Embedded Files
I led the UX design and data analysis for our team to build an assistant that would evaluate email content and sender details to increase send confidence and deliverability for first-time users (first 60 days).
ResultsÂ
→ After release, Email Assistant users achieved 90% first-time send success rate, a 6% increase in trial-to-pay conversion, and 15% higher campaign engagement.Â
→ Support volume for first-time users dropped by 35%.
Problem to solve
First-time users are unsure/unaware about email marketing best practices, which leads to hesitation in sending their first email to real contacts —in turn this impacts trial-to-pay conversion rates and customer support costs.
User goals
See what can be improved while creating emails with best practices in mind.
Avoid embarrassing mistakes—broken links, placeholder elements, spam reports.
Confidence in sending first email to see value Constant Contact can bring.
Business goals
Reduce support costs by surfacing recommendations frequently shared in customer support calls & chat.Â
Improve send rate for new customers in the first 60 days & trial-to-pay KPIs.
Integrate & train Recommendation and Insights AI model
Ensure high-quality emails come from Constant Contact IPs to maintain world-class sender reputation with ISPs.
My role & contributions
Best practices market research.
Negotiating time for discovery research.
Requirements building with PO & Support Ops
Concept ideation & flow diagrams.
Iterative prototyping in Figma.
Guiding UX researchers in designing studies.
Attending technical cross-functional review.
Annotated designs for development.
Data analysis for weekly leadership call.
UX quality assurance every sprint.
Key collaboration:Â
Advocated for 3 rounds of UX research when leadership initially pushed to ship faster. This research (25 participants) uncovered critical usability issues that would have increased support costs, validating the investment in validation before launch.
Early explorations
Refining designs
Multiple design iterations were informed by peer design review feedback, +20 research sessions across 3 rounds with 25 participants, and cross-functional reviews with Customer Ops.
Insights:
Positive marks should be shown by default to boost confidence.
Let users fix issues from Email Assistant instead of just providing info.
Access to Email Assistant in Editor & on Schedule pages optimizes for different user behavior (increased adoption).
 Key decisions:
Pushed for scrolling to broken links in editor before release (validated by UXR).
Add pre-header to Schedule page (bug).
Pushed for optimizing default content for all templates and republishing them to avoid false negatives (validated by UXR).
Pushed for analyzing all sent emails in the background to gather baseline data without needing to release a UI.
▸ Email Assistant summary screen from Figma UXR high fidelity prototype
â–¸ Email Assistant subject line score details
â–¸ Email Assistant sender information score details
Final design & dev handoff
I created a detailed UX implementation prototype with annotations. I tried my hand at mocking up each screen in React and linking to Gists in Github as a starting point for the dev team.Â
Sample Figma files
I created detailed implementation prototypes with React code snippets that served as working references for the development team, reducing implementation questions and accelerating delivery by 1 sprint.
“Hi Emmanuel, you got a big shoutout from our team for a very thorough work on the design along with code snippets you provided. Just wanted to say thank you!”
— Sr. Dev Manager for Email UI Team
Usage data
I put together a Google Sheet to visualize backend data from Splunk queries (provided by dev) and several Google Analytics dashboards for weekly readouts with Senior Leadership.
Future iterations
Leveraging advanced AI for helping users send more accessible emails (color contrast, text size, alt text, etc.), optimizing text content for engagement using historical data, and helping advanced users with dynamic content, variables and internationalizing their emails.
A modeless or inline solution should be explored— Initial release of the Email Assistant was limited by technical bandwidth for surfacing more contextual feedback to users in the editor.
Integrating a pre-send checklist in the email flow could greatly improve the quality of email sent through Constant Contact, since the current iteration is optional.
Expanding the assistant beyond email campaigns to optimize landing pages, social posts, surveys, events, ads, etc.
Report abuse