Service · web application development company

Web app development

User roles and workflows drive cost most, then third-party integrations, real-time features and compliance needs; a focused internal tool costs a fraction of a multi-tenant customer portal.

Product Engineering practice: abstract red and ink artwork from the Red Balloon reel

Spreadsheets, email threads and WhatsApp groups run more of Indian business than anyone admits, and the right web application development company turns them into one system people actually use. Red Balloon builds web apps for manufacturers, diagnostic chains, property operators, SaaS companies and startups: customer portals, internal tools, dashboards and the APIs that connect them. Custom web application development starts with mapping the workflow, roles and approvals, then shipping a usable first version quickly and improving it in short cycles. Where field teams or customers work on phones, progressive web app development gives an installable, offline-capable experience without the app stores. Code, cloud accounts and data sit with you from day one.

Where this works

Real-world situations we're built for

Illustrative · not a client
Manufacturing

Manesar auto-parts supplier tracks dispatches and quality checks in Excel files shared over email

What we'd do
Web app for dispatch planning and tablet QC checklists, with Tally integration for invoicing
What moves
Dispatch errors down, time from dispatch to invoice down
Diagnostics

Delhi diagnostic lab chain sends reports to patients and doctors manually over WhatsApp

What we'd do
Patient and doctor web portal with OTP login, report history and role-based access
What moves
Manual report sends down, doctor portal usage up
Co-living / property management

Gurugram co-living operator manages 1,200 beds on spreadsheets; rent tracking and complaints lag

What we'd do
Tenant and ops web app with UPI rent collection via Razorpay, complaint tickets and an occupancy dashboard
What moves
Goal: on-time rent collection up, complaint resolution time down, measured as on-time rent collection rate before and after, over a fixed period
How a project runs

5 phases, each with a deliverable

What happens, how long it takes, what you get and what we need from you.

Discovery

1–3 weeks

Workflows, user roles, data audit, integration review, performance and security needs

You get: Scope, backlog and estimate

We need: Process owners and system access

Architecture and design

2–4 weeks

Stack decision, data model, authentication, UX flows, Figma prototype

You get: Architecture document and approved prototype

We need: Hosting decisions and brand assets

Build

6–20 weeks

Sprints for APIs, integrations, admin tools, automated tests, CI/CD

You get: Staging environment and sprint demos

We need: Product owner and test data

Hardening and launch

2–3 weeks

Security testing, load testing, backups, monitoring, data migration; check first that a full restore from backup completes on a clean environment

You get: Production launch and runbook

We need: Cloud account in your name, go-live sign-off

Run, patch and extend

Ongoing

Monitoring, patching, feature sprints, cloud cost reviews

You get: Monthly report

We need: Priorities and user feedback

A typical engagement

From first call to result

Illustrative · not a client

Delhi industrial distributor, 60 staff, Tally for accounts, orders by phone and email, 2,000 active dealers

Weeks 1–2DiscoveryShadow the order desk, map dealer journeys, audit Tally data and price lists, define user rolesScope, draft data model, integration list
Weeks 2–3ArchitecturePick stack (e.g. Next.js, Node.js, PostgreSQL in AWS Mumbai), auth design, Tally sync approachArchitecture document and estimate
Weeks 3–6UX designDealer ordering flows, quick reorder, stock visibility rules, admin screens; tested with 6 dealersApproved Figma prototype
Weeks 5–13BuildSprints for catalogue, tiered pricing, cart, credit limits, Tally sync, WhatsApp order confirmations via a BSPStaging app demoed every sprint
Weeks 12–14Security and performanceOWASP ASVS checks, k6 load test at 3x month-end peak orders, backups, 180-day log retention per CERT-In directionsTest reports with fixes closed
Weeks 14–16Pilot100 dealers onboarded with order-desk support; most common blocker: dealer GSTINs or price tiers not matching TallyGo-live decision
Weeks 17–20RolloutAll dealers invited in waves, training videos, monitoring dashboardsLive app and support runbook
Where it starts450 orders/week by phone, 6% order errors, dealers call to check stock, no online order history
Where it endsA dealer ordering app connected to Tally, running in the distributor's own AWS account. Weekly watch: share of orders placed online, order error rate, p75 page load.
What we measure

The numbers we report on

Every metric has a formula, a target, a source and a reporting rhythm.

MetricHow it's calculatedTargetSourceCadence
Largest Contentful Paint (LCP)p75 time to render the largest visible element, from field data≤ 2.5 s at p75 (Core Web Vitals 'good')Chrome UX Report, web-vitals library sending to GA4Monthly
Interaction to Next Paint (INP)p75 latency from user interaction to next paint≤ 200 ms at p75 (Core Web Vitals 'good')web-vitals library, PageSpeed InsightsMonthly
UptimeMinutes available ÷ total minutes in the monthSet an SLO such as 99.9% and track the remaining error budgetBetter Stack, UptimeRobot or CloudWatchWeekly
API p95 latency95th percentile response time per key endpointRule of thumb: under 500 ms p95 for reads and under 1 s for writes; alert when a deploy adds 20% or moreGrafana, Datadog or New Relic APMWeekly
Error rateServer 5xx responses plus front-end exceptions ÷ total requestsRule of thumb: under 0.1% of requests; roll back any deploy that doubles the error rate within the first hourSentry, CloudWatchDaily
Workflow completionUsers completing the key workflow ÷ users who started itSet from launch baseline, then aim upward each quarterGA4 or PostHog funnelsMonthly
Cloud cost per active userMonthly cloud bill ÷ monthly active usersSet from baseline; investigate when it rises faster than usersAWS Cost Explorer or Google Cloud Billing reportsMonthly
Automated by default

What runs itself after launch

The manual work we take out of this service, and the tools we use.

When: Merge to mainTest and deployGitHub Actions or GitLab CI running tests, then deploying to staging and productionRemoves: Manual deploy errors and untested releases
When: Every pull request and weeklyDependency vulnerability scansDependabot, Snyk or npm audit in CIRemoves: Known vulnerable packages reaching production
When: Daily backup, monthly restore testBackups and restore testsManaged database snapshots (RDS, Cloud SQL) plus a scheduled restore jobRemoves: Finding out backups don't work during an incident
When: Downtime or error spikeUptime and error alertsBetter Stack or UptimeRobot plus Sentry alerts to Slack and an on-call phoneRemoves: Users reporting outages before you know
When: Pull request openedPreview environmentsVercel or Render preview deploys for each pull requestRemoves: Stakeholders seeing features only after they are merged
Rules we work to

Standards and regulations

OWASPOWASP ASVS and OWASP Top 10

ASVS gives testable security requirements by level; the Top 10 lists the most common web risks to test against

W3CWCAG 2.2

Accessibility criteria for forms, contrast and keyboard use; level AA is the usual target

MeitYDPDP Act 2023 and DPDP Rules 2025

Consent, purpose limits, security safeguards, breach notification and retention rules for personal data in the app; main duties apply from 13 May 2027

CERT-In (MeitY)CERT-In Directions of April 2022

Report specified cyber incidents within 6 hours and keep ICT system logs for 180 days within India

What's changing

Platform and policy updates that affect this work

We track: OWASP project news, web.dev blog, CERT-In advisories, Next.js blog.

GoogleInteraction to Next Paint replaced First Input Delay as a Core Web Vital

Heavy JavaScript and slow event handlers now count against measured user experience

React (Meta)React 19 released as stable, with Actions and Server Components support

New builds can use React 19 features; older libraries may need upgrades first

Vercel (Next.js)Critical middleware authorisation bypass CVE-2025-29927 disclosed and patched

Apps relying only on middleware for auth needed patching; keep access checks in server code too

MeitYDPDP Rules 2025 notified; main duties apply from 13 May 2027

Web apps need consent records, retention rules and breach response built in

Before you hire anyone

Questions to ask any agency, including us

Can we see the repository and documentation from a web app you handed over?

Why ask

Code quality and documentation decide your future maintenance cost

A good answer

A redacted repo walk-through with README, architecture notes, test suite and CI setup

Who are the engineers on our project, and how much of their time is ours?

Why ask

Engineers shared across many clients slow delivery

A good answer

Named roles, allocation, overlap hours and how replacements are handled

How do you estimate, and what happens when an estimate is wrong?

Why ask

How variance is handled matters more than the first number

A good answer

Story-level estimates, a tracked burn-down and agreed rules for overruns

What do you test automatically, and what manually?

Why ask

Without automated tests, every change risks breaking old features

A good answer

Unit, integration and end-to-end coverage goals, plus manual exploratory testing before releases

How do you handle production incidents after launch?

Why ask

Downtime costs revenue and trust

A good answer

Support hours, response times, a runbook and post-incident reviews shared with you

FAQ

Questions buyers ask us

Straight answers from the practice lead. Highlighted text marks Red Balloon figures we confirm before publishing.

How long does a web application take to build?

A focused web app with 3–4 user roles typically takes 10–16 weeks from discovery to production; platforms with many integrations or approval workflows take 5–9 months, delivered in releases.

  • Two-week sprints with a working staging build after each, so you review software, not slides.
  • Third-party approvals, such as payment gateway KYC or API access, often decide the launch date; we start them in week one.
What technology stack do you use for web applications?

By default, TypeScript with React or Next.js on the front end, Node.js (NestJS) or Laravel on the back end, PostgreSQL and Redis, hosted on AWS Mumbai or Hyderabad.

  • We match your existing .NET, Java or Python stack when your team will maintain the app after handover.
  • Infrastructure is written in Terraform and deployed through GitHub Actions, so environments can be rebuilt from code.
  • Monolith first. Split out a service only when one part needs independent scaling or its own release cycle; an early split mostly adds deployment work.
How do you secure a web application and meet Indian data rules?

We build to OWASP ASVS Level 2, encrypt personal data in transit and at rest, keep data in Indian cloud regions where required, and design consent, access and deletion flows for the DPDP Act and Rules.

  • CERT-In's 2022 directions require reporting specified cyber incidents within 6 hours and keeping system logs for 180 days within India.
  • An independent VAPT runs before launch, with a CERT-In empanelled auditor for regulated clients; high-severity findings are fixed before go-live, not deferred to the next sprint.
  • Most DPDP Rules 2025 duties apply from 13 May 2027, so consent notices and breach processes should be built in now, not retrofitted.
How do you measure whether a web application is working after launch?

We agree three to five measures before build, such as task completion time, weekly active users, error rate and support tickets, then instrument them in product analytics and monitoring from the first release.

  • PostHog or Mixpanel for product events, Sentry for front-end and back-end errors, and uptime checks with alerts.
  • Typical targets: 99.9% monthly uptime (about 43 minutes of downtime) and p95 API response under 500 ms.
  • A monthly review compares the numbers with the business case agreed in discovery.
Who owns the code, cloud accounts and IP of the web application?

Your company does: code lives in your GitHub or GitLab organisation, cloud and third-party accounts are opened in your company's name, and IP is assigned to you under the contract.

  • Handover includes architecture notes, environment setup, runbooks and admin credentials stored in your password manager.
  • Open-source licences are listed so your legal team can check obligations from GPL or AGPL components.
  • No proprietary framework of ours sits inside the code, so any competent team can take over.
  • Check the domain, payment gateway and Apple and Google developer accounts too; these are the accounts most often left in a vendor's name.
Back to the practice

Product Engineering →

Start with an audit

Ready for
lift?

A paid two-week Growth Audit covers your funnel, ads, site, tracking and automation gaps, and ends with a 90-day plan you can run with or without us. The fee is credited against your next engagement.

Prefer to talk? Call +91 97114 48083, or email info@redballoon.in.

A person from the team replies to every enquiry.

What do you need?
Your details are sent over a secure, encrypted connection and only used to reply to you.