The Pre-BFCM Account Audit for Meta, Google and YouTube

· Tutorials/Guides

The Pre-BFCM Account Audit for Meta, Google and YouTube

Last Black Friday, the median account in our dataset hit close to 4x its normal daily spend on its peak day. Across the busiest 30 days of the season, that same median account ran at nearly 2x its usual pace. Every bidding decision Meta, Google and YouTube make during that stretch happens at the highest volume of the year, with each one only as good as the data feeding it.

The worst time to discover your Meta purchase events are double counting is the morning of November 27. A few hundred Shopping products disapproved overnight for a price mismatch tells the same story on Google. Almost every BFCM problem we find in accounts traces back to something already broken in October, which is why this audit is worth running now while there is still time to fix what it turns up.

This checklist covers the 6 areas that matter most across all 3 platforms, which are tracking and CAPI health, Enhanced Conversions, Merchant Center disapprovals, feed accuracy, landing page speed and inventory sync.

The Audit Is Also How You Read BFCM Week Correctly

CPM rises across  pretty much every account during BFCM, because every brand in your category is bidding for the same people in the same week. That pressure pushes CPMr up even in accounts with no saturation problem at all, so a rising CPMr in late November tells you very little on its own.

CPMr, or Cost Per Mille Reached, is the cost to reach 1,000 unique people, calculated as CPM multiplied by frequency. Phil Kiel, our Director of Paid Social, built much of how we diagnose Meta accounts around this metric.

The signal lives in the split between its 2 inputs. When CPM climbs while frequency holds flat, you are looking at auction pressure, which usually eases after Cyber Monday. If frequency climbs while CPM holds flat, your budget is cycling through the same audience, which calls for exclusions and consolidation over patience. The headline number looks identical in both cases, so you can only make the call with a clean October baseline for CPM, frequency, CPMr and net new reach to compare against.

Google has its own version of this problem. Smart Bidding sets bids from conversion data, so purchases that are double counted or missing in October send the system into November learning from the wrong signals. On YouTube, video campaigns report unique users and frequency, which means CPMr can be tracked exactly as it is on Meta against the same kind of baseline.

None of those baselines are worth much if the tracking underneath them is wrong, which is where the audit starts.

Meta Tracking Breaks Quietly, So Check It Event by Event

The fastest sense check is to compare a week of Meta-reported purchases against backend orders for the same week. Meta will never match exactly, but a gap that has widened since summer is usually the first sign something has changed in the setup. From there, run these checks in Events Manager.

  • Check Event Match Quality on the Purchase event. A score of 6 or above is workable, with 8 or above counting as strong. Low scores usually mean email, phone or external ID are missing from the server-side payload.

  • Check the deduplication rate between pixel and Conversions API events. Both sources need a matching event ID, otherwise purchases get double counted or one source gets discarded.

  • Confirm every Purchase event carries value and currency. Value-based bidding and ROAS reporting both break without them, with hardcoded values turning up surprisingly often after a theme update.

  • Clear any active warnings in the Diagnostics tab, then confirm domain verification is still in place.

  • On Shopify, confirm the Meta channel's data sharing is set to Maximum, which is the setting that switches on CAPI.

At least 1 of these has usually drifted since the last time anyone looked, typically after a site change nobody connected to tracking.

Your Meta Catalog Needs to Update as Fast as Your Stock Moves

During BFCM, prices change on a schedule while stock sells out faster than at any other point in the year. A catalog that refreshes once a day will spend a meaningful part of Black Friday showing the wrong price or promoting products you can no longer ship.

  • Review rejected items and catalog issues in Commerce Manager, with closest attention on price and availability mismatches between the catalog and the site.

  • Move the feed to an hourly fetch, or a direct platform integration, for the full BFCM window.

  • Populate the sale price field rather than overwriting the regular price, so catalog ads can show the original price alongside the discount.

  • Build product sets that exclude low-stock SKUs, so spend shifts away from products before they sell out.

Advantage+ catalog ads deserve a specific look here. They are very good at finding buyers for the right product, though they also tend to drift toward warm audiences who were already going to purchase. In the numbers, that drift shows up as strong ROAS alongside a high CPMr. Check the new versus returning split on those campaigns in October, so you know what normal looks like before the sale distorts it.

On Google, the Most Common Problem Is Counting the Same Purchase Twice

Google accounts tend to accumulate conversion actions over time, with a Google Ads tag, a GA4 import and sometimes a Shopify app all tracking purchases. When more than 1 of those is set as Primary, Smart Bidding optimizes toward an inflated number. The account looks healthier than it is right up until the budget scales.

The same backend check works here. Pull a week of Google Ads purchases, then compare them against actual orders. If Google reports more purchases than the store took, duplication is the first place to look.

  • Open Goals, then Conversions, and confirm only 1 purchase action is set as Primary. Everything else tracking the same event should sit as Secondary.

  • Check the Enhanced Conversions diagnostics on that action. Anything flagged as Needs attention usually means hashed customer data is missing or malformed.

  • Confirm a transaction ID is passed on every purchase, since that is how Google deduplicates when a tag fires twice.

  • Check that conversion value reflects the real order value, including discounts. BFCM is when hardcoded or pre-discount values distort ROAS the most.

  • For any traffic from the EEA or UK, confirm Consent Mode v2 is implemented correctly, otherwise conversion modeling for those users degrades.

Fixing duplication now will make reported performance look worse for a couple of weeks. That dip is the account becoming accurate, which is far easier to absorb in October than during the sale.

Merchant Center Disapprovals Spike When Prices Change, Which Is Exactly What BFCM Does

Google crawls your product pages to compare them against the feed. When a sale price goes live on the site before the feed catches up, or the other way around, products get disapproved for price mismatch. In a large catalog, that can take out a big share of Shopping and Performance Max inventory in a single afternoon.

Work from the account level down to the items. An account-level issue such as a misrepresentation warning can suspend the entire feed, so check account status before working through item-level problems.

  • Clear item-level disapprovals in the Needs attention view, starting with price mismatch, availability mismatch, missing or invalid GTINs and landing page errors.

  • Turn on automatic item updates for price and availability, which act as a safety net when the feed lags behind the site.

  • Use sale price alongside a sale price effective date, so pricing switches at the right moment without a manual feed push in the middle of the sale.

  • Submit Merchant promotions well ahead of the event, since they need approval while review times stretch as BFCM gets closer.

  • Confirm shipping and returns policies are configured, because missing policy data affects eligibility along with the annotations that show on Shopping ads.

  • Add custom labels for margin, stock level and hero products, so Performance Max listing groups and Shopping campaigns can be split around what you most want to sell.

For anything that needs fixing without touching the primary feed, such as rewriting titles or adding labels, a supplemental feed is the lowest-lift route.

Inventory sync matters as much here as it does on Meta. Check how often the feed is fetched, or whether it runs through the Content API or a platform app. Then confirm out-of-stock products drop out quickly enough that Performance Max stops spending against things you cannot sell.

YouTube Inherits Every Google Tracking Problem, Plus a Frequency Question of Its Own

YouTube runs through Google Ads, so the conversion actions, Enhanced Conversions setup and Merchant Center feed you just audited are the same ones your YouTube campaigns rely on. Demand Gen and video campaigns with a linked product feed can only show approved products, which turns a feed problem on Shopping into a creative problem on YouTube.

YouTube is also the one part of Google where CPMr works the same way it does on Meta, because video campaigns report unique users and average frequency per user. That reporting makes it worth adding to your October baseline.

  • Add the unique users and frequency columns to your YouTube campaigns, then record CPM, frequency and CPMr weekly from now until the sale.

  • Review the engaged-view and view-through conversion windows in your conversion settings. BFCM is when view-based credit inflates the most, which makes it the easiest week to over-read.

  • Set frequency caps where the campaign type allows it, particularly on remarketing, so a small warm audience is not served the same ad dozens of times across the week.

  • Check where impressions are landing, since a growing share of YouTube viewing happens on TV screens where people watch but rarely click. Judge a CTV-heavy campaign on reach and assisted conversions over last click.

  • Upload BFCM video assets early, because ad review slows down for everyone in November.

The same saturation logic applies here as on Meta. If YouTube frequency is already climbing in October while CPM holds steady, the audience was too narrow before the sale even started.

Test the Sale Version of Your Landing Page Before It Goes Live

Speed audits typically happen on the page as it stands today. The page that actually runs on Black Friday usually carries a countdown timer, a discount pop-up, a free shipping bar and 2 or 3 new apps on top. Each of those adds load time on mobile, where the bulk of paid social traffic arrives.

On Meta, the ratio of landing page views to link clicks is the best in-platform proxy. If a meaningful share of clicks never turns into landing page views, people are leaving before the page loads, with that gap widening as site traffic peaks. Record the ratio now so you can spot it slipping during the sale.

On Google, the landing pages report shows a mobile speed score by URL, which helps you find the specific pages dragging down Shopping and Performance Max. For a fuller read, run the sale version of each key page through PageSpeed Insights on mobile and check it against the Core Web Vitals targets.

Metric

Good threshold

Largest Contentful Paint

Under 2.5 seconds

Interaction to Next Paint

Under 200 milliseconds

Cumulative Layout Shift

Under 0.1

If the site has a staging version, build the full BFCM page there in early November and test that. Without staging, switch the sale elements on for an hour on a quiet day to test the live page, which beats finding out on the day.

Spend Limits Sized for Normal Months Will Trip on Black Friday

We pulled the 2025 season from Data Druid, Taikun's proprietary account intelligence platform. Data Druid connects to client ad accounts and turns raw spend, creative and audience data into reports we can query directly. For this benchmark, we compared a normal August 2025 day against the 2025 BFCM season across 12 accounts with enough history to qualify.

The median account hit 3.8x its normal daily spend on its peak day, with the busiest 30 consecutive days running at 1.8x normal pace. Peak-day results for the middle half of accounts landed between 1.8x and 9.9x. One account reached 67x, which is why the median is the more useful number to plan around.

The 30-day figure matters more than it looks, because payment thresholds and credit lines are consumed by spend building up between billing events. A limit that comfortably covers a normal month can run out in the final week of November without any single day looking extreme.

  • Check account spending limits and campaign spend caps on Meta, raising them ahead of time instead of on the day.

  • Review payment thresholds and card limits on both Meta and Google, then confirm a backup payment method is attached.

  • Upload BFCM creative as paused ads at least a week before launch on Meta, Google and YouTube, so it clears review before it needs to go live.

  • Confirm admin access and two-factor authentication on Business Manager and Google Ads, so nobody gets locked out mid-sale.

  • Set up Smart Bidding seasonality adjustments in Google Ads for the core BFCM days. They are designed for short events with an expected jump in conversion rate.

  • Know where data exclusions live in Google Ads, so a tracking break during the sale does not teach Smart Bidding from bad data.

When to Run Each Part of the Audit

The order matters, because tracking fixes need time to settle before you set baselines. Feed fixes also need time to clear review before prices start changing.

  1. Now to October 16. Audit tracking on all 3 platforms, including CAPI, Event Match Quality, Enhanced Conversions, Primary conversion actions and the backend order cross-check.

  2. October 19 to 30. Clear Merchant Center and Meta catalog issues, set sale prices with effective dates, submit Google promotions and add custom labels.

  3. November 2 to 13. Test the sale version of every key landing page, upload BFCM creative as paused ads, raise spend limits and lock in weekly baselines for CPM, frequency, CPMr and net new reach.

  4. November 16. Freeze changes to tracking, site code and account structure, so nothing new breaks in the final 10 days.

  5. November 26 to December 1. Check CPM and frequency daily against the October baseline, using the split to decide whether a rising CPMr needs action or patience.

Pick the first section of this audit and run it this week, because every fix that lands in October removes a problem you would otherwise diagnose at 4x normal spend.

Read more about CPMr and how to calculate it attaikundigital.com/phil-kiel/cpmr-calculation.


Spend benchmark based on 12 anonymized accounts in the Data Druid dataset, comparing August 2025 daily spend against October 15 to December 20, 2025. A further 12 accounts were excluded for insufficient history.


Back to all articles