All Work
Case Study · StrongMind

LoudMouth

An intervention and engagement tool that delivered over 2 million personalized messages to 50,000+ students and parents, with a 98% opt-in rate.

Role
Lead UX DesignerResearch · UX Copy · Product Owner
Company
StrongMindIn-house · EdTech
Platform
SMS / Text-BasedTwilio · PowerBI
Focus
UX Writing · Nudge TheoryResearch · Data-Driven Design
LoudMouth messaging
Background

Engaging the disengaged

Imagine being a busy single parent trying to keep up with a child in a fast-paced virtual school, where falling behind by even a week often means not catching back up. Staying on top of grades, teacher contacts, and activity status feels impossible. LoudMouth was built to bridge that gap.

LoudMouth is an intervention and engagement tool that delivers actionable academic data directly to students and parents via text message, early, often, and without requiring them to log into anything. My role spanned research, UX copy, and at one point Product Owner.

Product vision: Loud Mouth seeks to engage the disengaged, reduce clerical work, and humanize the teacher.

Twilio · SMS API MTurk · Crowdsourcing PowerBI · Analytics UX Writing · Microcopy Nudge Theory · Behavioral A/B Testing · Optimization

The Product

Two million messages, zero interfaces

LoudMouth is entirely interface-less: it's a text message facilitator that talks to internal APIs to retrieve student data and deliver it at the moments we found had the greatest impact. Over nearly two years, it sent 2,000,000+ messages to over 50,000 students, parents, and guardians.

The message cadence was designed around behavior change principles:

  • System Welcome: sent at enrollment with opt-out info and expectations
  • Class Welcome: first Monday of the term with teacher name, class, and contact info
  • Grade Message: every Wednesday morning with current grade data
  • Activity Message: sent at customer-defined intervals to flag students falling behind
LoudMouth student message LoudMouth grade message LoudMouth activity message

Sample student-facing messages across the LoudMouth cadence

LoudMouth full message set

Attendance reminder through LoudMouth


Research & Iteration

Crowdsourcing clarity with Amazon MTurk

With each of 15+ messaging iterations, we ran studies to validate copy decisions. One persistent confusion: users couldn't tell whether "current overall grade" meant a running average or a final grade. We needed a fast, data-backed answer.

We turned to Amazon Mechanical Turk, sending a short survey with context and multiple-choice interpretations to 229 participants at $0.50 each. The goal was quantitative clarity on a specific verbiage question before committing to another iteration.

The message: "Your student's current overall grade is 86% in Algebra 2A as of 1/4/19 with 22 days left." Which statement best describes it?

Results (229 participants)
  • 141: "Your student's current grade in Algebra 2A is 86%." (winning interpretation)
  • 55: "Your student's final grade in Algebra 2A is 86."
  • 22: "Your student has completed 86% of Algebra 2A."
  • 10: None of the above
  • 11: Other (assignments, exams)

Removing the word "overall" reduced support calls about this specific confusion by over 70%. A single word change, validated in hours, with measurable impact.


Reporting

Data for administrators

Beyond the messages themselves, we built a PowerBI dashboard that surfaced response patterns and engagement trends, giving school administrators visibility into how students and parents were interacting with the system.

PowerBI dashboard

PowerBI dashboard providing engagement insights to administrators


Results

Measurable impact, real students

Within the first four months, clients reported increases in both course completion and seat time with no other obvious influencing factors. The product continues to grow every day.

2M+
Messages sent
50K+
Students & parents reached
98%
Opt-in rate
70%
Reduction in grade confusion support calls

Clients reported an increase in course completion rates by 5% and seat time by 12% within 4 months of deployment.


Learnings

What I learned

  • 1UX copy is design. A single word can cause consistent confusion at scale, and removing it can have a measurable, trackable impact on support load and user comprehension.
  • 2Crowd-sourced validation like MTurk is a powerful tool for copy and metaphor testing when you need quantitative signal fast without setting up formal research sessions.
  • 3Interfaceless products still require deep design thinking. The absence of a UI makes the words the entire experience; every character matters.