E-shop redesign

Summary

I led an e-shop redesign that removed the key barriers in the checkout flow and the confusion around discounts. It raised the conversion rate by +2.05 pp (a relative +183 %), lowered the bounce rate (−59.6 %) and cut discount-related support tickets (−33 %).

The starting point

Since its launch, the e-shop had barely changed in 7 years. It was neglected both technically and visually, which hurt revenue. So we chose a full redesign instead of small, incremental changes.

A secondary problem: support received a large volume of tickets about the same recurring issue — applying a discount in the cart didn’t work as expected, and people didn’t know they could add a gift to the cart.

Hypotheses

We expected a large redesign to lift the conversion rate at least to the category average and to improve how the brand was perceived — as modern and trustworthy.

H1 — Visually distinguishing the “add gifts and discounts” feature on the product page increases its recognizability and lowers the cognitive load of the task → fewer tickets (−50 %).
H2 — Introducing a modal that flags the option to add a discount / gift to the cart lowers the number of tickets for that issue (−50 %).

The solution

The analysis showed that users didn’t expect to have to activate a discount as a separate action in the checkout process. The system demanded steps that didn’t match their mental model (how discounts usually work in e-shops).

Based on this, I changed the button’s communication (visuals, copy) to match user expectations as closely as possible. For applying a discount to an order, I designed a check — a modal that flags the option to activate the discount for the order.

The result

After the redesign went live we saw the conversion rate rise by +183 % (i.e. by 2.05 pp). As a side effect, the bounce rate dropped from 63 % to 28 % (−59.6 %).

Six months after launch we ran a test aimed at reducing support tickets around applying a discount and gift to an order. It also came out positive — a −33 % drop in tickets.

Insight

Users treat a discount as a feature that should just apply — they expect the system to do it automatically. When the system requires an extra action without a clear mental model, users assume the discount is broken — and turn to support.