Straight answers on what a custom web app, mobile app, or CRM really costs in 2026 — from small SME builds to full operational systems. Written from the perspective of a Sri Lankan agency that sees both sides of the quote.

Every week someone asks us the same question in some form: "What should this cost?" Either they have a quote they think is too high, or they have a quote they think is suspiciously low, or they have no quote at all and want a sanity check before they go shopping.

This is not a generic "it depends" answer. This is the breakdown we actually walk clients through, with real USD ranges that reflect what small and mid-size businesses are paying in 2026. If you are an SME trying to budget, bookmark this.

First principle: you are buying a team, not a feature list

A website or app is not a product you pick off a shelf. It is a sequence of decisions made by people. When you hire someone to build it, you are hiring their judgement, not just their typing speed.

That is why two honest quotes for the same feature list can differ by 3x. The higher quote is usually buying you:

• Someone who has shipped this kind of system five times before and knows where it will break.
• Documented code your next developer can actually read.
• A real plan for what happens on launch day when something goes wrong.

You can buy the cheap quote and sometimes get lucky. We have also seen clients rebuild from scratch a year later because the cheap quote shipped unmaintainable code. That rebuild cost more than the expensive quote would have.

Ballpark ranges we see in 2026

These are honest ranges for work done by a competent agency or senior contractor. They are not marketplace minimums; they are not enterprise consultancy rates. They are what a serious, small-to-mid-size build looks like this year.

• Marketing / business website (up to ~10 pages, CMS, contact forms, blog): USD 800 – 3,500
• E-commerce store, self-hosted, up to ~100 SKUs, basic logistics: USD 2,500 – 8,000
• Bespoke web app for an internal team (CRUD, auth, dashboards, roles): USD 4,000 – 15,000
• Mobile app, cross-platform (iOS + Android), with a real backend: USD 6,000 – 20,000
• CRM or ERP module tailored to how your business actually operates: USD 8,000 – 30,000+
• Full operational platform (bookings + inventory + invoicing + reports): USD 15,000 – 60,000+

These are initial build costs. They do not include ongoing hosting, support, or the features you will want six months later (you always want features six months later, and that is healthy).

Why the range is so wide on each line

It looks fuzzy because the range is real. Three things push a project toward the upper end of its range:

1. Integrations. Hooking into an existing ERP, a payment gateway that is not Stripe, a third-party shipping provider, or a legacy database adds real hours. Each integration is a small project in itself. 2. Custom business rules. "Send invoice by default" is free. "Send invoice unless customer is on plan B and has an outstanding credit note from before 2024, in which case hold for approval" is a sprint. 3. Unclear requirements. If we cannot answer a feature question without asking you, that question has to become a meeting. Unclear scopes end up at the upper end purely because of the back-and-forth cost.

What makes a project cheaper (without cutting corners)

There are legitimate ways to reduce cost without shipping junk:

• Start smaller. A v1 that does the three most important things, ships in six weeks, and is used in anger, is worth ten times a v1 that does everything and ships in nine months.
• Use proven patterns. A standard admin panel, standard auth, standard Stripe checkout — we have already solved these. We charge less for them than for anything novel.
• Decide things. Faster decisions from your side really do reduce cost. A day waiting on approval costs you the same as a day of development.
• Pick a stack your team can maintain. If we hand you a system in a language nobody at your company understands, your effective cost is higher because you are locked into us.

What does not make a project cheaper

Things that feel like they should save money but usually do not:

• Giving the work to the cheapest freelancer on an offshore marketplace. This sometimes works. Most of the time the rebuild cost exceeds the original saving.
• Cutting QA. Bugs shipped to production cost about five times more than bugs caught in staging.
• Skipping documentation. You will regret this the first time your developer goes on leave.

How to sanity-check a quote you were given

If you just received a proposal and do not know if it is fair, ask three questions:

1. "What assumptions are you making, and what would make this more expensive?" 2. "Who specifically will be writing the code, and what have they built before?" 3. "What happens if I add one small feature halfway through — how do you handle scope?"

Good agencies answer all three without flinching. If the answers are vague, the quote is vague, and that is the real problem.

Where DevLK usually lands

For full transparency: most of our client projects land between USD 4,000 and USD 25,000 for the initial build. Under that and the problem is usually small enough that a no-code tool would be a better fit. Over that and we are likely replacing a full operational system, which is a different conversation altogether.

If you want an honest second opinion on a quote you have in hand, write in. No obligation. The worst that happens is you leave with better questions than you arrived with.

Original Source: https://devlk.com/blog-details.php?slug=custom-business-software-pricing-guide-2026