How Squarefare Cut Meal-Plan Time to Result by 98.4% With a Nutrition Web App

02The system before

The system before: one spreadsheet and manual follow-up

Every personalized plan depended on a single spreadsheet workbook holding a complex, WHO-provided formula that calculates a person's macronutrient needs from age, gender, height, weight, goal weight and workout intensity. Around it, the process was manual:

1. 01

   Collecting customer details

   The team contacted each customer to capture their current lifestyle, workout intensity, diet preference and fitness goal, then updated its data source by hand. That isn't viable at scale.

2. 02

   Calculating nutrient needs

   Working out how much of each nutrient every customer needs was complex and time-consuming.

3. 03

   Matching dishes

   Selecting dishes that met both nutritional requirements and personal preferences consumed staff time and was open to manual mistakes.

Personalization was the product, but the process behind it could only grow as fast as the team could contact customers and edit a spreadsheet.

03The risk

Why it was risky to change

The calculation is the core of what Squarefare sells. Moving it out of a spreadsheet had to make it more reliable, not just faster.

• RiskFormula fidelity

  The app had to reproduce the WHO-provided formula exactly, for every combination of age, gender, height, weight, goal and workout intensity.

• RiskData quality at the source

  Customers now enter their own details, so inputs had to be structured and validated rather than transcribed from conversations.

• RiskMenu constraints

  Suggestions could only include dishes actually on the menu, and had to respect each customer's meal preference.

• RiskFirst impressions

  The profile flow is one of the first things a new customer does, so it had to feel effortless.

04The architecture

The architecture: the nutrition formula as a production service

We engineered a web application that collects each customer's height, weight, gender, workout intensity and meal preference, and applies the WHO-provided formula to calculate their macronutrient needs.

The calculated requirements are then matched against the dishes available on Squarefare's menu, so each customer receives suggestions that fit both their nutrition targets and their preferences, with no spreadsheet or manual calculation in between.

What we engineered

• Customer profile capture

  A web flow that collects height, weight, gender, workout intensity and meal preference directly from the customer.

• Macronutrient engine

  The WHO-provided formula, implemented in the application, calculates each customer's macronutrient needs.

• Dish suggestions

  Calculated requirements are matched against dishes available on the menu, respecting meal preference.

• Consistent results

  The same inputs always produce the same result, removing the variation of manual calculation.

Key tradeoffWe implemented the formula in the application rather than keeping the spreadsheet as the calculation engine. It meant validating the logic carefully up front, and in return every result is consistent and the process scales with the customer base instead of the team.

05The results

The results

• 98.4%shorter

  Time to result was reduced by 98.4%.

• 42%more customers

  A frictionless customer experience grew the customer base by 42%.

The outcome has also drawn interest from more restaurants for onboarding.

06Is this your situation?

Is your organization facing the same pattern?

This architecture applies wherever critical expertise lives in a spreadsheet. It fits organizations where:

• A critical calculation, pricing or eligibility model lives in one spreadsheet that few people understand.
• Staff collect customer details by phone or email and re-enter them by hand.
• Personalized recommendations depend on manual work that limits how many customers you can serve.
• Results depend on who did the calculation.

An architecture audit identifies which of your processes run on spreadsheets, and what moving them into production software would return.

07Questions

Frequently asked questions