In a nutshell. Before Black Friday, start by testing the user journeys that drive sales: search and filters, product details and promotional prices, shopping cart, promo codes, payment, shipping, account, and guest checkout. Run through them once under real-world conditions, on both mobile and desktop, then automatically replay them with every deployment until the evening of November 27. A glitch in any of these seven user flows costs more on that day than during the rest of the month combined.
Why does Black Friday ruin what was working the day before?
The website as of November 27 is not the same as the one from October 15. In the meantime, the teams have added banners, pricing rules, promo codes, delivery time slots, and sometimes a new payment method. Each addition is small. Taken together, they change a lot on the same pages, often in the last few days.
Traffic does the rest. A glitch that used to affect a few customers a day now affects thousands in an hour. And it happens at the worst possible time: on a Friday night or Saturday, when the team is short-staffed.
Here’s an anonymized example. A retail chain with 220 stores, which generates about 30% of its revenue online, experienced an outage in its checkout process on a Saturday at 8:14 p.m. The outage lasted 3 hours and 18 minutes. Estimated cost: around 500 k€. Nothing out of the ordinary: a weekend deployment, a checkout process that hadn’t been tested.
Which routes should you try first?
Seven courses cover the essentials. Start with these, in this order.
1. Search and Filters
- The search returns results for the products featured in the campaign.
- The number of results displayed matches the actual list.
- The "Promotions," "Price," and "Availability" filters work and can be combined.
- The magnifying glass and the Enter key start the search as soon as you click.
2. Product Details and Special Offer Price
- The strikethrough price, the sale price, and the percentage are consistent across the list, the product page, and the shopping cart.
- The price format is the same everywhere: comma, space, euro symbol.
- No template variable appears in place of text.
- In-store inventory and availability are up to date.
3. Shopping Cart and Total
- The total is recalculated when you change a quantity or remove an item.
- Discounts apply only once.
- The shipping costs shown in the shopping cart are the same as those on the checkout page.
4. Promo Codes
- Each campaign code works, whether or not you have a customer account.
- An expired code or one that cannot be combined with other offers sends a clear message.
- Two codes cannot be added together when the rules prohibit it.
5. Payment and 3-D Secure
- Every payment method offered is processed through to completion, including the 3-D Secure step.
- If a payment is declined, the customer is redirected to the shopping cart without it being emptied.
- The confirmation page and order email will be sent.
6. Delivery and Pickup
- The time slots and deadlines shown are the same on the product page, in the shopping cart, and at checkout.
- In-store pickup offers the right stores and the right hours.
- Addresses outside mainland France are processed as expected.
7. Guest Account and Order
- You can create an account and log in on your mobile device.
- Ordering without an account takes you all the way to payment.
- The "Forgot Password" feature does indeed send an email.
Which defects are the most expensive?
Obvious glitches are easy to spot. Subtle issues last longer and drive visitors away quietly. Here’s what we’ve found on French e-commerce sites this year (anonymized):
- a price listed as 79.00 € on the price list and 79.00 € on the product page,
- a headline showing 385 results for 38 listed products,
- a visible template variable in place of the product name,
- two different delivery times in the same shopping cart,
- phrases left in English on a French page,
- a search magnifying glass that does nothing on the first click,
- a spelling mistake in the shopping cart.
None of them prevent a customer from placing an order. All of them cause the customer to have second thoughts at a time when they're already torn between two websites.
What should you do before, during, and after November 27?
Before: Pause and replay
- Set a deadline for production rollouts, ideally one week in advance.
- Go through the complete checklist on mobile and desktop, with the real promo codes.
- Automate the seven test runs so you can rerun them within minutes after each fix.
Meanwhile: Monitor the routes, not just the servers
- Keep the production lines running continuously at the manufacturing facility.
- An alert should be triggered when a route fails, not when a customer calls.
- Have someone on hand who can decide whether to backtrack until Sunday evening.
Afterward: Keep what you've used
- Make a note of any defects you find and any paths that have broken.
- Keep the automated scenarios—they'll come in handy for the sales and the next campaign.
- Be sure to factor in maintenance time. For us, a well-structured scenario takes about 5 minutes per month.
FAQ
When should you start planning your Black Friday campaign?
Three to four weeks in advance. It takes time to make corrections and then to play through the courses again after each round of corrections.
Do we need to test everything?
No. The seven courses listed above account for the bulk of the revenue. Test them thoroughly before expanding.
Are load tests enough?
No. They check to make sure the site can handle the traffic, not whether a promo code applies or whether a price is fair. Both are necessary.
Should we test on mobile devices?
Yes, that's a top priority. A large portion of daily purchases are made over the phone, and that's where display issues occur most frequently.
Can this checklist be automated?
Yes. The workflows are defined once, then re-run each time they go into production and run continuously in production.
Go through the checklist without losing sleep over it
With Mr Suricate, these customer journeys are automated without code, replayed before each deployment, and monitored in production throughout the campaign. When a customer journey changes, the AI suggests a correction to the scenario, and your team approves it.
Of these seven courses, how many do you still test manually?
