About Sleuth

Screenshot of the Sleuth app with the opened tracker, which displays the ability to save a symptom, behavior, treatment, measure, condition, specialist, organization and milestone

Sleuth by the numbers*

  • 62,000

    contributors

  • 5,100

    new answers daily

  • 285

    topics covered

  • 5,300

    app installs/month

  • 4/5 app store rating

  • 68%

    impression-to-install

*as of Dec 2023

What was Sleuth?

Sleuth created a radically different approach for tracking kids’ health and development.

Unlike feedback from annual health check-ups, Sleuth gave parents benchmarks based on the exact day of a child’s age. We filled the gaps in research covering how kids evolve – getting all the leverage we could out of data from 62,000 parents covering 285 topics.

At Sleuth, we focused on common challenges that cause parents to worry but don’t (initially) motivate them to seek experts’ advice: irregular sleep, challenging eating habits, speech development, anxiety, energy levels, tantrums, and so on.

It’s just amazing that you took a pretty hard moment for parents and are turning it into something so useful.

Mother of two in Menlo Park, CA

The Sleuth app had tight feedback loops that drove heavy engagement. Parents using Sleuth instantly saw benchmarks for the general population of U.S. kids for every question they answered.

Three smartphone screens displaying features of a child health and development tracking app. The first screen asks 'How strong is Sandra’s separation anxiety?' with options ranging from 'Not at All' to 'Very Strong.' The second screen shows Sandra's separation anxiety score as '75' with a note that it is 'Slightly High' compared to others, and includes a 'Separation anxiety forecast' suggesting trends over time. The third screen promotes the next recommended quiz titled 'Crying and Fits' with an option to start or defer the quiz.

We created consistent methods for measuring any symptom, behavior, and development (severity, strength, frequency, duration). This made for a much easier user experience. Sleuth’s questions were fast and familiar regardless of topic.

Three smartphone screenshots of a symptom tracking application for stomach aches. The first screenshot is for logging a new event, with fields to input the start time, severity, duration, and any notes or attachments. The severity scale ranges from 'Not at All' to 'Very Severe.' The second screenshot allows for more detailed input, with options to add specific locations on the body, frequency of the symptom, and related items from past entries. The third screenshot displays a summary of a logged event, including the start time, severity noted as moderate, duration of 3 hours, descriptive notes about the symptom, and an attachment showing an image.

Parents sought Sleuth for invaluable data that they could find nowhere else. By joining Sleuth and volunteering to contribute, Sleuth’s data grew by up to 5,000 answers per day.

Why we made it

  • Diagnose early

    As many as 50% of critical care interventions are missed by parents, teachers, and doctors at their ideal time.

  • Unresolved questions

    In Sleuth’s research, 94% of parents reported a “recent” concern about their kids’ health or development.

  • Continuous healthcare

    We treated health as a continuous process flowing from a child’s personal arc, rather than a response to incidents.

  • Local information

    Children’s caretakers are the unique go-betweens for doctors, schools, therapists, and children’s programs. They alone provide day-to-day child care.

  • Natural language

    Sleuth captured everyday & longitudinal observations in parents’ own words. This was essential to build an AI that communicated from their perspective.

  • AI in pratice

    Sleuth put machine learning and AI at the center of our designs. Our internal tests had strong diagnostic accuracy based on parents’ unstructured, everyday words.

Two product stages

Sleuth v1: Starting with storytelling

We built Sleuth v1 to support long-form storytelling by parents – as well as timelines of symptoms, diagnosis, treatments and outcomes.

Stories created a sense of community – and included essential detail. We discovered that stories frequently had important practical advice about how parents could approach doctors for their concerns and exactly whom to seek out.

In Sleuth V1, we also offered developmental and behavioral assessments. We reached out to leaders in pediatric research to get free access to authoritative, evidence-backed sleep, development, eating, and behavioral questionnaires.

A screenshot of a user interface from the Sleuth application showcasing a child's profile named 'Zuzu.' The profile includes tabs such as Updates, Journey, Expert Analysis, Assessments, and Medical Notes. A section titled 'Latest Update' includes a narrative about Zuzu's second opinion from NYU related to Asthma, Sneezing, and Behavioral Therapy. On the right, there's a detailed report titled 'Well-Being of Young Children: Developmental Milestones (12 months) Overview,' which compares a child named 'Michael's' scores in areas such as cognition and motor coordination against average scores. Charts and score ranges for different categories such as swimming and attention are visualized below.

More than 90% of people who visited Sleuth’s website either clicked on profiles or completed assessments.

Sleuth v1 was built with funding provided by Antler.co and Newlab.com in partnership with Ridgebox.com.

Sleuth v2

Sleuth v2 reflected parents’ demand for more quizzes and more quantitative feedback.

We used classic qualitative coding and ML-based clustering with OpenAI’s text-embeddings to surface 250 foundational topics in 30,000 child health histories.

A flow diagram with three interconnected circles. The first circle says 'Reviewing 30,000 child health histories'. The second circle states 'Coding and clustering to identify usefully distinct topics'. The third circle mentions 'Likert surveys on 285 topics from 0 to 17 years of age'. Arrows point from the first to the second circle, and from the second to the third, indicating a sequential process.

Version 2 included:

  • Assessments: 70+ original, extensively validated short quizzes covering 285 topics
  • Logging/Tracking: Tools for tracking events related to any symptom, behavior, treatment, program, or specialist. Parents could upload files/video and take notes.
  • Library of symptoms, conditions, and treatment data: A library of data and and summary statistics built from our existing content.
  • Discussions: Tools to allow parents to discuss any topics in the extensive library.
A composite image showing different features of the 'Sleuth' app interface. The main section is a webpage titled 'Welcome to the Sleuth Library' where parents find answers for kids' health and development, with categories like 'Cold, Cough & Allergy' listing various symptoms. Below, two smartphone screens display app functionalities: one for discussions among parents, showing comments and a prompt to share experiences; the other features a selection of topics under 'Mind,' 'Body,' and 'Social' categories, with images representing anxiety, acne, and other issues. There's also a body diagram for selecting symptoms by location.

Sleuth v2 was built with funding provided by Brooklyn Bridge Ventures, Math Venture Partners, Graham & Walker, John Friedman, Gary Vaynerchuk, and other outstanding angel investors.

Tech Infrastructure

We built Sleuth with several considerations:

  • established architecture and frameworks;
  • ability to quickly scale horizontally for millions of users if necessary, without over-engineering in the meantime;
  • flexibility to move data/cloud, switch out tools and architecture;
  • prioritizing quality and development speed.

We kept a healthy balance of essential testing and continuous delivery for a small team. At the same time, there were numerous technical improvements we shelved for a later date.

A detailed diagram of Sleuth's technical infrastructure divided into four segments. The first segment 'Hosted on Digital Ocean' outlines the development and production environments including Rest API with Python and Django, PostgreSQL database, Redis for caching, RQ Scheduler, and Nginx. The second segment 'Hosted on Vercel' covers the Serverless API with Node.js, Web App with TypeScript and React, and third-party integrations. The third segment 'Mobile Apps' shows Android and iOS app development with TypeScript and React Native. The fourth segment 'Developer and Internal Tools' lists continuous integration and delivery tools like GitHub Actions, Fastlane, Docker, and content moderation tools like Sentry and Retool.

Marketing

Sleuth ran an efficient marketing engine working closely with a UK-based agency called Favoured. We noticed that few healthcare organizations use social media ads to target specific health symptoms and developmental concerns, even though many of these affect 5% to 20% of all children at some point in their lives. This created an opening for exceptionally efficient marketing that scaled horizontally across thousands of specialized keywords from “tantrums” to “motor skills”.

Three smartphone screenshots displaying social media ad sets for 'Sleuth', a child health tracking app. The first ad emphasizes speech skills assessment through fun quizzes, the second celebrates developmental milestones with a picture of a father and child, and the third ad addresses turning anxiety into victories, featuring an image of a worried child. Each screenshot includes 'Sleuth' branding, app download prompts, and social media engagement metrics such as likes, comments, and shares.

Activity on Sleuth

Sleuth’s growth took off in the 3rd quarter of 2023. Our parent community answered 28,000 questions per week in the app, beating our goals of pacing at 1 million answers per year before the end of 2023.

A line graph titled 'Quiz Topic Completed [Total Events]' showing a sharp increase in events for the Sleuth app. The timeline starts on September 18 and ends on December 3, 2023. Initially, the line is flat, indicating stable event counts. However, there is a steep climb starting in November, peaking at 28,262 events in the last recorded week, which is a 23.75% increase from the previous week. The chart highlights significant growth in user engagement with the quizzes.

Sleuth organized its 70+ quizzes into different themes, and about 60% of people get very into those quizzes – quickly completing 5 or more.

A bar chart showing the percentage of quizzes completed on the Sleuth app for five quiz milestones. Each bar represents the proportion of users who completed that number of quizzes. The first bar shows that 100% of users completed 1 quiz (5,619 users), while the subsequent bars show a decrease with each additional quiz completed, with 81.49% completing 2 quizzes (4,579 users), 90.39% for 3 quizzes (4,139 users), 90.48% for 4 quizzes (3,745 users), and 91.08% for 5 quizzes (3,411 users). The overall completion rate across all quizzes is 60.7%.

We routinely saw individuals who answered hundreds of questions per day.

Here were some of the biggest Sleuth community topics:

  • Crying, fits, and tantrums (especially in young kids)
  • ADHD-related worries: restlessness, difficulty sitting still, difficulty paying attention
  • Child anxiety-related concerns: separation anxiety, constant worry
  • School-related concerns: anxiety about school, difficulty focusing on homework, etc.

In the media

Sleuth received favorable coverage in:

Harvard Business Review Forbes Crunchbase PIX Scary Mommy CBS LA
ABC Building while flying Out of patients we need to talk
The Uma Show clear health costs Mom 2.0 Business Insider Just Grind The Kara Goldin Show Ellevate

Who we were

Core team

  • Picture of Alex Leeds

    Alex Leeds

    Co-Founder & CEO

  • Picture of Sehreen Noor Ali

    Sehreen Noor Ali

    Co-Founder & President

  • Picture of Georgiy Tarasov

    Georgiy Tarasov

    Founding Engineer

  • Picture of Alyona Korysta

    Alyona Korysta

    Design Leadership

Partners

  • Picture of Alexander Polekha

    Alexander Polekha

    Software Engineer

  • Picture of Igor Veselov

    Igor Veselov

    Software Engineer

  • Picture of Jade Glaze

    Jade Glaze

    Engineering Manager

  • Picture of Jenna Cunningham

    Jenna Cunningham

    Digital Art

  • Picture of Regina Dacio

    Regina Dacio

    Social Media

  • Picture of Sam Battles

    Sam Battles

    Performance Marketing (Favoured)

Advisers

  • Picture of Camille Proctor

    Camille Proctor

    Condition Strategy

  • Picture of Chethan Sarabu

    Chethan Sarabu

    Pediatrics & Product

  • Picture of John Friedman

    John Friedman

    Strategy & Finance

  • Picture of Shaun Hinklein

    Shaun Hinklein

    Growth Marketing

Really, it has been other parents that have helped me to figure out this world. They have suggested treatments. They have told me things I should ask doctors about when I didn't know as much. And now… I find myself answering questions for others.

Donia, Mother in of two in FL