On November 27, 2026, Black Friday falls on a Friday. Cyber Monday on November 30. For e-commerce stores, that’s typically 8 to 25% of annual revenue concentrated over 5 days. And every year, the same incidents repeat: stores crashing at 6 PM Friday, struck-through prices not compliant with the Omnibus directive, phantom stocks sold then canceled, GA4 conversions collapsed by server-side saturation.
Stores that get through Black Friday incident-free don’t do so by luck. They follow a playbook starting at D-30 and covering five areas: infrastructure, catalog, marketing, compliance, monitoring. This article assembles the complete checklist, calibrated for PrestaShop 8 and 9 in 2026.
D-30 to D-21: infrastructure and capacity
Server capacity audit
- Measure current traffic (unique visitors/hour, requests/minute, observed peak hour).
- Estimate Black Friday peak: +300 to +800% of average traffic at peak hour (6-10 PM Friday evening, 11 AM-2 PM Saturday). Reference: Adobe Digital Insights statistics and industry data.
- Verify current hosting can sustain that peak. Shared hosting (o2switch, OVH Performance) generally can’t hold a ×8 peak — dedicated VPS or cluster needed.
Realistic load testing
Reference tool: k6 (or Locust). Scenario to script: 70% anonymous catalog visitors, 20% cart-add, 10% complete checkout. Target: hold 3× the estimated peak without TTFB degradation beyond 800 ms. Test from outside (not from the same datacenter as the server).
Pre-warmed cache
Configure a cron that crawls the 500 most-visited URLs the day before Black Friday to pre-warm the cache (Varnish, edge CDN, Redis object). Cold page at peak = store going down.
Scaling plan
- On VPS Hetzner / Scaleway / OVH: switch to a higher plan the day before (doubled CPU + RAM). Marginal cost €30-100 for 7 days, extension until December 2-3.
- On AWS / GCP: properly dimensioned Auto Scaling Group, validated health check.
- CDN: switch to Pro/Business plan if necessary for uncapped bandwidth.
D-21 to D-14: catalog and product preparation
Price audit and Omnibus directive
This is the most-controlled compliance by the French DGCCRF in 2026. For each promo product, the struck-through price must be the lowest reference price practiced in the last 30 days. Concretely:
- Audit
ps_specific_pricetable: identify already-active promos. - Document the reference price (lowest of last 30 days) for each product going on Black Friday promo.
- Configure the struck-through price explicitly with this reference price, not with an artificially inflated catalog price.
For reference, the foundational article on the Omnibus directive is scheduled on the blog for June 22: reference price compliance has become the #1 target of e-commerce inspections and remains sanctioned up to €300K for legal persons.
Pricing rules and promo code preparation
- Create all
specific_priceandcart_rulein a staging environment, test, then export/import to production. - Code precise start/end dates (e.g., November 27 0:00 to November 30 23:59).
- Verify rules don’t unintentionally stack (priority, exclusivity).
Stock and logistics audit
- Identify products at risk of running out during Black Friday: top 100 catalog.
- Verify real-time stock updates (ERP sync, deactivation on stockout, back-order handling).
- Alert logistics partner: prepare for ×5 daily volumes over 4 days.
D-14 to D-7: marketing and email warming
Email deliverability warming
If the newsletter hasn’t sent in over 30 days, Gmail and Outlook anti-spam filters will treat a large Black Friday send as suspicious. Warming strategy:
- Progressive sends to the most engaged base at D-14, D-10, D-7.
- Growing volume: 10%, 25%, 50% of the base, then 100% on D-day.
- Monitor bounce rate and inbox placement (via Mailgun or Postmark).
Ads campaign preparation
- Brief Meta Ads and Google Ads on budget increase for November 27-30.
- Verify conversions properly tracked via server-side Conversion API (otherwise apparent ROAS skewed).
- Activate custom audiences (abandoned cart, 30-day visitors, top-spenders) at D-7.
Capture lever activation
- Newsletter popup with « Early Black Friday Access » offer — capture from D-21 to D-7.
- Web push notifications: dedicated campaign on November 26 evening (D-1).
- SMS to opt-in base: send November 27 morning.
D-7 to D-1: final checks
Deployment freeze
No module, theme, or custom code deployment from D-5. Any change must be staging-validated and deployed before. Exception: critical security hotfix.
Failover plan
- Document the technical on-call number for November 27-30.
- Prepare a customer communication plan in case of incident (« temporary slowdown » banner, reassuring email for shipping delays).
- Full backup at D-2 (DB + files), stored off-site, restoration tested.
Final compliance verification
- Visual audit of a dozen promo product pages: struck-through price, reference price, displayed savings coherent.
- Up-to-date T&Cs, specific Black Friday mentions (offer duration, exclusions).
- Return policy clearly visible — post-Black Friday returns are massive.
D-day: real-time monitoring
Dashboard to watch
- TTFB and 5xx error rate via UptimeRobot, BetterUptime, or Grafana.
- Server CPU/RAM load (htop, Netdata, Datadog).
- PHP-FPM queue (
pm.status_pathexposed). - Redis / CDN cache hit ratio.
- Real-time GA4 conversions, compared to back-office volume.
- Stock fill rate on top 50 products.
Human on-call
At minimum one developer on-call Friday 6-11 PM and Saturday 10 AM-2 PM (peaks confirmed over 5 years of data). With SSH access, sudo, and back-office access. Without on-call, an incident at 7 PM Friday can kill 30% of Black Friday revenue.
D+1 to D+15: the post-Black Friday phase
Return management
Black Friday return rate is typically 1.5× to 2× the annual average (impulse buying, unwanted gifts). Prepare:
- Customer service capacity: +50 to +100% tickets over 2-3 weeks.
- Automated return workflow (prepaid label, fast refund).
- Proactive D+7 communication: « how to return if needed ».
Post-Black Friday fraud spike
Fraudsters exploit volume to slip through. Chargeback rate typically rises 50-100% in December year-over-year. Particularly watch express-shipping orders to unusual addresses.
Post-mortem analysis
At D+15, produce a report:
- Total Black Friday revenue, compared to 2025 and average daily revenue.
- Conversion rate by channel (email, ads, organic, direct).
- AOV by segment (new vs returning).
- Technical performance (uptime, average TTFB, incidents).
- Real margins after promos and acquisition cost.
- Top 5 lessons for Black Friday 2027.
The takeaway
Black Friday isn’t a marketing event — it’s a technical, legal, and logistical operation requiring 30 days of serious preparation. Stores that come through in 2026 all share:
- Infrastructure load-tested at 3× estimated peak.
- Clean, pre-warmed cache absorbing anonymous pages.
- Omnibus compliance audited and documented for each promo.
- Progressive email warming to avoid spam on D-day.
- Human on-call during the peak, with documented failover plan.
- Briefed logistics and anticipated return handling.
Without this playbook, Black Friday becomes a gamble: sometimes it works, sometimes the store loses part of its revenue and brand reputation. With this playbook, Black Friday becomes a reliable annual growth lever. The difference: 30 days of preparation starting October 27.