FeedCount

FeedCount: from a daily problem to a clearer product

I defined, designed and built a focused feeding-planning MVP, then created a public landing page to explain how its estimate works, who it is for and where its limits are before people begin onboarding.

Project type
Independent solo project
Role
Research · Scope definition · UX design · Frontend development
Stack
SvelteKit · JavaScript · PWA · localStorage · Netlify
Outcome
Live app and public product landing page
Desktop viewport of the FeedCount landing page, showing its intended audience and a product example
The landing page introduces FeedCount’s audience alongside a real product example.

Defining the useful problem

I already had an app that recorded feeding reliably. The remaining friction was translating those records into a practical decision: how much formula might remain for the day, and how could it be split across the remaining feeds?

Research changed my initial assumptions about the inputs needed for the estimate. This shaped the product definition and helped me separate the calculation task from the much broader problem of building a baby tracker.

Choosing a deliverable scope

I limited the MVP to setup, daily intake, a remaining-volume estimate and suggested bottle splits. Accounts, caregiver sharing, cloud sync, reminders and wider tracking features stayed outside the first release.

That scope made it possible to build a useful tool within limited development time. It also gave me a clear basis for evaluating feature requests: would they improve the core daily task enough to justify the additional complexity?

Making implementation trade-offs

I chose a mobile-first PWA so the product could be used from a phone without app-store distribution. Local storage and static hosting kept the first version simple to deploy and maintain.

Keeping data on the device also avoided introducing a backend for child-related feeding information at the MVP stage. The trade-off was that data would not automatically sync between caregivers or devices.

The interface separates relatively stable setup information from the inputs used during the day, with the estimate and bottle options presented around the immediate task.

FeedCount setup screen with baby profile and formula settings
Setup: saved profile and formula settings.
FeedCount main screen showing today’s intake and remaining estimate
Daily use: intake and the remaining-volume estimate.
FeedCount bottle split screen showing alternative bottle options
Bottle planning: options for the remaining feeds.

Launching and iterating

I took the project through problem definition, UX design, frontend implementation and deployment. After launch, I continued refining input behaviour and information hierarchy, maintaining the work in a GitHub backlog.

The aim of these iterations is to improve clarity within the existing scope before adding more features.

FeedCount GitHub issues backlog showing post-MVP improvement tasks
The GitHub backlog records interface improvements after launch.

Explaining the product before onboarding

After the MVP launched, the app’s main screen was also its homepage. New visitors had to enter the tool before they could understand where its estimate came from or decide whether FeedCount suited their situation.

I created a separate public landing page and moved the tool to /app. The page first makes the intended audience clear: babies under 12 months who are exclusively formula-fed. It then shows the calculation openly: baby’s weight × an age- and sex-based reference energy factor ÷ the prepared formula’s calories per 100 ml × 100.

The page also explains the boundaries of that model before inviting someone into the app. The result is a planning reference, not an exact requirement or medical recommendation, and the calculation does not subtract breast milk or solid foods. This gives visitors the context to decide whether the product fits before starting setup.

A three-step walkthrough then shows the product flow: enter the baby’s details and prepared formula calories, save familiar bottle sizes, and plan the remaining feeds. The landing page carries the app’s visual language forward with the system font, warm beige light theme, soft navy dark theme and real product screenshots.

The app remains directly available at /app for existing users and PWA installs, while the public homepage gives new visitors a clearer way to understand the product first.

Desktop viewport showing FeedCount’s formula, reference inputs and estimate limitations
The calculation is visible before onboarding, with its reference inputs and assumptions.

Delivered

  • A live mobile-first PWA
  • A focused calculation and bottle-planning workflow
  • A public landing page that explains the estimate, its limits and the product workflow
  • Local setup storage and static deployment
  • A public code repository and improvement backlog

The product now pairs a focused, direct-to-task app with a public explanation of its audience, calculation and limits, helping visitors make an informed decision before onboarding.

FeedCount provides practical estimates, not medical advice or a prescribed feeding target.

Have a role or project in mind?

I’m open to permanent roles, contract projects and part-time engagements. Get in touch to discuss what you’re working on.