed → garbage lookalike. Top-LTV seeds compound returns over time. Are value-based custom audiences and Value-Based Lookalikes in use for ecom, with LTV data passed to Meta? Value-based optimization shifts spend to the customers worth keeping. Are geo targets defined precisely — non-serviceable areas excluded, budget allocated by region performance? Default country-wide targeting wastes spend on regions you can't or won't serve. Is Advantage+ Placements on by default, with manual placements only where data justifies overriding it? Manual placement decisions usually cost performance unless you can prove otherwise.
Does every ad have a clear hook in the first 3 seconds — verbal pattern interrupt, visual hook, or text overlay? 70%+ of viewers decide whether to watch in the first 3 seconds. Is there a deliberate mix of UGC, founder-led, studio, and static creative — not over-reliance on one format? One format always fatigues first; format mix is your hedge. Are videos delivered in 4:5, 1:1, AND 9:16 — and statics in 1:1 and 4:5 at minimum? Letterboxed or center-cropped creative in Reels is a frequency killer. Are captions or SRT/embedded subtitles on every video for sound-off feeds? 85%+ of Meta video is watched without sound — no captions = no message. Is the CTA on every ad single, clear, and action-oriented (Shop Now, Get Quote, Book Call) — not 'Learn More' by default? 'Learn More' is the universal sign of an ad with no offer. Is brand recognition strong in every creative — logo, palette, fonts — without overpowering the hook? Unbranded ads can't compound brand equity, no matter how well they convert. Is creative refresh cadence enforced — winning ads paused at frequency >2.5 or CTR drop >25% within 7 days? Frequency creep is invisible until CPMs spike and ROAS collapses. Is the winning-ad scaling protocol documented — duplicate at 2–3× budget into new ad sets, not budget-edits on existing? Budget edits reset learning; duplicate-and-scale preserves it. Is there a static-to-video ratio at the ad-set level (we recommend 60/40 video/static for prospecting)? Static-only ad sets miss the algorithmic boost Meta gives video right now. Does every video have a custom thumbnail/cover frame — not Meta's auto-pick? Auto-thumbnails routinely pick a face mid-blink or a blurry frame. Is the primary text under 125 characters to avoid the 'See more' cut on Reels and Feed? Anything past 125 chars rarely gets read on mobile. Is the headline tested as a separate variable, with multiple headlines per ad and Advantage+ creative optimization on? Headlines are the single most testable creative element and the easiest to ignore.
Is blended ROAS (Shopify/CRM revenue ÷ total ad spend) tracked weekly — not platform-reported ROAS alone? Platform ROAS double-counts; blended is the only number CFOs trust. Is MER (Marketing Efficiency Ratio = total revenue ÷ total marketing spend) tracked with an agreed leadership floor? MER is the north star metric Meta and Google can't game. Is new-customer vs returning-customer revenue split tracked separately? If new-customer share is dropping, ROAS is flattering you with repeat business. Is a post-purchase survey running (Fairing, Knocommerce, DIY) asking 'How did you hear about us?' Survey attribution is the cheapest reality-check on platform claims. Is lightweight Marketing Mix Modeling in place (Recast, Pecan, or spreadsheet) for channels >$20k/mo? MMM is no longer just for enterprise; lite versions deliver 80% of the value. Is a weekly creative report produced — top 10 ads by spend with %, ROAS, frequency, and refresh status? Without it, creative refresh decisions are vibes-based. Are Meta ad sets respecting the 7-day learning phase — no edits or restarts on ad sets with <50 conv/7d? Editing in learning resets it. Most accounts do this constantly. Are you reaching ≥95% statistical confidence before declaring a creative or ad-set winner — not 'looks like it's winning' after 3 days? Premature winners ship a noisy result and lose the real one. Is budget pacing tracked daily against monthly cap, with a documented response for ±10% drift? Pacing drift is the most common cause of end-of-month panic-spend. Is at least one incrementality / lift test or geo holdout running per quarter on top-spend channels? Lift tests are the only way to know what your ads actually cause vs correlate with.
Has the landing page been reviewed against Meta + Google ad policies (claims, before/after, restricted categories) in the last 30 days? Policy drift causes the slow-strangulation account death you can't appeal. Are all health/financial/legal claims supported by disclaimers, footnotes, and accessible substantiation pages? Unsubstantiated claims are the #1 trigger for permanent disable. Have restricted-category settings (housing, employment, credit, alcohol) been correctly applied where required? Wrong category settings = auto-rejection, or worse, shadow restriction with no notice. Are admin accounts on Meta and Google using 2-step verification, tied to non-personal Gmail / work email addresses? Personal Gmail + no 2FA = one phishing email from losing the entire account. Is a backup payment method on file for BOTH platforms — and is the billing user a different person from the admin? Separation of billing and admin is a basic financial-controls safeguard.