Apps that actually ship.

The SMB app trap

A small business owner has a clear idea: a booking app, a customer portal, a field-tech dispatcher, a loyalty thing. They get quoted USD 90,000 by an offshore agency and USD 240,000 by a local boutique. Both promise "a full app." Both produce 14 months of meetings and a beta nobody uses.

The trap is treating an app project the way an enterprise IT department does — full feature spec up front, big-bang launch, custom backend for everything. SMBs cannot afford that and do not need it.

The IF Digital model: ship the smallest version that solves the actual workflow, get it into 50 real users, and let the data drive the next sprint. Most of our apps cost a third of what enterprise shops quote and launch four times faster.

Do you actually need a native app?

We ask every prospective client this before scoping. The honest answer changes the budget by USD 20,000–60,000. Here is the decision tree we use:

The cases where native is genuinely worth it: field service apps that need offline sync with complex conflict resolution, apps that integrate with BLE hardware, consumer social apps where push notification CTR drives retention, and apps where the App Store placement itself is a distribution channel (i.e., people search the App Store for your category).

We will tell you which bucket you are in during the first call — before any money changes hands.

PWA vs React Native — the real trade-off

The framing that most agencies avoid giving you:

• PWA: built on the web stack you already have (or we build). Installs from a browser prompt on Android (instant) and iOS (Add to Home Screen, one tap). Cost: ~USD 8,000–18,000 for a focused 6–10 screen app. No app store approval. Updates ship in minutes, not 48-hour review cycles. Covers roughly 80% of SMB use cases.
• React Native (Expo):true native on iOS and Android. App Store and Play Store listings — which means distribution, but also Apple's 30% on in-app purchases, USD 99/year developer fee, and 1–3 day review latency on every release. Cost: USD 22,000–60,000 for the same feature set. Justified when you need deep hardware access, offline-first sync, or the App Store as a growth channel.

Our default is PWA. We escalate to React Native when the client's use case genuinely requires it and we can quantify the ROI of the extra spend. We have talked more than a few clients out of React Native builds that would have been the wrong tool for their actual problem.

What's included in every engagement

• Paid discovery sprint with a fixed-scope SOW output — job-to-be-done map, screen inventory, data model, and integration list before any dev starts.
• TypeScript codebase with documented handover — a competent developer can take over without a debrief from us.
• Sentry error monitoring + PostHog product analytics wired at launch — you see real usage data from day one, not just server logs.
• 30-day post-launch sprint included — hot-fix window, first user-feedback synthesis, one focused improvement sprint, and 60/90-day roadmap.
• Infrastructure under USD 200/month at launch — Vercel + Supabase + Cloudflare R2, scales to 10k MAU without a re-architecture.
• Full code and repo ownership transferred to your org — GitHub repo, env vars, deployment access — no retainer lock-in.

MVP-first philosophy: ship v1 in 6–8 weeks

The worst app projects we have seen share one trait: the client waited 18 months to show the product to real users. Every assumption baked into the spec at month one was wrong by month twelve. The features they built last were the only ones users cared about.

Our default is a disciplined v1 in 6–8 weeks:

• One workflow, end-to-end. Not ten features at 60%. One workflow at 100%, polished enough that a stranger can use it without explanation.
• 50 real users before v2 scoping. We do not write the v2 SOW until we have session replays, support tickets, and at least two weeks of production usage to look at.
• Explicit deletion criterion for v1 features. The SOW names what we are cutting from v1 and why. This is the hardest conversation to have and the most valuable.
• Infrastructure that starts small and scales. Supabase on the free tier covers the first 500 MAU. The architecture does not change to hit 10,000 MAU — only the plan does.

This approach typically costs 60–70% less than a full-spec build and produces a product that users actually return to, because it was shaped by real usage instead of committee decisions.

App shapes we ship for SMBs

1. Customer portals

Login, dashboard, document upload, history, status updates, messaging with the team. Built as a PWA installable from your domain. Common users: trades businesses, accountants, lawyers, healthcare practices, coaches. Replaces email-and-PDF chaos with a place your customers actually return to.

Stack: Next.js + Supabase Auth + Postgres + S3-compatible storage. Launch range 6–9 weeks, USD 18,000–32,000.

2. Booking and scheduling apps

Customer-facing booking flow + provider dashboard + admin. Calendar sync (Google, Apple, Outlook), payment hold or full charge at booking, SMS + email reminders, cancellation policy enforcement. Often replaces a Calendly + Acuity + Square stitch that costs USD 250/month and does not match the actual workflow.

Stack: Next.js + Cal.com fork or custom + Stripe + Twilio. Launch range 8–12 weeks.

3. Internal operations apps

The unglamorous stuff that runs the business — dispatch, inventory, commission tracking, field service forms, route planning. Often an Airtable-and-Zapier mess that breaks every other week. We rebuild it as a focused web app with mobile-first forms and an offline mode for field staff.

Stack: Next.js or React Native + Postgres + background jobs (Inngest or Trigger.dev). Launch range 6–10 weeks.

4. Lightweight SaaS launches

You have a customer segment, you have a wedge product, you need an MVP that takes payment and runs autonomously. We ship in 8–12 weeks: marketing site, auth, subscription billing, the core feature, and basic admin. Enough to validate, not enough to sink budget on speculative features.

Stack: Next.js + Stripe Billing + Postgres + the AI provider that fits (Anthropic Claude, OpenAI, or open-weights via Together / Groq if cost is the driver). Launch range USD 22,000–60,000.

The discovery sprint is mandatory

Every app engagement starts with a paid 5–7 day discovery. Outputs:

• Job-to-be-done map. The one workflow the MVP solves, written as a sequence of user actions. Anything not in this map is out of scope for v1.
• Screen inventory. Sketches of every screen the MVP needs, with state transitions. Typically 6–14 screens.
• Data model. Tables, columns, relationships. We show you the ERD so you can sanity-check it against your business.
• Integration list. Stripe, Twilio, Google Calendar, QuickBooks, whatever connects to your existing stack. Each one scoped and risk-rated.
• Fixed-scope SOW. A timeline with milestones, a fixed price, a payment schedule, and a written deletion criterion for anything we drop from v1.

Discovery is USD 1,200–2,500 depending on scope. Credited 100% if you proceed with the build. If we discover the project does not make business sense, we tell you and walk away — you keep the discovery document.

The stack and why we standardize

Standardization is how a small team ships fast without cutting corners. Our 2026 default for SMB apps:

• Frontend (web): Next.js 16 with App Router, TypeScript, Tailwind for utility CSS where appropriate, React Server Components by default.
• Frontend (mobile): React Native via Expo. Single codebase for iOS and Android. EAS Build for CI to App Store and Play Store.
• Backend: Postgres on Supabase or Neon. Auth on Supabase Auth, Clerk, or Auth.js depending on feature needs. Background jobs on Inngest. File storage on Supabase Storage or Cloudflare R2.
• Payments: Stripe for everything that takes money. Subscriptions via Stripe Billing. Marketplace flows via Stripe Connect.
• Observability: Sentry for errors, Axiom or Datadog for logs, PostHog for product analytics.
• Deployment: Vercel for web. EAS for mobile builds and OTA updates.

Every component is replaceable. None of it locks you in. The whole stack costs under USD 200/month at launch and scales smoothly to ten thousand active users without a re-architecture.

What's in the 30-day post-launch sprint?

Launch is the easy part. The first month is when most app projects fail — not because of bugs, but because nobody planned the operational side. Our standard 30-day post-launch sprint is included in every engagement:

• Day 1–7: hot-fix window. We sit on Sentry alerts around the clock. The vast majority of issues are wrong-config or environment mismatches — not real bugs — and fix in under an hour. We stay on Slack.
• Day 8–14: first user-feedback synthesis. We pull PostHog session replays, cluster the friction points by frequency, and produce a prioritized list. We also pull your support inbox (email, Intercom, whatever you use) and look for repeating themes. Most of the time, 80% of friction traces to two or three screens.
• Day 15–21: one focused improvement sprint. We pick the single highest-leverage fix from the feedback synthesis and ship it. This is deliberately one thing — not a v1.1 feature release. Usually a checkout simplification, an onboarding clarification, or a confusing error state that silently drops users.
• Day 22–30: handover, docs, and roadmap. We finalize the technical documentation (architecture, ERD, env vars, deployment runbook), record a Loom walkthrough of the codebase, and produce a 60- and 90-day roadmap with you. You decide whether we stay on retainer or your team takes over.

The 30-day sprint is not a maintenance contract. It is the window that turns a working v1 into a product users actually return to. Every IF Digital app engagement includes it at no extra charge.

When IF Digital is the wrong fit

• You want a clone of a successful app.Cloning someone else's product is a strategy to lose. We will not build it.
• You have no users yet and want a 30-screen MVP.Wrong shape. Start with 4 screens, 50 users, and learn.
• You need a regulated medical or financial app.HIPAA, PCI-DSS Level 1, FedRAMP — we do not serve those compliance regimes. Pick a specialist agency.

Next step

Tell us the one workflow your app would replace. A booking flow, a client portal, a dispatch board, a subscription product. We will respond with a one-page sketch, a rough budget range, and an honest opinion on whether you should build at all — or whether a PWA or no-code tool gets you 80% of the outcome at 20% of the cost.