Petteri Helttula Design & Code
Back to projects

UX Audit / Checkout Flow

You Pressed Checkout. So Why Are You Still Shopping?

Overview

A self-initiated audit of two conversion-critical moments in Elisa's online store: the upsell screen after checkout and the checkout page where product reassurance becomes too weak.

5 min read

My role

UX research Checkout audit UX writing UX critique

Timeline

01/2026

Project Takeaways

What I preserved

Elisa still needs room for add-on revenue. The recommendation is not to remove upsells, but to stop making them a blocking step.

What I challenged

The flow asks the user to evaluate accessories, subscriptions, insurance, and streaming before confirming the product they came to buy.

What I'd do next

I would validate the redesign with real shoppers and compare completion rate, add-on attach rate, backtracking, and confidence before payment.

Problem framing

I entered Elisa's purchase flow as a regular customer trying to buy a phone. From the cart, I wanted to return to the product page to double-check the device, but there was no clear link back. Pressing checkout felt like the next logical move.

Instead of moving toward payment, the flow opened a full upsell page. At that point I had already decided to buy an expensive phone, but the interface shifted the task from complete the purchase to review more products. That moment of hesitation became the focus of this audit.

Product tension: upsells can make sense commercially, but checkout is a fragile moment. The design needs to support add-on discovery without weakening the user's original purchase intent.

Research

This was a heuristic audit rather than a measured research study. I reviewed the flow as a first-time shopper and focused on the points where confidence, momentum, or task clarity dropped.

  • The audit covered the upsell page shown immediately after pressing checkout.
  • It also covered the actual checkout page, where the selected phone was reduced to a small text line in the order summary.
  • I evaluated the flow against purchase intent, cognitive load, reassurance, and whether the interface matched the user's expected next step.

Analysis

The main problem is not that Elisa offers add-ons. The problem is that the add-ons become the dominant experience at the exact moment the user expects forward progress.

Checkout is a confirmation phase. The user needs to know what they are buying, what it costs, what happens next, and whether they can safely continue. A full-screen upsell interrupts that mental model and adds several new decisions before the original one has been completed.

The interruption screen turns checkout into another product evaluation moment.

The second issue appears after the user pushes through the upsell screen. On the checkout page, the phone itself almost disappears: no strong image, weak product detail, and little visual reassurance. For a high-price purchase, that is the wrong tradeoff.

What I ruled out

My first idea was to move the upsells into the left side of the checkout page, below the login prompt. I ruled that out because the left side changes meaning for logged-in users: it becomes the place for contact details, account information, and existing subscriptions.

Putting commercial upsells there would mix two different jobs: account completion and add-on shopping. The right-hand order summary is a cleaner fit because it keeps the add-ons tied to the purchase decision without competing with form completion.

Findings

  • The checkout button sets a clear expectation. When the next screen is a product discovery page, the flow breaks that promise.
  • The upsell screen creates avoidable decision load. The user is asked to consider a subscription, streaming service, accessories, insurance, and chargers before continuing.
  • The primary product loses visual priority. The phone should be the anchor of the checkout, but it becomes easy to miss.
  • The cart creates a small navigation dead end. If users cannot easily return to the product page, checkout becomes the fallback path for verification.
The original checkout confirms payment details, but the selected phone has little visual weight.

Solution

The cleaner direction is to remove the standalone upsell interruption and move add-ons into the checkout page as a secondary, optional part of the order summary.

The order summary should lead with the product the user chose: image, model, storage, color, price, and a clear edit or back-to-product option. Under that, Elisa can show a compact add-on module for relevant recommendations, such as protection, case, charger, or subscription options.

This keeps the commercial opportunity alive, but changes the behavior from gate to option.

Final design

The redesigned checkout would place the phone summary at the top of the right column, visible before any add-on recommendation. The left column stays focused on login, customer details, delivery, and payment.

Recommended add-ons would sit below the product summary in a collapsible section. The default state should be calm and skimmable, with copy that makes the choice feel optional rather than required.

The proposed checkout keeps the phone visible first and moves recommendations into a secondary, optional position.
  • Order summary: "Your phone" with image, model, color, storage, and price.
  • Optional module: "Recommended add-ons" with a clear expand/collapse control.
  • Microcopy: "You can continue without add-ons" near the action area if Elisa keeps a stronger upsell treatment.
  • Navigation repair: a visible link from cart and summary back to the product page.

Results

This was a self-initiated audit, so there are no live conversion results. The output is a design hypothesis: if upsells move from a blocking screen to a contextual checkout module, users should reach payment with less hesitation while Elisa keeps a relevant add-on surface.

The next step would be a lightweight validation round: compare the current flow against the proposed version with 5-8 shoppers, then A/B test checkout completion rate, time to payment, add-on attach rate, and backtracking from checkout.