đš Workshop Recording from 06 November 2024 â eCom Unity x DRIP, with Max Rast (Klar) and Samu Hess (DRIP).
tl;dr
Loss aversion is the main lever during sale events â shoppers fear missing a deal 2.3Ă more than they value gaining one.
Top product page tests: selling-out-fast badge (~6.7% uplift), sellout risk indicator (~6.25%), and discount countdown timer.
Cart urgency message ("products not reserved") drives ~3.6% uplift and is one of the easiest wins.
Use bandit tests during peak events to shift traffic toward the winner dynamically â minimizing losses, maximizing revenue.
The psychology behind Black Friday
Black Friday works because of loss aversion. Kahneman's research shows that humans value a potential loss 2.3Ă more than an equivalent gain. During a sale event, the fear of missing a deal at its lowest price is a stronger motivator than the deal itself.
That said, price isn't everything. A family of five won't buy a two-seat sports car no matter how big the discount. The customer still needs to find the right product, size, and color. The tests below amplify urgency and perceived value at the moment the customer is already close to buying.
Product page tests
These are the highest-priority tests to run during any sale event, ranked by impact and ease of implementation.
1. Selling-out-fast badge (~6.7% avg. uplift)
Add a small label to on-sale products: "Items on sale are selling out quickly â don't wait too long." Takes roughly 2 hours to set up in a testing tool. Low effort, high impact, and the first test Samu recommends for any brand running a sale.
2. Sellout risk indicator (~6.25% avg. uplift)
A collapsible section on the PDP that shows a sellout risk level (High or Medium). Example copy for High: "This product is currently being purchased by many users. Based on sales from the past 6 hours, it is likely to sell out today." Use Medium for fast-moving products with sufficient stock, and skip it entirely for products that won't sell out. One dishonest warning destroys repeat-purchase trust permanently.
3. Discount countdown timer
A small, subtle timer tied to the current discount â placed above the price or the discount badge. Use real timers with a clear end time: 24 hours for Black Friday itself, then extend through the weekend with a stated reason ("extended until Sunday"), then "last chance" on Monday. Never use session-reset timers that restart on every visit â customers check on multiple devices and will notice. Longer durations consistently outperform short ones because they also convert the second and third touchpoints.
4. Amplified perceived discount (AOV lift)
Show the old price on the left, the new price on the right, and put a visible spacer or arrow between them. The increased visual distance makes the discount look bigger, and the product appears more valuable. In tests, conversion rates stayed similar but AOV went up â customers shifted toward higher-priced products.
5. Discount tabs by variant
When a product has many variants at different discount levels, add tabs showing the discount per variant so customers can instantly find the best deal. Reduces friction and increases the likelihood they convert on a higher-discount option they'd otherwise have missed.
Collection page tests
Discount filter (+2.09% avg. uplift)
Add a filter by discount tier (10%, 25%, 50%) on collection pages. One important insight from tests: sending paid traffic directly to the 50%-off collection performed worst. Customers who arrived expecting "up to 50% off" landed on that page, found nothing they liked, and left without exploring. Sending them to the 10â25% pages kept them browsing longer and converted better.
Sale navigation menu (+1.82% avg. uplift)
Add sale-specific entries to the homepage navigation: discount tiers (10%, 25%, 50%), plus "Low in Stock" and "Best Sellers." Low stock and bestsellers combine urgency and social proof in one click. Straightforward to implement as separate collection pages with an AB test on the menu items.
Cart tests
Cart urgency message (+3.59% avg. uplift)
Show a message in the cart: "Products in your cart are not reserved." One of the highest-impact and easiest cart tests to run during a sale. Works particularly well when some products genuinely do sell out during the event.
"Your Black Friday savings" label
Show customers exactly how much they're saving, with the event name attached: "Your Black Friday savings: âŹX." The event-specific version outperformed the generic "Your savings" in tests. It anchors the deal to a specific, time-limited moment instead of making it feel like a permanent price.
Running tests safely during peak events
The common fear is that a bad test during Black Friday is too costly. The math says otherwise: the absolute revenue impact of a -4.5% result on a regular day vs. a peak day is the same â it's just compressed into fewer hours. More traffic means faster statistical significance, which means you can catch and stop bad tests sooner.
The approach Samu's team uses during sale events:
80% bandit tests â dynamically shift traffic toward the better-performing variant as data comes in. This maximizes revenue from winners and limits exposure to losers.
20% classic A/B as a parallel backup â bandit test statistics can fail under extreme traffic spikes (two tools have been crashed by volume alone). The classic A/B runs independently as a safety net and source of clean learnings.
Set a minimum sample before the bandit starts shifting â e.g. at least 2,000 transactions per variant. Too early and the statistics are noise.
These tests apply equally to any sale event â Singles' Day, birthday campaigns, seasonal clearances. If Black Friday is too close to test in time, run them at the next one.