Every year I talk to a few brands in the second week of December who want to know what actually worked over Black Friday, and every year some of them cannot answer, because something in their tracking broke in early November and nobody noticed until the sale was over. The fixes are rarely hard. They just have to happen in October, while you still have normal traffic to test against and time to look at the results.
This is the list I go through. It assumes you already have an attribution tool in place. If you do not, it is late to add one for this year, because a new pixel needs a few weeks of normal data before its numbers mean much.
1. Freeze your UTM conventions now
BFCM is when the most new links get created by the most people in the least time. Email, SMS, influencers, affiliates, and paid social all launch campaigns in the same week, and someone always invents a new utm_source. Write the conventions down, share them with everyone who builds links, including agencies and creators, and spot-check a sample of live links each day of the sale. My structure is in how to structure UTM parameters for Shopify.
2. Place a real test order through checkout
Not a preview. A real order, from a phone, through the same checkout customers will use, with a discount code applied. Then confirm the purchase event fired once in each ad platform and in your attribution tool, with the right revenue. Checkout changes, new apps, and theme updates in the weeks before a sale are the most common reason tracking breaks at the worst moment. If you are on Shopify, read does checkout extensibility break attribution tracking before your test.
3. Check that server-side events are flowing
Meta’s Conversions API and Google’s enhanced conversions matter most during BFCM, because that is when ad platform algorithms are spending your biggest budgets of the year on the data you send them. Look at event match quality in Meta and confirm the server events are deduplicating against browser events, not doubling them. A doubled purchase event during a sale inflates every platform number you look at. Background is in do you need server-side tracking for accurate attribution.
4. Map every discount code to a channel
Sitewide codes spread everywhere. Coupon sites and browser extensions pick them up within hours, and suddenly orders that came from an extension look like they came from your email. Keep a sheet of every code, who it was issued to, and which channel it belongs to, and use unique codes for influencers and affiliates. Then you can break revenue out by code after the sale and see what each one really brought in.
5. Update your post-purchase survey
If you run a how-did-you-hear-about-us survey, check that the options cover what you will actually run this year, including any new channels, TV or podcasts, and creators. Gift buyers often found you differently from regular customers, so this is a good time to add a short question about whether the order is a gift. Survey answers are the best read on what drove new customers during a period when pixel data gets noisy.
6. Save last year’s numbers as a baseline
Export last year’s BFCM week by day: revenue, orders, new customer share, spend by channel, MER, and attribution by channel. Write down what else was different then, such as a different offer, a different email list size, or a channel you have since stopped. Without that, the post-sale review becomes a debate about whether this year was good, instead of a look at what changed.
7. Pick the one number you will read every day
During the sale, platform ROAS will look fantastic, because everyone is buying and every platform is claiming those orders. Decide before November which number you will use to make live budget calls. For most brands, I would use MER plus new customer revenue from your attribution tool, and treat platform ROAS as a within-platform signal only. I covered why in how to read ROAS during a sale or promotion. And do not change attribution models or windows in the middle of the sale. Every number afterward becomes impossible to compare.
Where your attribution tool helps
An independent attribution tool is what lets you see the sale without each platform’s overlapping claims. ThoughtMetric, for instance, can break the same orders out by channel, campaign, discount code, and new versus returning customer in one view, which covers most of what this list sets up. (Disclosure: ThoughtMetric sponsors this site.) Whichever tool you use, test it now, not on the Friday.
The October checklist
- UTM conventions written down and shared with every link builder.
- A real test order placed on mobile with a code, and verified in every platform.
- Server-side events confirmed and deduplicated.
- Every discount code mapped to a channel, with unique codes for creators.
- Post-purchase survey options updated for this year’s channels.
- Last year’s BFCM saved by day, with notes on what was different.
- One daily decision number agreed, and no model changes until December.
Leave a comment