Increasing Online Wholesale Orders via A/B Testing

Overview

Liberty Furniture, a furniture manufacturer and wholesaler, primarily sells to retail stores and showrooms. Traditionally, retailers placed orders via phone or email with their Liberty sales rep—a slow and inefficient process at scale. To address this, Liberty invested in driving retailers to use its online ordering platform. To boost adoption and address stagnant usage, I leveraged rapid experimentation to enhance the UX and streamline user journeys.

Current State

Below are screenshots of the current starting state of the checkout experience.

Assumptions

• The current 'Need Help?' section has low visibility on most pages. Moving it will reinforce the connection to offline processes with sales reps and ease internal concerns.
• Reducing the number of steps to complete an order will improve conversion rates and increase online order processing.
• Improving information hierarchy for action items on each page will boost awareness of optional fields and actions.
• Enhancing the contrast between the checkout and review screens will improve clarity.
• Revamping the confirmation and receipt page for a feeling of “completeness” will gain trust:
  • Simplify terms like “website confirmation number” to avoid confusion about additional steps.
  • Ensure the view/print as PDF button is fully visible.
  • Organize scattered information for better readability.
• Updating typography across the site will create a more professional and polished appearance.

Experiment

After forming my assumptions, I made changes to each page based on the existing content and known design principles.

Step 0: Shopping Cart (Pre-checkout)

Breaking out the shopping cart from the checkout stepper keeps the user from from feeling like they've already begun the checkout process.

Hypothesis for B Test:

• Moving optional fields above the action button to encourage entering them in before moving on
• Changes to the layout of the 'Need Help?' card make it easier to scan through

Step 1: Checkout

Since we removed the shopping cart as a step, the start of checkout is now step 1, shortening the perceived time it takes to check out.

Hypothesis for B Test:

• Stepper is the focal point to keep the user engaged
• Keeping form fields consistent on each page, and moved above form buttons
• Information hierarchy of billing and shipping information for clarity

Step 2: Review & Confirm

Consolidating the review and confirmation screens into one steps into one to speed up the time it takes to checkout.

Hypothesis for B Test:

• Turning input fields into read-only text gives the user a sense of completion
• Making the 'Need Help?' section easy to access on every screen of the process, near the action buttons
• Typography changes also make this screen easier to read through

Step 3: Receipt

Consolidating the receipt step into one also allowed for a quicker experience, since viewing the receipt is optional.

• Changes to the typography and layout of information make this screen easier to scan
• Inclusion of the 'Need Help?' section on the receipt page also keeps the sales rep involved if there are any questions about the order or changes that need to be made at this stage
  
• Updating the PDF button will make it more noticeable, so it's not resembling the site's table headers

Experiment Results

Using the AB Tasty integration with Hotjar, I was able to track metrics for both the A test (the cart at its existing state) and the B test. Before I gathered the results of the B test over the course of a few weeks, we began receiving positive feedback from retailers and sales representatives about the improved experience. That was a good leading indicator for our results:

• Percent of orders processed through online site - the B test processed 30% of orders, while the A test only 18%. 
• Checkout conversion rate - rate was already high at 72%, but raised 2 points with the changes to the checkout experience. This could be due to other factors rather than changes to the checkout experience.
• Increased usage of optional fields  - 13% more notes were added to orders on the B test.
• Heat map activity on 'Need Help?' card - comparative rise in heat map activity on section.

New Checkout Experience

Takeaways

This project deepened my understanding of our users and strengthened my confidence in A/B testing, using metrics to justify - rather than dictate - design decisions. Seeing design principles and my own judgment drive a measurable improvement in the checkout experience was especially rewarding. A key challenge of this project was securing leadership buy-in for the A/B test instead of implementing changes without a comparison. I overcame this by presenting a detailed experiment plan and demonstrating its value. Ultimately, I was able to refine my skills in A/B testing and am proud of the new checkout experience launch!