Digital trade-in
Goal
Reduce the amount of unsellable items we receive from customers through our online trade-in program.
My role
Sole designer; collaborated w/ 1 PM, 7 engineers.
Impact
72%
decrease in item rejection
rate due to eligibility.
Problem
We reject 1 in every 4 items that people sent us through the mail-in trade in program.
This leads to poor user satisfaction because they sent in their item expecting to get a credit but end up not getting anything. It is also costly for the business since we incur costs to process items that we ultimately cannot resell.
Redefine project goals
I joined the project after the product requirements were already defined - the main goal of the redesign was to increase conversion and reduce customer service inquiry rate. However, I noticed that the item rejection rate was alarmingly high, we were rejecting over 20% of trade-in items that were mailed to us.
While conversion could always be improved, the more salient problem at hand was the item rejection rate. I worked with my product manager to realign the project goals to focus on reducing item rejection rate.
So why was the rejection rate so high?
The 21% rejection rate broke down to
- 16% rejected due to item condition
- 5% rejected due to item in a category that we don’t accept for resale
We suspected that the current trade-in initiation flow did not sufficiently communicate to people what items are accepted vs not.
The redesign
We added more steps in the trade-in initiation flow to make sure people are sending in items that will be accepted.
Our hypothesis was that users wouldn’t mind putting in a little more effort in giving us more information about their items, and by doing so it would actually increase affirmation that their items would be accepted.
Prioritizing style number lookup
In the new flow, the user will start their trade-in initiation by entering the style number because this is the most effective way for the user to find the exact catalog match of their item.
Matching to an exact catalog record would:
- confirm that the item is eligible for trade in
- provide the user with the most accurate payout estimate
Moving eligibility confirmation to earlier in the flow
Previously the eligibility criteria was hidden in the last step of the process right before the user finalizes the trade-in. Given the high rejection rate, it’s safe to assume that people weren’t actually reading through the criteria at the end of the flow.
I gave this information more emphasis by moving it earlier in the flow and into its own screen. This would allow the user to confirm if their item is eligible for trade-in credit before spending time filling out the entire form.
Asking for condition grading
In the new flow, we will ask the supplier to identify any flaws on their item from a list of flaw types.
This information allows us to give the customer a more accurate payout estimate based on the condition of the item, as well as serve as a second net to catch for ineligibility due to excessive wear.
I was the most unsure about this part of the new design.
I worried whether suppliers would feel confident enough to condition grade their own items, and if not, would the cost of a decrease in trade-in conversion be worth it. Therefore, I focused the first round of user testing on validating this feature.
User testing findings
- Some participants didn’t expect the condition grading step and thought the first set of eligibility questions were enough
- They didn't necessarily mind the effort of filling out the information, but rather they didn’t feel confident that they can properly assess their item’s condition
I reduced the cognitive load in this step by limiting the flaw selections to the 3 most common flaw type and worked with our photos team to retake images that showed the severity levels in a similar fabric.
Results
After the redesign launched, we saw a decrease from 5% to 1.4% of item rejection rate due to category ineligibility.
However, we saw no change in the item rejection rate due to condition. When we looked into the site data, we saw that most people were skipping through the condition grading page - they were selecting that their item has no visible flaws but we ended rejecting their items. For the next iteration, we would focus on optimizing the condition grading step of the flow.
Made by Katie • 2024