Setting Buy with Prime merchants up for success.
Role: Lead designer
Team: 1 PM, 8 engineers
Surfaces: Desktop and mobile browser
In April 2022, Amazon introduced Buy with Prime, which allows merchants to offer Prime benefits directly on their e-commerce sites, including fast, free delivery, hassle-free returns, and a seamless checkout experience.
The merchant setup process is complex, involving product catalog management, UI configuration, form entry, and account linking.
Buy with Prime needed a merchant setup experience that would pull all these services together.
Context
Background
Merchant setup began as a simple list of tasks, but as our original (highly confidential) product became more defined, it grew into an unmapped maze of dead ends and dependencies.
Then the org pivoted to a new product and decided to launch an invitation-only experience within 3 months.
Our new product, Buy with Prime, had better product market fit and allowed us to quickly launch an incremental product (which means we’d finally be able to collect data and talk to real merchants). However, my team had 3.5 weeks to go from requirements creation to design handoff.
Our P0 performed about as well as expected. Amongst the papercuts, I saw an overall need for a more guided setup experience over the open-ended, freeform experience we launched.
Problem
How might we create a simple, guided setup flow that helps merchants:
Feel confident in the look, feel, and functionality of their shopper experience
Understand how the Buy with Prime tool impacts other parts of their existing e-commerce business configuration
Challenges
Split ownership of the setup flow
Based on our original charter, feature teams owned the setup pages related to their feature* (e.g. Catalog team owns all Catalog-related pages). For each change I proposed, the feature team had to prioritize that work against their own roadmap.
Highly confidential nature
Until we launched, we couldn’t get feedback from actual merchants.
Volatile org roadmap
Constant priority shifts and aggressive launch dates resulted in a constant battle for tech resources.
Prime shopping expands beyond Amazon.com
April 20, 2022
Process
Discovery workshop
I set up a three-day cross-functional workshop, including POCs from the six feature workstreams involved in the setup flow.
Workshop result #1: Stakeholder alignment on reframing success as building merchant confidence instead of speed to completion.
Workshop result #2: Identification of areas of opportunities, low-hanging fruit, risks, and areas for further investigation, prioritized against complexity/effort and customer impact.
This set the stage for a guided model that biased towards clarity and merchant education instead of reducing steps and clicks.
North star
I began working on a north star “guided onboarding” concept that addressed the most important improvements we identified.
Guide and explain instead of upsell
Digestible chunks instead of long, dry forms
Consistent UI between steps
Avoid navigational pogo-sticking
Add sales tax setup
As the owner of our console’s homepage, I added a post-launch task section, which allowed us to 1) more rigorously define pre-launch tasks and 2) continue guiding the merchant after launch.
Leadership review
I led the first low-fidelity design review with Buy with Prime leadership.
It was an influential step in formalizing low-/mid-fidelity leadership reviews so that at earlier stages of product definition, leaders focused on the right discussion points and designers saved time and energy otherwidse spent on refining fidelity, visuals, and copy.
Screenshot of workshop session brainstorming merchant setup goals, with clusters around Set expectations, confidence and Speed
Screenshot of workshop session mapping areas of opportunities against impact vs. effort/complexity
Process
Design
For a public launch, all workstreams needed to ensure their end-to-end CX could be completed by merchants self service.
Although we would have to cut scope to meet a new aggressive deadline, this was the perfect opportunity to drive forward my proposal of a guided, linear setup flow.
While onboarding a new PM and TPM into the space, I met with feature owners to advocate for a consistent, guided setup flow—including the changes they would need to make to their CX, due to our split tech ownership model.
“Guided onboarding” P0 improvements:
1. Funneled setup experience
Hiding the console’s navigation prioritized guidance over the freeform flexibility that resulted in setup abandonment. I also added setup-specific navigation to show progress through setup, and “side quest” ingresses for important actions that needed to be available at any time (customer service contact).
2. Sales tax setup
I added a high visibility ingress to set up sales tax and clarified merchant obligation, because merchants were missing the step (assuming Amazon would charge sales tax for them), putting them at risk of being audited.
3. Navigation from step to step
I added a consistent, visible navigation to take merchants from one onboarding step to the next, moving towards a model that would remove the need to return to the homepage between each step.
4. Copy improvements
Our copy needed to guide and educate instead of upsell. Merchants wanted clarity on why we asked for certain data and how their choices would affect their shopper experience and other parts of their e-commerce configuration.
The merchant setup homepage lists the four steps required to set up Buy with Prime
The first setup step requires entering business details
Outcomes
The “guided onboarding” launch increased setup completion from 43% to 71%.
“This was one of those launches where the data was night and day different between pre- and post-launch; big congrats to the UX team (especially Vicki) on delivering such high-impact design.”
- Senior manager, PM-T
Post launch
Improvements
To scale my influence and push for crucial improvements, I synthesized existing research, hunted down and analyzed sales onboarding calls, pulled in design heuristics, and led a workshop to gather inputs from sales, research, design, and product. I produced a document detailing all known papercuts, deferrals, and gaps of the merchant setup CX. One year later, the team is still using this doc.
Before transitioning to a new space, I created other artifacts to help guide my team towards next steps: a phased, modular north star, refreshed tenets, and design decks to make the case for tech resources to solve two crucial merchant pain points.
I also began working with my engineering team to propose single-threaded ownership across the setup CX, which would unlock our ability to drive greater consistency and ship improvements at greater velocity.
P1 design proposal to improve product-mapping CX (copy redacted)
Learnings
Refine north stars only to the point of usefulness. Diving too deep into details meant that priorities and resourcing would change before I finished.
Push through tenets early on, especially for a setup flow where crisp, tight-knit alignment is needed from all stakeholder teams.
Work more closely with tech during design discovery to get crisp on the right-sized improvements and get increased buy-in early on, especially with constant PM turnover.
Have any questions?