The DTC Playbook Search ⌘K
The DTC Playbook
Contents Glossary Benchmarks CalculatorsBeta Tools Health Check Dashboard

The DTC Playbook is a collection of learnings, frameworks and stories from my journey co-founding Quad Lock, and scaling to $200M in revenue and a $500M exit. - Rob Ward

Silent explainer video. It presents The DTC Playbook's tagline - Build a brand. Scale it. Sell it? - and introduces the playbook: a free, single source of truth for direct-to-consumer founders, written by Rob Ward, who bootstrapped Quad Lock to $50M+ in revenue before a $500M exit. It shows the Health Check that diagnoses what to fix in your business, and the sections, checklists and tools that show you how to fix it.

Home / Growth Channels / Peak Season & Promotions
S31 · Growth Channels

Peak Season & Promotions

BFCM, Sales Events, Promotional Calendar

Section 31 / Growth Channels / by Rob Ward
Free to read or create a free account to score your brand and see exactly what to fix first. No Subscription. No Credit Card. No Upsell. 100% Free.
TL;DR
  • Peak is won months before it starts: lock your calendar, margin floor, inventory and creative early.
  • Engineer offers like bundles, gift-with-purchase and tiered thresholds under any sitewide headline, never a flat percent-off alone.
  • Build the list and warm audiences in October; email and SMS drove roughly 42% of BFCM 2025 revenue.
  • Treat BFCM as customer acquisition: plan the second-purchase journey before the sale and track that cohort separately.
On this page
Principle
Win Peak Before It Starts

The biggest revenue weeks of the year are won in the months before them. Lock your calendar, your margin floor, your inventory and your creative early, and peak season becomes a system you execute, not a scramble you survive.

Black Friday and Cyber Monday (BFCM) is the biggest commercial event of the year for most DTC brands. Shopify alone did $14.6B in GMV across the 2025 weekend, up 27% year on year, and peaked at $5.1M of sales per minute. That is the prize. It is also the day your margin gets shredded, your site gets stress-tested, your 3PL gets buried, and your support inbox catches fire. All at once.

This section is about running peak season like an operator, not a gambler. It pulls together the promo calendar, the offer design, the phased timeline, the paid and owned-channel orchestration, and the ops readiness that has to sit underneath all of it. Email and SMS get their own deep treatment in Section 12: Email & SMS; the margin maths lives in Section 26: Finance & Unit Economics. This section is the conductor that brings them together for the one period of the year where getting it wrong is most expensive.

Read next
Read This Alongside

This section ties directly into Section 12: Email & SMS (your highest-ROI peak channel), Section 21: Retention & Loyalty (what you do with the cohort afterwards), and Section 26: Finance & Unit Economics (the margin floor that should govern every offer).

The Annual Promotional Calendar

Most brands discount reactively. Sales dip, someone panics, a 20% off code goes out. Do that all year and you train your customers to wait, cannibalise your full-price sales, and make BFCM feel like nothing special, because you've been running BFCM-depth offers every six weeks.

The fix is a 12-month calendar built in advance, where every promotion has a defined slot and a defined job. Lay it out once, agree it across marketing, finance and ops, and stop improvising.

For each promo on the calendar, define:

FieldWhat it means
MomentBFCM, Boxing Day/EOFY, a product launch, a regional holiday, an evergreen sale window
Offer typeBundle, gift-with-purchase, tiered threshold, free-shipping floor, sitewide percent-off (rare)
DepthThe actual discount or value given away
Margin floorThe contribution margin you will not go below, per SKU
Channel mixWhich of email, SMS, paid, organic carry it
Inventory tied to itThe specific stock committed, so you don't promo something you can't ship

The single most important entry on the calendar is the one that isn't a promo: your no-promo baseline. Decide upfront how many genuine sitewide events you'll run in a year. For a brand protecting equity, that's one to three - Quad Lock ran one. Everything else is full price, or an engineered offer that doesn't read as a blanket discount. That scarcity is what keeps BFCM special.

Warning
Calibrate the calendar to your category before you copy the once-a-year discipline.

If you're an apparel brand, markdown isn't a promo decision, it's part of the operating model: seasonal stock has a shelf life, so your calendar needs planned end-of-season clearance windows per drop (kept visually and structurally separate from your brand-moment sales) or ageing inventory will force the reactive discounting this section warns about. If you're a consumable brand, demand is steady-state and a deep sitewide sale mostly lets existing subscribers stock up at a discount - engineer peak offers around acquisition instead: first-subscription incentives, bundles and gift formats, with your subscriber base fenced off from offers that merely cannibalise the reorders they were already going to make.

Warning
Reactive Discounting Is a Slow Leak

A brand with no calendar discounts ad-hoc all year, cannibalises full-price revenue, erodes margin, and arrives at BFCM with a customer base that has been trained to only buy on sale. The calendar isn't bureaucracy. It's the thing that protects your pricing power for the one weekend that actually matters.

The Global Sales-Event Calendar

If you sell into more than one region, a single US-centric calendar leaves money on the table and annoys half your customers. The peaks don't line up, the dates move by country, and some of the biggest events barely register in the US at all. Map the whole year by event, then localise it per market.

EventWhenWhere it matters mostNotes
Valentine's Day14 FebUS, UK, EU, AUGifting categories only
Mother's DayUK: mid-Mar (Mothering Sunday) · US/AU: 2nd Sun MayUK, US, AUThe date is different by country - don't send the AU offer to the UK in May
EOFY (End of Financial Year)30 JunAustraliaA major AU sale event in its own right; tax-driven, locally often rivals BFCM
Amazon Prime Day~Jul (+ an Oct "Big Deal Days")US, UK, EU, AU, JPAmazon-led, but a halo moment - run a parallel DTC sale off the demand
Father's DayUS/UK: 3rd Sun Jun · AU: 1st Sun SepUS, UK, AUDate differs by country again
Singles' Day (11.11)11 NovChina / Asia (growing globally)The single biggest shopping day on earth
Click Frenzymid-NovAustraliaAU-specific online sale event, just before BFCM
Black FridayDay after US Thanksgiving (late Nov)GlobalThe anchor of the whole season
Cyber MondayMon after BFGlobal
Green Monday2nd Mon DecUSThe last big push before shipping cut-offs
Boxing Day26 DecUK, AU, NZ, Canada (NOT the US)The post-Christmas equivalent of BFCM; huge in AU/UK
Chinese New Yearlate Jan/FebChina / AsiaA demand event AND a supply event - factories shut for weeks, so plan inventory around it

Then localise down to each market - offers, inventory and shipping cut-offs mapped to that region's real peaks, not a US default:

MarketThe peaks that actually matter
USPrime Day, BFCM + Cyber Week, Green Monday, the Christmas run-in
UK / EuropeBFCM, Boxing Day, regional Mother's Day (March)
AustraliaClick Frenzy, BFCM, Boxing Day, and EOFY (June) - four big moments, not one
China / AsiaSingles' Day (11.11), 6.18, Double 12, Chinese New Year

And if you sell through wholesale or retail partners as well as DTC, your promo depth has to be coordinated with theirs, or you'll undercut your own stockists and start a channel-conflict fight you don't want.

Rob's take

Quad Lock only ever discounted once a year, around Black Friday and Cyber Monday. We'd done it since before we were in retail and distribution, so by the time we had wholesale and retail partners it was simply business as usual - one short, sharp window a year to tell the Quad Lock story at scale, then straight back to full price. Partners don't always love it; a stockist can feel undercut when your own DTC site runs the same sale. What keeps the peace is communicating it early and being straight about it: this is the one promo we run, it's a few days and not a month, and then it's done. It wasn't always received well, and it's a balancing act - but you're far better off setting the expectation up front than springing it on them.

BFCM Offer & Margin Design

The default move is a deep sitewide percent-off. On its own, with nothing engineered underneath it, it's the bluntest tool available: a flat 30% off the whole store hits your best-margin and worst-margin SKUs identically, gives away margin to people who were going to buy anyway, and teaches everyone to wait for the next one. The 2025 winners protected margin while still feeling generous - the read from Klaviyo's 2025 season was that loyalty and AI outperformed discounting outright, and around 40% of consumers said they actually expected fewer discounts than before, so deep percent-off for its own sake isn't even what the market is demanding. The fix isn't necessarily to scrap the sitewide headline - it's to make sure real offer engineering sits underneath it.

The shift is from blanket discounts to engineered offers:

1
Bundles / ritual sets
Package complementary products at a set price. Raises AOV and moves more units while protecting per-unit margin.
2
Gift-with-purchase (GWP)
Add a low-cost, high-perceived-value item over a spend threshold. Feels generous, costs you cost-of-goods not margin-off-revenue.
3
Tiered spend thresholds
Spend $X get Y% or a gift; spend more, get more. Pulls basket size up rather than just cutting price.
4
Free-shipping floors
A shipping threshold set just above your current AOV nudges people to add one more item.
5
VIP / early-access perks
Exclusivity for loyalty members and subscribers instead of a universal doorbuster.

Why bundles beat blanket percent-off: the same headline discount number can produce a completely different basket. A 30% sitewide code just shaves 30% off whatever the customer was already going to buy; a 30% bundle gets them to buy more and anchors the value on the set, not the markdown. The gap Triple Whale measured between the two is stark:

~$31.50 -> ~$89
Bundle AOV lift from a 30%-off engineered bundle (Triple Whale). The same "30% off" as a sitewide code, but it grows the basket instead of just discounting it.
Rob's take

For more than a decade, Quad Lock ran the same headline every year: 30% off sitewide, Black Friday and Cyber Monday, globally - and that was the only real promotion we ran all year. That headline covered our hero SKUs too - the sanctioned once-a-year exception to the protect-your-hero rule in Section 26: Finance & Unit Economics; the scarcity is what made it safe. The flat sitewide number was deliberate. Our range had so many possible bundle combinations that we could never have hand-built a bespoke offer for each one, so a clean 30% covered the whole catalogue at once. But the headline was never where the work happened. Underneath it we engineered upsells and built bundles, and on the bundles we'd often give a bigger dollar discount than the 30% - then show the customer exactly how many extra dollars they were saving by taking the bundle. That dollar framing, not the percentage, is what pulled AOV up.

The reason we could go deeper than 30% on a bundle without wrecking margin is that a bundle ships as a single parcel. More units, the same pick-pack-and-post cost - so the per-order economics actually improve as the basket grows, even as the headline discount on it deepens. The percentage gets the click; the bundle protects the margin.

There's a product-planning corollary to all of this. Bundles only pull existing customers back if there's something genuinely new worth coming back for. The bundles that worked hardest were built around products released that year that applied to a broad slice of the customer base - which meant peak planning really started months earlier, in the product roadmap. If you want existing customers to re-buy at peak, plan at least one broadly-relevant new product to be in-market before the season, not after it. (See Section 5: Product and Section 6: Merchandising & Assortment for the range planning this leans on.)

Set The Margin Floor First

Before you commit to any offer, set a hard contribution-margin floor per SKU, and model the blended margin after returns and shipping. Not the headline discount. The number you actually keep once the parcel has shipped, some of it has come back, and the discount has been applied. Peak-period returns run hot, especially on gifting, so a margin that looks fine on the order line can go underwater once Q4 returns land.

Engineered, margin-protected offer
  • Bundles, GWP, tiered thresholds
  • Hard margin floor set per SKU
  • Blended margin modelled after returns and shipping
  • Lifts AOV and units, feels generous
A flat percent-off as the whole offer
  • A sitewide % with nothing engineered underneath
  • No per-SKU margin floor
  • Margin given away to people who'd have paid full price
  • Trains customers to only buy on sale

The Phased Calendar: Tease, Early Access, Peak, Extension, Post-Peak, Retention

Black Friday stopped being a weekend a while ago. It's now a multi-week marathon that runs from early October through Cyber Week and beyond. Around 32% of consumers start their holiday shopping before November even begins, and in 2024 the two weeks running into BFCM grew roughly 45% year on year. If you concentrate everything on the Friday-to-Monday window, you miss a third of the demand and pay the most expensive ad prices of the year to do it.

Run the season in distinct phases, each with its own job:

1
Phase 1 - Warm-up / teaser (October)
Build the list and warm the audience. Run list-building popups, capture VIP sign-ups, tease that something's coming. This is where you convert cold traffic to owned before prices spike.
2
Phase 2 - Early access (early November)
Open the sale early to VIP, email and SMS subscribers. Rewards loyalty, pulls demand forward, and takes load off the main launch.
3
Phase 3 - BFCM peak (Black Friday to Cyber Monday)
The main event. Daily sends, full paid spend, the headline offers live.
4
Phase 4 - Extension / Cyber Week (post-Cyber)
"Last chance" and extension offers for the people who didn't move during peak. Cheaper incremental sales off an already-warmed list.
5
Phase 5 - Post-peak push (Green Monday + shipping cut-offs)
Green Monday and the holiday shipping cut-off push. Urgency switches from "best price" to "order now or it won't arrive in time."
6
Phase 6 - Retention (December onward)
The second-purchase journey for the cohort you just acquired. See the cohort section below.

The whole point of phasing is that you do the hard, cheap work early, building the list and warming audiences in October, so that when peak hits, your spend is converting warm traffic instead of buying cold clicks at peak prices. Spread the offers, spread the spend, spread the load.

Rob's take

At Quad Lock we planned BFCM by pulling last year's data and mapping out exactly how sales came in day by day, alongside what we did on each day. Emails, social posts, performance campaigns, all of it. We put it on Miro as a graph with every initiative pinned to its date. Email, social, performance, each team contributed. Then we'd build this year's plan on top of it. If we wanted X% above last year, we could see exactly where the extra had to come from. Everyone understood the full picture and the timing of every piece before the sale kicked off. (How the email and SMS calendar specifically gets planned against last year's send-by-send data is in Section 12: Email & SMS.)

The other thing that map gave us was the ability to smooth the load, not just plan the sales - to spread demand so no single moment overwhelmed the site, the checkout or the 3PL. The mechanics of that, and the Black Friday that taught us to do it, are under fulfilment and site readiness below.

Peak season is the most expensive auction of the year and the fastest place to fatigue your creative. Two things kill under-prepared brands: they run stale ads into the ground, and they spend at elevated CPMs against cold audiences they should have warmed weeks earlier.

The fixes:

01
Pump out creative
Increase creative output 3-4x normal volume to beat fatigue, leaning on your proven evergreen winners with a simple BFCM overlay rather than building everything from scratch.
02
Warm audiences early
Launch ad sets 3-5 days before the sale, even on engagement objectives, so you're retargeting a primed audience at peak instead of paying to acquire cold.
03
Segment and retarget
Use RFM segmentation, scale retargeting against your warmed list, and pre-load Advantage+ / PMax with assets before the rush.
04
Pace and kill
Have daily budget-pacing and kill rules ready in advance, so a stale ad set or a blown CPA gets cut fast instead of bleeding budget for a day.

Budget for elevated CPMs. Triple Whale's aggregate paid-social CPM hit around $22.26 at BFCM 2025, up about 7.6% year on year across their 50,000+ DTC brands. (That's an aggregate benchmark, so treat it as a directional planning figure rather than gospel for your specific account, your category and CPM history will differ.) The point stands regardless of the exact number: prices go up at peak, and the cheapest incremental conversions are the ones you've already warmed.

Warning
Don't Buy Cold At Peak

The classic mistake: do no list-building or audience-warming in October, then launch on Black Friday and pay the year's highest CPMs to cold traffic. By the time you're warming an audience on Black Friday, it's too late. The warm-up is an October job, not a peak-weekend job.

Rob's take

The reflex is to assume peak just means expensive - CPMs climb as everyone piles into the auction, so your CAC climbs with them. It doesn't have to. When intent is this high and the offer is genuinely compelling, conversion lifts enough that CAC can hold flat or even come down, expensive CPMs and all. The lever is timing. We'd start warming the audience before the sale, front-running it with story and hype while CPMs were still cheap, so that when peak hit we were harvesting an audience we'd already primed rather than buying cold clicks at the year's highest prices. The other thing working in our favour: because our Black Friday offer was never any deeper than it always was, there was nothing to wait for. The customers who'd otherwise sit on the sidelines watching for a better discount knew this was as good as it got, so they moved.

For the full Meta and Google playbooks that this peak overlay sits on top of, see Section 14: Meta Ads and Section 15: Google Ads. For channels beyond the big two, see Section 16: Other Channels.

Email & SMS At Peak

Owned channels are the highest-ROI, CPM-proof lever you have at peak. While paid prices spike, the marginal cost of an email or an SMS to a list you already own is effectively zero. That's why email and SMS drove roughly 42% of total BFCM revenue in 2025 (around 43% on the peak days themselves). These are the cheapest incremental sales available to you all weekend.

~42% of BFCM revenue
Email and SMS combined drove around 42% of total BFCM 2025 revenue (43% on peak days). Klaviyo alone delivered 22.7B messages and $3.8B of attributed revenue across the weekend.

The full flow and campaign mechanics live in Section 12: Email & SMS. For peak specifically, the job is twofold: plan the broadcast calendar, and run email and SMS together.

The broadcast calendar mirrors the phased timeline: teaser, early-access, launch, daily peak sends, last-chance, and extension. On top of that, refresh your core flows for the season, welcome, abandoned cart, browse abandon and post-purchase, so the automations that run 24/7 carry the peak offer too.

Run email and SMS as one orchestrated programme, not two silos. Cross-channel shoppers (people you reach on both) placed around 11% more orders and viewed around 71% more products than single-channel shoppers in 2025. SMS scaled hard as a peak channel: send volume was up 34% and text revenue up 25% year on year. But it's the most intimate and most easily abused channel, so segment by RFM, suppress recent buyers, and stay strict on SMS consent and quiet-hours rules.

+11% orders / +71% products viewed
Cross-channel (email + SMS) shoppers vs single-channel at BFCM 2025 (Klaviyo). Orchestrating the two together, not running them separately, is what produces the lift.

Inventory & Demand Forecasting

A stockout on Black Friday forfeits your best-margin sales day of the year. Overstock leaves you discounting dead inventory in January. Both are expensive, and both are avoidable with a proper forecast.

Forecast at SKU level, not brand level. A brand-level number hides the fact that your hero SKUs will sell out while your long tail sits there. Build three scenarios from last year's peak, your year-to-date growth rate, and the specific promo plan:

1
Base
Last year's peak adjusted for YTD growth. The number you'd bet on.
2
Optimistic
Stronger growth plus the lift you expect from this year's offers and audience-building.
3
Stretch
The upside case. What you'd need on the shelf if everything overperforms.

Then get the operational details right:

  • Place POs against early cut-offs. Factories and inbound freight back up well before peak; warehouses are overloaded by mid-October. For October/November delivery, planning typically starts mid-year. Miss the cut-off and the stock doesn't arrive in time, full stop.
  • Hold safety stock on hero SKUs. Your best sellers are where stockouts hurt most. Over-index safety stock there.
  • Pre-agree substitution and oversell rules. Decide in advance what happens when a SKU runs out, before you're making the call live at 2am on Black Friday.
  • Write the backorder messaging now. If you're going to oversell, the customer-facing comms need to be ready and honest.
  • Model elevated peak return rates into your net-inventory plan. Gross units sold isn't net units kept.

The forecasting, PO-timing and inbound-logistics mechanics this leans on are covered in full in Section 7: Supply Chain & Ops - peak just compresses the timeline and raises the stakes.

Rob's take

We'd be planning and placing Black Friday stock more than six months out - it has to be made and shipped in, and demand doesn't stop at Cyber Monday: you run straight into Christmas, and then Chinese New Year shuts the factories, so you need enough on hand to carry all the way through. The live part was the marketing-to-ops loop during the sale itself. If ops flagged that a hero SKU was about to run thin, we didn't keep pouring spend into it and sell out - we'd move the budget onto another product that was converting nearly as well and had the stock behind it. Better to redirect demand than to advertise your way into a stockout on your biggest weekend.

Warning
Forecast At SKU Level Or Don't Bother

Forecasting at brand level is how you stock out of the hero SKU that drives the weekend while sitting on a pallet of slow movers you'll be discounting in January. The forecast has to go down to the individual SKU, with safety stock weighted to your proven winners.

Fulfilment, 3PL & Shipping Cut-Offs

Peak can surge daily order volume to roughly 3-5x normal, with spikes up to 10x on the biggest days. Unprepared ops break under that. And when ops break, you don't just lose the order, you generate refunds, chargebacks, support load and one-star reviews during your highest-traffic window of the year. The damage compounds.

Lock the operational pieces early:

ActionWhy / when
Reserve 3PL capacityLock it early (often by September). Capacity gets allocated to whoever booked first.
Agree seasonal labourWarehouses need ramped headcount; agree it before the labour market tightens.
Pre-negotiate carrier surchargesPeak carrier fees get set in advance; negotiate before they're imposed on you.
Use multi-node fulfilmentShipping from multiple locations cuts transit times and de-risks any single node.
Publish shipping cut-off dates everywhereOn the PDP, in the cart, in email, in ads. They create urgency AND set honest expectations.
Build a surge contingency planWhat happens if volume blows past the stretch case. Agree it before, not during.

The shipping cut-off dates do double duty. They protect you from over-promising delivery you can't make, and they're one of the most effective urgency levers in the back half of the season, when "order by Tuesday to get it for Christmas" converts harder than another percent off.

Rob's take

Even on Shopify Plus, our site went down during one Black Friday. We'd just sent over a million emails driving traffic to the sites. A couple of minutes after we went live, a heap of other big Shopify stores launched their sales at the same time and the whole platform buckled. Chaos.

That experience changed how we ran every major sale afterwards, and the fix was operational rather than just technical. The principle became smoothing the demand curve across the whole season, so no single moment overwhelms any one part of the stack - the website, the checkout, the payment processor or the 3PL. VIP early access pulls a chunk of demand forward by a day or two; paid gets sequenced to fire after email and organic have taken the first load; and the big owned sends get staggered rather than dropped all at once (the email-send side of that - why a list at scale has to go out in waves - is in Section 12: Email & SMS). As you scale, the problems don't go away, they just become bigger problems, and you're increasingly dependent on platforms you don't control. Smoothing the curve beats building for the spike, every time.

Customer Service & Post-Purchase Surge

Support volume rises with order volume. If your response times blow out at peak, you convert excited buyers into refund requests and churned cohorts. The fix is to staff and script ahead of the surge, not scramble during it.

01
Scale capacity
Ramp staffing (in-house or BPO) on a timeline that has people trained before peak, not hired during it. Layer in AI chat and macros and a solid help centre.
02
Pre-write the templates
Macro responses for the top peak issues: WISMO ("where is my order"), discount-code failures, stock questions, delays.
03
Extend the returns window
A longer returns window for December gifting reduces friction and CS load on gifts bought weeks before they're opened.
04
Make exchanges frictionless
A clean exchange flow protects the relationship and recaptures revenue instead of refunding it away.
Rob's take

Our CX team treated peak as a known event, not a surprise - hire and train ahead of the wave, roster up hard for the period, and watch first-response time and CSAT as the live dials. When those two start sliding you're going underwater, and at peak the gap between a fast, helpful reply and a slow one is the gap between a kept sale and a refund.

A robust, up-front FAQ on shipping, returns and promo terms deflects a big chunk of the WISMO and "how does this code work" volume before it ever reaches a human. For the broader retention and support philosophy this sits inside, see Section 22: Customer Support & Experience.

Tech & Site Readiness

A site crash or a broken checkout during peak-minute traffic is a direct, unrecoverable revenue loss, at the most expensive traffic of the year. Shopify peaked at $5.1M of sales per minute in 2025; that's the scale of demand your storefront has to absorb without buckling.

The readiness checklist:

1
Load-test the storefront and checkout
Test against a multiple of your expected peak volume before the season, not during it.
2
Code freeze
Freeze risky deploys 2-3 weeks before peak. The last thing you want is a fresh bug shipping into Black Friday.
3
Pre-stage the promo mechanics
Auto-applied codes, bundles and gift-with-purchase all built and live in advance, not configured the night before.
4
QA every offer across devices
Test discount stacking, edge cases and every offer combination on mobile and desktop. Broken or stacked codes erode margin and break checkout.
5
Optimise mobile and accelerated checkout
Most traffic is mobile; Shop Pay was around 32% of all BFCM 2025 orders (up 39% YoY). Fast pages and accelerated checkout convert the peak rush.
6
Have a rollback plan
If something does go wrong, know how to revert fast.
Warning
QA The Discount Stack

Overlapping codes, a bundle that doesn't apply, a gift-with-purchase that silently fails. Mis-QA'd promo mechanics either give away margin you didn't intend or break the checkout at the worst possible moment. Build the offers early and test every combination across devices before the freeze.

For the storefront and checkout fundamentals this builds on, see Section 11: Website & Conversion.

Cohort Quality & The Second Purchase

This is the discipline most brands skip, and it's where the real money is. Discount-acquired buyers churn fast and have lower LTV. If you treat BFCM as a one-off sale instead of a customer-acquisition event, you've bought a cohort at your worst margin of the year and then done nothing to turn them into repeat buyers. That's a waste.

Plan the December-to-February second-purchase journey before BFCM, not after:

  • Post-purchase flows that surface replenishment and cross-sell to the new cohort. (The mechanics are in Section 12: Email & SMS.)
  • Loyalty enrolment to convert a discount buyer into a member.
  • A non-discount re-engagement offer, content, early access, a gift, so their second interaction with you isn't another markdown that re-anchors them to discount-only buying.

Then track the BFCM cohort separately from your normal-period cohorts:

MetricWhat it tells you
Repeat-purchase rate at 30 / 60 / 90 daysWhether the cohort is coming back at all
D7 / D30 retentionEarly engagement signal
90-day LTV / CAC vs non-promo cohortsWhether you actually acquired profitably or just bought a sale

Weight your effort toward existing customers, not just new acquisition. Repeat shoppers were the single biggest BFCM 2025 growth driver, with roughly 45% higher GMV than new shoppers, and they were the fastest-growing segment. Your owned list is both your cheapest peak channel and your highest-quality cohort. For the full retention and loyalty engine, see Section 21: Retention & Loyalty.

~45% higher GMV
Repeat shoppers vs new shoppers at BFCM 2025 (Klaviyo), and the fastest-growing segment. Peak is as much a retention event as an acquisition one.

AI Discovery & Agentic Commerce

This is the new one, and it's moving fast. AI-driven traffic to retail sites was up around 805% year on year on Black Friday 2025 (holiday-season generative-AI traffic up around 693%), and those AI referrals convert meaningfully higher than the baseline. People are increasingly starting their shopping inside ChatGPT, Gemini and other AI surfaces, and the platforms (Shopify, Adobe Commerce) are committing to agentic-commerce standards so those assistants can find, compare and even check out products directly.

What that means for you: if your catalogue isn't machine-readable, you're invisible to the fastest-growing discovery channel in commerce.

The standards this section told you to watch have now landed, and the shape of the channel is settled for the moment: discover in the AI, buy on your site. Shopify's Agentic Storefronts (default-on for eligible US-selling merchants since March 2026) syndicate your catalogue into ChatGPT, Perplexity, Copilot and Google's AI surfaces via Shopify Catalog; the buyer clicks through and checks out on YOUR store, with your theme, your upsells and no platform fee. The other model - checkout inside the chat - was tried and retreated: OpenAI's in-chat Instant Checkout launched with a 4% fee in January 2026, attracted only a few dozen Shopify merchants, and was wound back by March. Walmart reported in-chat checkout converting roughly 3x WORSE than handing the customer to their own site. Your store is still where the transaction happens - which means everything this playbook teaches about conversion still applies to the AI-referred customer.

1
Verify your AI channel settings before peak
Shopify admin, Sales channels, then Agentic: confirm which AI surfaces carry your catalogue, exclude what shouldn't be there, and check your policy pages are complete (an eligibility requirement). The full checklist lives in Section 16: Other Marketing Channels.
2
Treat product data as the ranking lever
Feed-sourced product data dominates the top slots in AI shopping answers, and results are organic - no paid placement exists. Complete titles, attributes, live pricing and review schema are the whole game (GEO: generative engine optimisation).
3
Get your promos machine-readable too
AI assistants surface offers and price drops at peak. If your BFCM pricing and inventory aren't accurate in the feed, the assistant either skips you or, worse, quotes stale prices.
4
Track AI as its own channel
GA4 has a native "AI Assistant" channel and Shopify tags AI-channel orders in admin (Section 27). Watch conversion, not just sessions - AI-referred visitors convert well above baseline.
Info
Treat AI Discovery As A Peak Channel

The instinct is to file "AI search" under "experimental, deal with it later." The data says otherwise: AI-referred retail traffic grew roughly 14x from late 2024 to mid-2026 and now converts 40-50% BETTER than non-AI traffic. It's still small in absolute terms - which is exactly why it's cheap to win now. Getting your product data clean and machine-readable is table stakes, and most of that work pays off in regular search too.

Measurement & The Post-Mortem

Peak-period blended ROAS is flattering. It's inflated by organic and repeat demand that would have come anyway, so if you trust platform-reported ROAS at peak, you'll over-credit ads and mis-plan next year. Decide how you'll measure before the season, not after.

Pre-define the KPI set and the dashboard:

  • Blended MER / ROAS across all channels, not platform-reported numbers.
  • Contribution margin after discount, returns and shipping, the number you actually keep.
  • New-vs-returning split so you know whether you acquired or just harvested.
  • AOV and CPA by channel.

Then, for attribution, use incrementality, geo or holdout tests rather than trusting platform ROAS at peak. The platforms will happily take credit for sales that were always coming.

After it's over, run a structured post-BFCM debrief while it's fresh. What offers, creative and channels drove genuinely profitable revenue (not just top-line). How good was the cohort. Where ops failed. Then write it down and update the playbook for next year, so the whole organisation compounds its learning instead of relearning the same lessons every November.

Warning
Blended ROAS At Peak Is A Vanity Metric

Judging BFCM on peak blended ROAS without incrementality or margin-after-returns leads you to over-invest in channels that mostly harvested organic and repeat demand. Measure contribution margin after everything comes back, and use holdouts to find the genuinely incremental spend. That's what tells you where to put the budget next year.

For the attribution and incrementality methods this leans on (geo holdouts, MMM, incrementality testing), see Section 27: Measurement & Data; for the unit-economics and margin framework, see Section 26: Finance & Unit Economics.

Peak Season Checklist

Section checklist
24 items
Sign in to unlock this 24-point checklist and track your progress across every section.
Sign in to unlock →
Share this section

Found this useful?

If this helped, send it to a DTC Founder or Operator who’d benefit. The Playbook is FREE — the only way it grows is word of mouth.

Free access · No subscription

Go from reading to doing.

You've read the section. Sign in to score your ecommerce brand with the Health Check, track your progress across 472 checklist items, and get your tools and history in one dashboard. Free, and always will be.