Black Friday falls on November 27. In a few weeks, your site will be frozen, your campaigns locked in, your stock in place. And the bugs already running silently on your site will still be there, just more expensive.
Many e-commerce teams believe a peak period "drowns" small problems in the volume. It's the opposite: the peak doesn't dilute a leak, it multiplies it. In most cases, monitoring and dashboards miss it, and customer service finds it weeks later.
The math no one does before the peak
A bug doesn't hit a fixed number of visitors. It hits a percentage of visitors: those on a given browser, a given sequence of steps, a given payment method. That percentage stays stable during Black Friday. Traffic, on the other hand, triples.
The loss is simple to calculate: sessions × share of visitors affected × conversion rate × average order value. Take an average site and a bug that blocks purchase for 2% of visitors:
| Normal day | Black Friday | |
|---|---|---|
| Sessions | 20,000 | 60,000 |
| Visitors hit by the bug (2%) | 400 | 1,200 |
| Conversion rate | 2.5% | 3.5% |
| Average order value | €70 | €70 |
| Loss for the day | €700 | €2,940 |
The same bug costs four times more. During the peak, the affected visitors are buyers with card in hand, one click away from a competitor.
Three peak bugs, three true stories
These cases come from conversations with French e-commerce teams. None of them took the site down, and all of them were expensive.
1. The Christmas product pages. A leading French e-commerce player, on one of its European markets. Just before Christmas, a regression hits 5% of product page traffic, on a specific sequence of steps and a specific device configuration. Result: 1,000 orders lost every two weeks, and a market at -20% of its target. The bug lived for weeks before anyone saw it.
2. The "forgot password". Same company, ahead of another peak period. Password reset stops working. Loyal customers, the ones who buy the most, can no longer log in. A double hit: lost orders and an overwhelmed customer service team. The bug was found by chance, while watching session recordings. Without that stroke of luck, how long would it have lasted?
3. The year-end redesign. An e-commerce brand launches a redesign before year end. A bug slips in during deployment and stays live through the entire year-end season. No tool caught it; customer service did, after a stream of messages from unhappy customers. Between launch and fix: the most profitable period of the year.
Why your tools don't see them
These three bugs have one thing in common: they're invisible to the tools you already have.
- The site isn't down. Technical monitoring watches availability, latency and server errors. Here, everything responds normally.
- Dashboards stay green. A 5% drop on one segment gets lost in the overall Black Friday traffic surge.
- Analytics tells you what happened. You have to know what to look for, segment, dig. No one has that time in the middle of the peak.
Marketing has tools to understand what happened, but none that show what it is losing right now.
The rule: install the sensor before freezing the site
In October, teams push their last releases: new checkout flow, revamped product pages, Black Friday promo codes. Then the site is frozen. It's the riskiest moment of the year, and often the least watched.
A release without a point of comparison is a loss you'll never see. If you don't know how each journey behaved before, you won't know what broke after, or what quietly degraded.
Waiting for the peak to start monitoring means discovering the problem when it costs the most. The right time is now: eight weeks ahead, when you can still fix things safely.
8-week checklist: the 5 journeys to check
For each one, ask yourself the same question: if this journey broke for 5% of my visitors tomorrow, how long would it take to find out?
- Login and forgot password. This is where your loyal customers go, the ones who buy the most during peak periods.
- Product pages, on mobile. Test on several devices and browsers, not just the team's iPhone.
- Promo codes. Check that the Black Friday codes apply, and that old ones aren't still circulating. If a code has already escaped its audience, see what a leaking promo code costs you per day.
- Buy now, pay later and alternative payment methods. A pay-in-4 option can fail one time in two on mobile, while every server call responds normally.
- The checkout flow, end to end. Especially if it was changed this year, or runs on a different URL from the rest of the site.
For each journey, record a baseline today: success rate, volume, associated revenue. That's what will let you put a figure on a loss on the big day. For measurement and performance, see also our BFCM 2026 technical checklist and the cost of a checkout outage.
The question to ask your team this week
Your IT team is already checking whether the site will handle the load. Ask a different question: "If a journey broke without anyone noticing on Black Friday, who would see it, how fast, and how much would it have cost us?"
If the answer is "customer service, in a few days", you know what you have left to do before November 27.