Updating The Knot’s Registry Store to address global shipping delays due to Covid-19
ABOUT THIS PROJECT
TIMELINE
Q1 2022
CONTRIBUTORS
Adi Rustgi,
VP of E-Commerce
Wes Tyler,
Engineering manager
Juliet O’Reilly,
Senior UX Writer
Sarah Selph,
Senior UX Researcher
Liaison in Merchandising
What is The Knot?
The Knot offers a robust suite of wedding planning tools for engaged couples.
What’s a Registry?
Couples add gifts in advance of their wedding for guests to shop from. Guests typically purchase a Registry gift within three months of the wedding.
What’s this project about?
In 2022, The Knot Registry Store was experiencing up to 70% out of stock rates, and couples were abandoning their registries rather than replacing out of stock gifts.
We tailored our interventions to work dynamically for 2 cohorts based on time to wedding, and saw big results.
What was the existing workflow like?
As couples browse and add items to their Registry, it was not made clear when one or more of these items is unavailable. And if a Registry has mostly unavailable gifts, guests don’t have much to choose from.
We hypothesized that the high level of abandonment came from folks spending time browsing, adding gifts, and only realizing later as they look at the “Manage Registry” page that potentially many of their selected gifts are actually out of stock. This negative surprise could impact their impression of the store altogether.
What constraints were there?
Due to the nature of The Knot’s contracts with suppliers, we would not be able to show an estimated availability date. This was the most common solution that other businesses were utilizing according to my market research at the time.
Another popular solution was to label these items as ‘backordered’, which implies that it will be back and is less scary than ‘out of stock.’ However, we could not confirm that any item would be back, so we could not use this term either.
We also had limited engineering bandwidth, and had to be scrappy in imagining solutions that could be built simply.
Screenshot from Figma
We released three interventions, customized for couples based on their proximity to their wedding date.
Tactic #1
Improvements to 'manage registry' screen
We added customized guidance at the top of the screen to help couples how they can best handle the shortages.
For couples farther out from their wedding, we could encourage them to ignore the out of stock status as the item would be likely to come back in stock.
For couples within three months of their wedding, we encouraged couples to remove out of stock gifts and add available gifts.
We also updated the tag from ‘Out of stock’ to ‘Temporarily out of stock’ and added a customized tool tip reflecting that time sensitive guidance. (Pictured above.)
Last, we added logic for users within three months of their wedding to bring all out of stock items to the top of the screen so that they would be easier to remove.
Old experience
New experience
Tactic #2
Replace instead of delete a gift
In order to delete a gift, a user must open a side panel which includes a replacement option carousel. This carousel had a less than 1% conversion rate, so we moved this carousel to appear after the user has already deleted their item.
Tactic #3
Make it easier to shop available gifts, without hurting brand perception
Reflecting stock availability top of funnel was deemed too risky, so we opted to add a subtle but powerful filter to the shopping experience.
This would be especially useful for the cohort of couples close to their wedding date.
We conducted usability testing with a focus on nuances within copy options. Users reported that the increased guidance was helpful, and the further transparency gave them a positive impression of The Knot.
Ultimately, we saw big results from the a/b tests:
- 30% increase of removals of out of stock items by couples close to their weddings
- 16% - 20% increase in adds from the replacement carousel
- 15% increase in the average amount of in stock items in couples’ registries
Were there next steps for this project?
I noticed that the algorithm for replacement recommendations needed to be made more robust, but we lacked the engineering bandwidth to execute on this during the project. I communicated this finding to the team with a strong emphasis on the importance of this improvement.
© 2035 by Leah Kasten