Checkout
optimisation:
from chaos
to conversion

74% of users quit at the last step. It's not a lack of interest, it's a cognitive confusion. The wrong architecture. Trust signal is missing. Four sprints later: +28% conversion rate.

+28%

Conversions · 60 days from launch

−40%

Checkout support ticket

45g

ROI recovered

* Real data. Company name omitted for confidentiality reasons.

74% dropped out
at the final step.

A B2C SaaS platform with quality traffic and a solid product had an invisible problem: the buying funnel lost almost three out of four users in the last step. Not on the homepage, not on the product page, the last step before payment.

The heatmaps showed users scrolling up and down, clicking non-interactive spots, leaving without a clear pattern. The session recordings revealed only one thing: Structural confusionI 'm sorry . Too many simultaneous choices with no hierarchy, zero trust signal at the time of increased buying anxiety, an architecture designed to showcase the catalogue, not to guide the user towards a decision.

The team had already tried cosmetic fixes: change the color of the CTA button, add a promotional banner, reduce text. None of them worked. The problem wasn't visual, it was structural.

The diagnosis: the user did not abandon due to lack of interest. He would give up because the product didn't guide him, he would let him choose for himself at a time when he needed to be handled.

Sector B2C SaaS
E-commerce
Role Lead UX
Designer
Duration 4 Sprint
~ 8 weeks
Methodology Heatmap, Session
Recording, A/B Testing

What wasn’t working
and why.

Every element of the original checkout had a reason to exist. The problem was not the existence of individual elements; it was their hierarchy, their order, and what was missing in moments of greater decision-making friction.

Before the procedure

✕All options shown in parallel without progression: plan, frequency, payment method and promotions on the same screen
✕Zero trust signal: no warranty, no security badge, no indication of what happens after payment
✕Single generic CTA: "Buy now" without the context of what you were buying
✕Hidden order summary , users didn 't remember what they were buying in the last step .
✕No recovery from abandonment: once lost, the user was lost forever

After the operation

✓Progressive disclosures: one step at a time, with visible progress and the possibility of going back without losing data
✓Contextual trust signals: satisfied guarantee, secure payment badge and social proof placed at the exact moment of anxiety
✓CTA specifies: "Active [Name Plan] · €X/month", the user knows exactly what he is about to do
✓For example, the use of a computer's memory and the use of a computer's memory.
✓Automatic cart retrieval e-mail 1h after abandonment for identified users

4 sprints.
One clear method.

No complete redesign. No refactor from zero. Surgical intervention on the existing funnel, validated with data before each production step.

01

· Discovery week 1-2

Heatmap analysis and session recording over 2,000 sessions. Identification of abandonment points and construction of the friction map. Interviews with six users who had left. Just one question: "What stopped you?"

02

· Design week 3-4

I'm re-drawing the information architecture of the funnel. Low-fidelity wires tested with the team. Iteration on 3 variants of the step progression. Selection of the variant with the best score on internal task completion test.

03

Prototyping and testing · 5-6 weeks

High fidelity prototype with Figma. Moderate test with eight target users. Measuring task completion rate and time-on-task for each step. Found and corrected a critical friction not visible in the wires: the payment frequency selector blocked 30% of the sample.

04

A/B Testing and release · Week 7-8

A/B testing on real traffic: 50% control, 50% variation. Statistical significance achieved in 11 days. Full release on the winning variant. Post-release monitoring for 30 days with alert for anomalies.

Key insight

The problem was not the design of the individual elements; they were all acceptable. The problem was Cognitive sequence: the product asked the user to make 5 decisions at once during a time of high anxiety. No human being can do it with serenity.

The Unexpected Discovery

During the tests, 30% of users stuck to the "payment frequency" selector not for price confusion: Fear of commitmentI 'm sorry . Solution: First show the cancellation warranty, then the selector. This change alone has shifted the completion rate by 12%.

What we didn't do

We haven't redesigned the landing page. We haven't changed the pricing. We haven't added any features. Traffic was already good. The product was already valid. He was just missing the lead at the critical moment.

60 days.
Real numbers.

+28%

Conversion rate checkout

−40%

Support tickets related to the purchase

45g

To recover the investment of the project

−52%

They drop into the critical step of selection plan

The +28% conversion rate was not the result of a marketing trick or a discount. It was the result of removing friction from a path that already existed. Traffic was unchanged. The product was unchanged. The price was unchanged. Only the sequence in which information was presented to the user was changed.

The ROI recovered in 45 days includes only the cost of the project, without calculating the compound value of users acquired in the following months thanks to the optimized funnel.

60 minutes to
find out where.

The same approach that produced these results is available as UX Business Audit. I'll identify the three critical points of your funnel in one session and give you the priorities for intervention in writing within 24 hours.