DIRECT ANSWER

Build the web system first if your problem involves money or records: fees, donations, attendance, receipts. Build the app first only if your problem is reaching people on their phones every day. For a sekolah agama, madrasah, or mosque office, the problem is almost always money and records, so the web system comes first and costs less.

What does each do better?

The strengths split cleanly, as the table below shows:

NeedWinnerWhy
Fee billing, receipts, reportsWeb systemStaff work at desks; parents pay from a link, no install needed
Daily engagement, push notificationsAppLives on the phone; notifications reach people without them asking
Reaching every parentWeb systemA link in WhatsApp opens for everyone; an app must be installed first
Prayer times, qibla, offline useAppDaily-habit features belong on the home screen
Admin, roles, audit trailWeb systemOffice software is web software everywhere

Why does the web system come first?

Three reasons, listed below:

  1. The money problem is the urgent one. Late fees and unrecorded donations cost the institution every month. Engagement is nice; collections keep the lights on.
  2. No adoption barrier. A web system works the day it launches, because parents click a link they already received in WhatsApp. An app must be found, installed, and kept: a real barrier for the least techie parents, who are exactly the ones whose fees arrive late.
  3. The app inherits the backend. Built in this order, a later app is a front end on a system that already holds the students, bills, and records. Built app-first, the "backend" gets improvised inside the app and must be rebuilt when the office needs real reports.

What does each cost?

At DuaLab, an institution web system is RM 20,000 and a single-purpose app RM 15,000, each a fixed quote with the repository in your name from day one. The web system's scope (billing, FPX payments, receipts, attendance) is described under Islamic institution systems, and the whole decision usually settles itself in the free scope call: describe the problem, and the right first build is obvious within the hour.