Kahawai · E-commerce UX · Post-purchase Flow · 2023
How a single modal cleared 70 units of dead stock in 3 weeks without a discount
Role: UX Designer, Conversion Optimisation
Type: E-commerce · Behavioural Design
Tools: Figma, A/B Testing
Year: 2023
The problem
Kahawai's core product, vehicles covers, was selling consistently. But two secondary products, specific motorcycle covers and bike covers, had been sitting in inventory with no meaningful turnover. The team was producing them in bulk alongside the main line, absorbing cost without return.The business needed a way to move this stock without discounting the main product or disrupting the existing purchase flow.
The behavioural insight
Rather than pushing discounts or creating a separate campaign, I looked at the psychology of the purchase moment.
Three principles shaped the solution:
Designing the intervention
The solution was a post-purchase modal, triggered immediately after the customer confirmed their order, before reaching the final confirmation screen. Two combo options were presented simultaneously: Car cover + motorcycle cover, or car cover + bike cover.
Presenting two options was deliberate. A single upsell feels like a push. Two options shift the frame: the customer is now choosing between offers, not deciding whether to engage at all.
The Screens
Testing the dismiss CTA
The dismiss action was the variable I was least certain about. I tested two versions:
Version B consistently outperformed. The text CTA forced users to consciously articulate the cost of declining — activating loss aversion at the moment of decision. The icon simply disappeared; the sentence created a micro-moment of hesitation that the offer could fill.
Outcome
Both secondary SKUs sold out within 3 weeks of the modal going live. The business shifted its production model from bulk to combo-focused manufacturing, a structural change driven entirely by a UX decision.
Business impact
A UX intervention solved what a pricing or marketing change hadn't. The combo modal didn't just move stock, it changed how the business planned production going forward.
What I'd do differently...
The modal was placed post-checkout — after the user had already committed. That timing worked, but I'd now test introducing the combo offer earlier: at the product page level, before the customer has mentally closed the purchase. Earlier placement could capture more users and allow the combo to influence the initial decision, rather than adding to a completed one.
I would also instrument the flow more precisely — tracking scroll depth on the modal and time-to-dismiss, to understand whether the hesitation the text CTA created was the right kind of friction.