The BFCM 2026 paid media playbook for ecommerce.

Black Friday 2026 falls on November 27. Thanksgiving is the Thursday before, November 26, and Cyber Monday is November 30. As I write this at the end of September, there are about eight weeks to go.

That sounds like plenty. It is not, because the work that decides the weekend is not the work that happens on the weekend. By the Thursday, the offers are set, the creative is what it is, the feed is approved or it is not, and the tracking either works or it does not. What is left is pacing and nerve. This guide is the plan in the order it needs doing.

The calendar

Weeks 1 and 2: set the economics before the budget

Every decision on the weekend depends on one question: what return does this offer need to pay for itself? The answer changes during a sale, because the discount comes straight out of margin. A product that breaks even at 2.5x at full price can need 4x or more at 25% off.

Here is the arithmetic for one illustrative product, not a client result. It sells for $100 with $55 of contribution margin after product cost, shipping, and payment fees. Every dollar of discount comes out of that $55, so the margin shrinks faster than the price does.

Read the last column before the ROAS column. At full price, paid media can spend $55 to win the order. At 40% off, it can spend $15. The platforms will not know that unless you tell them, which is why the targets and budgets later in this plan depend on this table. Run your own numbers in the breakeven ROAS calculator.

Do this per offer, not per store. Then decide what the event is for. There are two honest answers and they lead to different plans.

• →Profit on the weekend. Every sale should clear its breakeven. Bid to margin, limit budget on the deepest discounts, and accept a smaller top line.
• →Customers for next year. Buy new customers near or below breakeven and earn it back on repeat purchase. This works only if you know your repeat behavior by cohort, and it should come with a ceiling on new-customer CAC.

Most brands do some of each. Write down which one each offer is for, because on the Saturday someone will ask why a campaign is spending at 2x ROAS, and the answer should already exist.

Which answer fits depends on what is holding the business back the rest of the year. A brand short of new customers can justify buying them at breakeven in November. A brand whose constraint is margin should not use its biggest week to make that constraint worse. The ecommerce growth ladder sets out the constraint at each revenue band, and it is worth reading before you choose.

Use last year as your benchmark, carefully

Your own history is the best forecast you have, and it needs adjusting before you use it. Pull last year’s spend, orders, conversion rate, and average order value by day for the four weeks around the event, and by hour for the peak days. Then write down what was different: the offer, the prices, the stock, the channels you ran, and any tracking changes since. Last year’s Black Friday conversion rate is a good starting point for this year’s seasonality adjustment only if the offer and the site are comparable.

Line the calendar up by event day, not by date. Thanksgiving falls on November 26 this year, so compare this Black Friday with last year’s Black Friday, not with last year’s November 27. The same applies to Cyber Monday and the days around them.

Weeks 2 to 4: audit tracking, then freeze it

The most expensive thing that can happen on Black Friday is not a weak offer. It is a broken purchase event during the one week bidding matters most. Smart Bidding and Meta’s delivery both learn from the conversions they receive. If purchases stop reporting, the platforms conclude your ads stopped working and act on it, and you will not notice until someone compares the ad account with the store.

1. Compare purchases reported by each platform with store orders for the last four weeks. The platforms will claim more than the store in total. What you are looking for is a stable ratio, not a match.
2. Place test orders and check the value, currency, and order id that each platform receives.
3. If you send purchases both from the browser and from your server, confirm they share an event id so the platform counts each order once. Meta documents how its deduplication works.
4. Confirm consent settings behave as intended in the regions you sell to.
5. Then freeze. No new tags, no consent banner changes, no checkout app swaps, no theme releases that touch the cart, until December.

Weeks 3 to 5: creative and the feed

Creative

Offer creative needs to be live before the peak, not launched on it. New ads need time and spend to find their audience, and the last week of November is the most expensive week of the year to pay for that learning. Launch offer variants in late October and early November at modest budgets, find the winners, and scale those.

• →Lead with the offer and the product, clearly. Clever concepts lose to clarity in a crowded week.
• →Prepare versions for each phase: early access, the weekend, Cyber Monday, and a last-chance version.
• →Make enough variety to rotate. Frequency climbs fast when every brand is spending at once, and creative fatigue shows up in days, not weeks.
• →Have the post-weekend creative ready too, so December does not start with nothing new.

The Merchant Center feed

For Shopping and Performance Max, the feed is the ad. Three things to settle in October:

• →Sale prices with effective dates, so the discount shows in the ad at the right moment and not before.
• →Promotions loaded in Merchant Center, following Google’s promotions specification, with time left for review. A promotion that is still pending on the Friday is a promotion that does not exist.
• →Disapprovals cleared. Check the products that matter most for the event, not just the account-level count.

If you are carrying extra stock for the event, make sure the hero products are clearly separable in your campaign structure, so you can push them without pushing everything else. The Shopping feed optimization guide covers how to build that layer.

November 1 to 22: warm demand before you harvest it

The brands that do well on Black Friday usually did well in the three weeks before it. The weekend converts people who already know you. If the only prospecting happens on the Friday, you are competing for cold attention at peak prices.

• →Keep prospecting running at a planned level through early November, even as costs rise.
• →Grow the email and SMS list with early-access offers. People on your own list are the cheapest customers you will reach that weekend.
• →Watch new-customer counts, not only ROAS. Warming spend looks inefficient on the platform and shows up on the weekend.

The peak: pacing budget and bidding

Budgets

Plan the daily budget for each day of the peak in advance, with a written ceiling. Google Ads can spend up to twice the average daily budget on a single day, while never charging more than 30.4 times the daily budget across the month, so a budget that looks capped can move faster than expected on a heavy day. Check spend at set times through each day rather than waiting for the evening.

Raise budgets in steps you decided beforehand, not in reaction to a good hour. A campaign that is limited by budget while holding its return is the one to feed. A campaign that is spending more and returning less has found the edge of its audience.

Three principles cover most of it. Shape the month as a curve rather than a switch, so spend rises through November as warm audiences grow. Put every increase behind the offer with the most room in the table above, not the deepest discount. And write down, before November starts, the condition that stops extra spend, because the moment to decide it is not the moment it arrives. The peak season budget planner lays out November spend by week and marks where the next dollar stops paying for itself.

Seasonality adjustments

Smart Bidding learns from recent conversion rates. On the peak days, conversion rates jump, and the algorithm takes time to catch up. Google’s seasonality adjustments exist for exactly this: you tell the bidding system that conversion rates will be higher for a set window, and it bids accordingly.

• →Google describes them as ideal for short events of one to seven days and warns they may not work as well beyond 14 days. Use them for the peak days, not for the whole of November.
• →They apply to Search, Shopping, and Display campaigns on Target ROAS or Target CPA, and to Performance Max on any bid strategy, per Google’s documentation.
• →Set the expected change from your own history. If last year’s Black Friday converted at about twice a normal day, that is your starting point, not a generic figure.
• →Do not stack them with large target changes at the same time. Make one move and watch it.
• →Remove them when the event ends. If you forget, the system keeps bidding for a conversion rate that is gone.

Targets

If you bid to a ROAS target, the target should reflect the offer’s breakeven, which you already worked out. A target set for full-price margin will overspend on deep discounts. A target set too high will stop the campaign spending on the most valuable days of the year. The tROAS versus tCPA guide covers when each works.

What not to do on the weekend

• →Do not restructure campaigns. Combining or splitting campaigns mid-peak resets what the system has learned.
• →Do not launch large amounts of untested creative. Rotate what you prepared.
• →Do not judge an hour on its own. Look at the day against the same day last year and against the plan.
• →Do not assume a drop is demand. Check store orders first. If the store is taking orders and the platform has stopped counting them, the tag broke.

Meta during the peak

Meta’s auction gets crowded in the same week as Google’s, and its delivery system reacts to sudden changes in a similar way. The same principle applies: decide the moves in advance and make fewer, larger, planned ones.

• →Raise budgets in planned steps rather than many small reactive edits. Meta says ad sets usually leave the learning phase after about 50 results in the week after their last significant edit, and that frequent budget changes can send them back into it at the worst possible moment.
• →Keep the purchase event deduplicated. If you send purchases from both the browser and the server through the Conversions API, each pair must share an event ID, or Meta will count the peak twice and your reported ROAS will look better than the store.
• →Decide what to do with recent buyers. Prospecting campaigns that keep reaching people who bought last week are spending acquisition budget on customers you already have.
• →Watch frequency by day. When every brand is spending at once, the same person sees the same ad quickly, and the rotation you prepared in October is what keeps performance from sliding.

If something breaks during the weekend

Something usually does. A tag stops firing after a theme update, a promotion is disapproved, a payment provider slows down. Write the order of checks before the weekend so nobody improvises it at midnight.

1. Check store orders against the same hour on a normal day. If orders are fine, the problem is measurement, and budgets should not move.
2. If orders are down, test the checkout yourself on mobile and desktop before looking at campaigns.
3. Check Merchant Center and the ad platforms for disapprovals on the hero products.
4. Check spend against the day’s plan. A campaign racing ahead of plan with falling orders is the one to slow down first.
5. Write down what happened and when. The January review will need it.

Watch the store, not just the platforms

During the peak, read the ad platforms and the store side by side. Platform-reported revenue will run well ahead of what the store books, because several platforms claim the same orders. That is normal. What you are watching for is the ratio changing suddenly, which almost always means a measurement problem. Why Meta, Google, GA4, and Shopify report different revenue explains how to read the gap.

Two numbers to check every few hours: spend against the day’s plan, and store orders against the same hour on a normal day and last year. If both are on track, leave it alone. If spend is racing and orders are not, find out why before adding budget.

Subscription and replenishment brands

If you sell a subscription or a product people reorder, the event works differently. The discount is not only on one order. It can set the price a customer expects for months.

• →Decide whether the offer applies to the first order only or to the subscription. A discount that carries forward changes the lifetime value of every subscriber acquired that weekend.
• →Model breakeven on the first two or three billing cycles, not the first order, using your own retention history.
• →Expect event subscribers to behave differently from full-price ones. Measure their month-two and month-three retention as a separate cohort before deciding the event worked.
• →Keep existing subscribers out of acquisition campaigns where you can, so the budget buys new subscribers rather than discounts for people already paying.

December: hold, extend, and wait for returns

The weekend is not the end of the event. Offers often extend through the following week, and there is a second window for holiday gifting in early December with delivery cut-offs to plan around. Keep the best creative running, bring retention into the plan, and bring seasonality adjustments and targets back to normal.

Then wait. Returns from the event arrive through December and into January, especially in apparel. Judging the weekend on the Monday after is judging it on gross sales, before the business has found out how many of those sales it keeps.

January: judge it on what the store kept

Platform ROAS for BFCM is always flattering. The platforms credit themselves for purchases from people who were waiting for the sale anyway, and several platforms claim the same orders. The review should answer different questions.

The last question takes months to answer, and it is the one that decides next year’s plan. Customers acquired on a deep discount sometimes buy again at full price and sometimes never return. If you know which, you know how much to spend acquiring them next November. The ecommerce unit economics guide covers the cohort view.

What most peak advice gets wrong

The most repeated piece of BFCM advice is to spend hardest on the days conversion rates are highest. It sounds like arithmetic. It leaves out the other half of the sum.

A higher conversion rate means more orders per click. It says nothing about what each order is worth. On the peak days the discount is usually at its deepest, so the contribution per order is at its lowest, and every other brand in the category is bidding for the same people at the same time. Look at the table again: the same product that could carry $55 of ad cost in October can carry $15 at 40% off. A conversion rate that doubles does not rescue an order whose margin fell by more than half.

The better question for each day is what the next dollar earns after the discount. Often the answer is to put paid budget behind the shallow offers and the full-price bestsellers, and to let the email and SMS list carry the deepest discounts, because reaching your own list costs almost nothing. The deep offer still sells. Paid media just stops buying it at a loss.

When this plan is wrong

This plan assumes a brand that discounts, has at least one prior peak to learn from, and has the stock to meet demand. When any of those is false, parts of it change.

• →You do not discount. A full-price brand can treat the week as expensive attention and hold spend flat, or even lower it, while competitors bid up the auction. The event is not compulsory.
• →This is your first peak. Without your own history there is no honest number for a seasonality adjustment. Skip it, or use a small one and watch hourly, rather than borrowing a figure from somewhere else.
• →Your hero products will sell out. Extra spend on a product that will be out of stock by Saturday buys disappointed customers. Plan budgets from stock, not from last year’s demand.
• →You sell a subscription. The first order is not the unit of value, and the section above on subscription brands matters more than the pacing sections.

What to do this week

It is the end of September. Four things are worth finishing before the week is out, because everything else in the calendar depends on them.

1. Pull last year’s spend, orders, conversion rate, and average order value by event day, lined up on Black Friday rather than on the date.
2. Build the breakeven table above for every offer you are considering, using your real contribution margin.
3. Name the tracking freeze date and the person who owns it, and tell the developers now.
4. Put a date in the diary for loading Merchant Center promotions, with room for review before the peak week.

The short checklist

• →Breakeven ROAS worked out per offer at the discounted margin.
• →A written purpose for each offer: profit now, or customers for next year.
• →Tracking audited against the store, deduplicated, and frozen until December.
• →Offer creative live and tested before the peak week.
• →Sale prices and promotions loaded in Merchant Center with time for review.
• →Prospecting and list growth running through early November.
• →Daily budget plan and ceilings written down for each peak day.
• →Seasonality adjustments scheduled for the peak days only, and removed after.
• →Hourly checks of spend and store orders during the peak.
• →A January review on net revenue, margin, and new customers.