2025 vs 2026 planning for Holiday season for DTC brands

The BFCM window closes on 2 October
Paid media

The BFCM window closes on 2 October

Meta CPMs across our book are up 35% year over year. Black Friday still pays for itself, but only for accounts that were built before October ended. Here is the math we are running with our own clients.

Every year the same conversation happens in the second week of November. A brand realises Black Friday is close, the offer is not costed, the creative is not tested, and the pixel has been quietly broken since the summer. We take the call, we do what we can, and everyone agrees it would have gone better with more time.

This year the cost of being late is higher than usual, and we can show it rather than assert it. We pulled July and August across our book of beauty and skincare DTC brands and compared it against the same two months of 2025. Three numbers stood out.

up 35%
Meta CPM, July to August, year over year.
up 77%
Cost to acquire a new customer over the same window.
down 24%
Site orders, even as total demand grew.

Demand did not disappear. It moved.

The instinct when site orders fall 24% is to conclude the consumer has pulled back. That is not what the data says. Blended sales across site and Amazon were up 20.7% over the same window. Amazon revenue nearly tripled.

What changed is where the order lands and what it costs to buy. And that growth was purchased, not free: Amazon total advertising cost of sale went from 2.25% to 8.94%. The brands holding up best are the ones treating marketplace and site as one budget with one target, rather than two teams defending two numbers.

Blended across our book of beauty and skincare DTC brands. Windows: 1 July to 31 August, 2025 and 2026. Source: Triple Whale.
Metric20252026Change
Site orders60,76546,107down 24.1%
AOV$66.07$71.28up 7.9%
Blended ROAS7.065.04down 28.6%
Spend as % of sales14.2%19.9%up 5.7 pts
New customer CPA$17.79$31.52up 77.2%
Meta CPM$13.28$17.94up 35.1%
Meta CTR1.15%1.49%up 0.33 pts
Google ROAS7.118.33up 17.2%
Email share of sales34.5%22.1%down 12.5 pts
Blended sales, site plus Amazon$5.27M$6.35Mup 20.7%

Two things in that table deserve more attention than they usually get. First, Meta click through rate went up, from 1.15 to 1.49%. The creative is working harder than it did last year. The cost increase is auction pressure, not a creative failure, and treating it as a creative failure sends teams chasing the wrong fix.

Second, email went from 34.5% of sales to 22.1%. Owned channel is carrying a third less of the load heading into the most email dependent quarter of the year. Every point that email gives up is a point paid media has to buy at a November CPM.

Black Friday still pays. That is the part people get wrong.

The reasonable conclusion from rising costs would be to spend less in the most expensive week of the year. Our data says the opposite, with a condition attached.

Here is peak week last year, 24 November to 1 December, against the October baseline:

Same book. Peak week is 8 days, 24 November to 1 December 2025. Baseline is October 2025. Source: Triple Whale.
MetricOctoberPeak weekChange
Revenue per day$52,344$217,9514.2 times
Meta CPM$14.59$25.68up 76.0%
Meta CPC$1.07$2.01up 87.7%
Meta CPA$25.32$19.30down 23.8%
Meta ROAS3.194.68up 46.7%
Blended ROAS5.007.23up 44.6%
New customer CPA$29.60$23.53down 20.5%
Discounts as % of sales25.6%57.5%up 32 pts

Read the third and fourth rows together. CPM rose 76% and cost per acquisition still fell 24%. That gap is conversion rate. During peak week the same click is worth roughly twice as much, which is why the most expensive media of the year is also the most efficient.

The offset is real, but it is not distributed evenly. It goes to accounts that were already built.

Conversion rate at that level depends on things that take weeks to assemble: warm audiences with enough volume to target, creative that has already been through a test cycle, campaigns that have cleared the learning phase and are not still recalibrating during the highest traffic hours of the year. A brand that turns campaigns on in the third week of November pays the 76% and captures none of the 24%.

The last row is the other warning. Discounts ran at 57.5% of sales during peak week, against 25.6% in October. Whatever a brand thinks its offer will be, the category is going deeper. That is a margin decision, and it has to be made before the offer is announced rather than negotiated in the middle of the week.

Working backwards from 27 November

Black Friday 2026 is Friday 27 November. Cyber Monday is 30 November. Published Meta guidance puts the Q4 preparation window at 8 to 10 weeks, and the component parts explain why: video viewer audiences need 4 to 6 weeks to accumulate usable volume, lookalikes off an email list need 1 to 2, and every new campaign needs to clear learning before CPMs move.

Counting back from 27 November gives two dates that matter:

18 Sept
Last date that gives an account the full 10 week ramp.
2 Oct
The 8 week floor. Tight, but workable with focus.
After that
Buying peak CPMs on a cold pixel to gather data you should already have.

What has to be true before you can trade the week well

This is the diagnostic we run on every call between now and the start of October. 6 things, in the order they have to happen.

  1. The offer is costed against contribution margin. Category discounting hit 57.5% of sales last peak week. Until you know what the promo costs you, you cannot know what you can afford to bid.
  2. Pixel and conversion data are clean. Everything downstream depends on this, and it is the one item that genuinely cannot be fixed in November.
  3. Creative has already been through a test cycle. Winners identified in October, scaled in November. Peak week is not the moment for a first concept round.
  4. Warm audiences are accumulating now. Video viewer pools need 4 to 6 weeks. Starting in November means starting cold, at the worst possible price.
  5. Email and SMS are carrying their share. Email dropped from 34.5 to 22.1% of sales across our book. That gap gets paid for in paid media unless it is closed first.
  6. Inventory and fulfilment are confirmed. Nothing wastes peak spend faster than the hero SKU selling out on day two.

None of this is complicated. It is just sequential, and the sequence takes longer than most brands allow for. If you are reading this in September, you have time for all six. If you come back to it in the first week of November, you have time for one.

Pennock

Want the version of this run on your account?

We work only with beauty, skincare and lifestyle DTC brands. Twenty minutes, your numbers, and an honest answer on whether the window is still open for you.

nikki@pennock.co
Notes and sources
Performance figures are blended across the beauty and skincare DTC brands Pennock manages, pulled from Triple Whale on 2 September 2026. Date windows and day counts are stated with each table. No individual brand figures are disclosed.

External sources: Shopify BFCM 2025 merchant data ($14.6 billion GMV, up 27% year over year, cosmetics the top selling category, $114.70 average order value across 81 million buyers). NRF 2026 retail sales forecast, published 18 March 2026 (4.4% growth to $5.6 trillion). Meta seasonal campaign guidance on Q4 CPM inflation and ramp timing.
Nikki Lindgren