Home/Guides/Custom software cost

// Guide

How much does custom software cost? What actually drives the number

There’s no honest single answer to “how much does custom software cost?” because the price is whatever it takes to design, build, test, launch and then look after the specific workflows your people need. Six things set the number: how many workflows and screens you need, who uses it and on what devices, which systems and data it connects to, whether you build new or improve what you have, what happens after launch, and the pricing model. This guide shows how each one moves the cost and how to get estimates you can actually compare.

If you’re asking “What will this cost us, and why are the quotes so far apart?”, the second half matters as much as the first. Quotes for “the same” project differ because each vendor assumed a different project. Once you know which assumptions drive the number, you can control most of them.

The short answer: custom software cost depends on things you can control

Custom software development cost follows effort, and effort follows decisions you make: what goes into the first release, who has to use it, what it connects to and how it’s looked after. The same drivers apply whether you’re pricing a custom web application, a customer portal or an internal tool. Here’s the quick version of what affects software development cost, and which way each driver pushes it.

Cost driver
Pushes the cost up
Keeps the cost down
Scope
Many workflows, exceptions and reports in the first release
One high-value workflow first, the rest in later phases
Users and devices
Many roles with different permissions, native apps, offline use
Few roles and one responsive web app for every device
Integrations and data
Systems with no API, messy data from several sources
Documented APIs and one clean source of existing data
New vs existing
Rebuilding everything at once
Improving the parts of your current system that hurt most
After launch
No plan for maintenance, so problems pile up
A planned budget for hosting, updates, support and change
Pricing model
A fixed price on a vague scope, padded for risk
A model matched to how well the scope is known

Scope: count workflows and screens, not features

A feature list like “user management, dashboard, reporting, notifications” is a poor basis for an estimate, because each item can mean two screens or twenty. Engineers estimate from workflows: a job a person does from start to finish, such as “a customer reorders from a past order.” Each workflow breaks down into screens, the rules behind them and the exceptions.

What to count before you ask for a quote

  • Workflows. Each start-to-finish job, written as “who does what.”
  • Screens per workflow. Forms, lists, detail pages and confirmation steps.
  • Rules and exceptions. Approvals, limits, “unless the account is on hold,” and what happens when information is missing. This is where a lot of the effort hides.
  • Reports and exports. Only the ones someone actually reads, and who reads them.
  • Notifications. Every email or text the system sends, and what triggers it.

Example (hypothetical): a distributor’s customer reorder portal

Not client data. A regional industrial-supply distributor takes orders by email. Inside sales re-types each one into the order system, and customers phone to ask where their orders are. The wish list says “customer portal.” Counted as workflows, it looks like this:

Workflow
Who does it
Screens
Phase
Reorder from a past order
Customer buyer
4
1
Check order status and download invoices
Customer buyer
3
1
Approve orders over a customer’s credit limit
Inside sales
2
1
Manage customer companies and their users
Admin
3
1
Build a new order from the full catalog
Customer buyer
5
2
Sales reports by customer and product
Sales manager
2
2

Phase one is four workflows and 12 screens, and it removes the re-typing and the status calls. The catalog and reports wait until customers are using the portal. Deciding what has to work on day one, and what can follow, is the biggest lever you have on cost.

Phasing in practice

We’re modernizing Amusement Advantage’s existing contractor portal page by page, each page tested on a staging copy before production, instead of holding everything for one big launch. Read the Amusement Advantage case study.

Users, roles and devices

Every distinct type of user adds work. Each role needs its own permissions, often its own screens, and its own test cases: someone has to check that one customer can’t see another’s invoices and a sales rep can’t approve their own order. Three roles is a different project from eight.

Devices matter just as much. “Our contractors need to do this from their phones” changes the design of every screen: bigger touch targets, shorter forms, camera uploads and testing on the phones people actually carry. Planning for phones from the start avoids redesigning screens later. If you already have a desktop-only system, a mobile-friendly retrofit can be an alternative to a rebuild.

Responsive web app, native app or both

Option
What it is
What it means for cost
Responsive web app
One browser-based application that adapts to any screen size
One thing to build, test and maintain; updates reach everyone at once
Native iOS and Android apps
Apps installed from the App Store and Google Play
More to build and test, a store review for each release and developer account fees
Both
A web app for desk users plus native apps for people in the field
The largest scope; worth it for device features or offline work a browser can’t handle

If native apps are on the table, budget for the store accounts too: Apple’s Developer Program is 99 USD per membership year, and a Google Play developer account carries a one-time US$25 registration fee. Our guide to mobile web apps vs native apps for field teams walks through the trade-off.

Other user requirements that add scope

  • Poor signal or offline use on a plant floor, in a basement or on a remote site. Saving work on the device and syncing it later is real engineering work.
  • More than one language for screens, emails and error messages.
  • Accessibility for customer- or public-facing portals. If you need to meet the W3C’s Web Content Accessibility Guidelines (WCAG) 2.2, say so up front; it affects the design and testing of every screen.
  • Audit trails that record who changed what, and when.

Integrations and existing data

Connections to the systems you already run can be the least predictable part of an estimate, because the effort depends on what’s on the other side. Before anyone quotes, list every system the new software must read from or write to, and how data gets in or out of it today.

What’s on the other side
What it usually means for effort
A documented REST API with a test environment
The most predictable case: engineers can build and test against it safely
An API with thin or outdated documentation
Extra time to find out how it really behaves, and more testing
Direct access to a database such as SQL Server
Workable, but the team must learn its structure without breaking other programs that use it
File exports and imports (CSV or spreadsheets)
Simple to start, but it needs rules for timing, errors and duplicates
No interface at all
Someone keeps re-typing, or a way in has to be built first

Data migration and reporting

Moving existing data is a job in its own right: finding every source, mapping old fields to new ones, cleaning duplicates and deciding what to do with records that don’t fit. One tidy table is quick. Five spreadsheets that disagree about customer names are not.

If people already work from dashboards, say which tool. Storing data so that tool can read it, such as SQL Server tables that Tableau reports on, is part of the design, not an afterthought. Our guide to eliminating double data entry shows how to scope these connections, and our API and database integration page explains how we build them.

Building new vs improving or taking over what you have

You don’t always need a new system. If the current one still matches how you work but is hard to use, slow to change or unusable on a phone, improving it in stages usually means a smaller first project and less disruption than replacing it. If the business rules have changed underneath it, or nobody can change the code safely, rebuilding some or all of it may be the better spend.

Where you’re starting
What drives the cost
Sensible first step
Spreadsheets, paper and email
Discovering the real process and all its exceptions
Map the workflows before anyone designs screens
A system that works but is hard to use
How much of the existing code can be kept as each screen is improved
A prototype someone built internally
Hardening it: security, error handling, backups, hosting and support
Compare what it does today with what production use demands
A stalled project someone else started
How much of the existing code is sound (only a review can tell)
Get an assessment before agreeing on a price to finish it

Any price for finishing someone else’s code that comes before anyone has read the code is a guess. Our legacy modernization and software project rescue pages explain how we approach each situation.

Not sure what your first phase should be?

Your first call is with an engineer, who can talk through your workflows and systems with you and suggest what a sensible first piece of work looks like.

Talk to an engineer

The costs after launch: hosting, maintenance, support, change

Software your operations depend on keeps costing money for as long as you use it. Budget for four things:

  1. Hosting. Servers or cloud services, databases, backups, file storage and email or text sending. Cloud bills grow with usage; AWS publishes a pricing calculator for modeling them.
  2. Maintenance. Security patches, framework and library updates, certificate renewals, and keeping up as browsers, phone operating systems and the systems you connect to change.
  3. Support. Answering user questions, fixing bugs and responding when something breaks in the middle of a shift.
  4. Change. Your process will change: new products, new rules, a new customer requirement. Each change is a small project of its own.

A low build price with no plan for maintenance isn’t a low cost. Our web application maintenance checklist lists what needs regular attention.

At Effortless Presence, the same crew that builds the software hosts and maintains it, so changes come from engineers who already know your system. See maintenance and support.

Pricing models and how they shift risk

The pricing model doesn’t change how much work there is. It changes who carries the risk when the work turns out bigger than expected, and that shows up in the price you’re quoted.

Model
How it works
Who carries the risk
Fits when
Fixed price
One price for a defined scope; changes go through change orders
The vendor, who builds that risk into the quote
The scope is small, well understood and unlikely to change
Time and materials
You pay for the time worked at agreed rates
You, so you need clear visibility into progress and spend
The scope will evolve, or the work is discovery, rescue or ongoing change
Capped time and materials
Time and materials with a not-to-exceed amount per phase
Shared: flexibility for you, a ceiling on your exposure
You need flexibility and a firm budget limit
Phased
A short scoping phase, then each later phase estimated from what it found
Lower for both: each estimate rests on better information
Larger projects, takeovers and modernization
Ongoing retainer
A recurring arrangement for maintenance, support and small changes
Set by what the terms include and bill separately
After launch, for software you rely on every day

No model is cheaper by default. A fixed price on a vague scope either carries a large contingency or grows through change orders. Time and materials can cost less on a well-run project, but only if you can see where the time goes. Our guide to fixed price vs time and materials covers the contract clauses that protect you under each model.

How to get software development quotes you can compare

Every vendor fills the gaps in your request with its own assumptions. Close the gaps: send the same written brief to every vendor and ask each to answer in the same structure.

What to put in the brief

  • The problem in your own words: “Our existing system is difficult to use,” or “We need software that follows our specific process.”
  • Who uses it, roughly how many people, and on which devices.
  • The workflows, each marked must-have, should-have or can-wait.
  • Every system it must connect to, and what existing data has to move.
  • Constraints: security reviews, where it must be hosted, your own target date and, if you’re willing, a budget range.
  • What you expect after launch: hosting, maintenance, support and who handles changes.

Our software RFP template and project brief generator walks you through each of these and produces a brief you can send to any vendor, including us. Nothing you type is sent anywhere unless you choose to send it to us.

What to ask every vendor to spell out

Ask them to state
Why it matters
Assumptions and exclusions
Two quotes that leave out different things can’t be compared
How testing works, and who tests on real devices
Testing is one of the easiest things to leave out of a quote
Data migration and each integration as separate lines
The least predictable work stays visible
What happens after launch, and what it costs
The build price is only part of the cost
How changes are estimated and approved
Change is coming; the rules for it belong in writing
Who writes the code, and whether you can talk to them
The people who estimate the work should be the people who do it
Whose name the code, accounts and hosting will be in
You want to know before you sign, not when you change vendors

Compare the answers line by line, not just the totals; our guide to choosing a software development company has more questions worth asking. Still deciding whether to build at all? Start with custom software vs off-the-shelf or the build vs buy calculator.

Our custom business software page explains how we build around the way your company works, and the guides library has more plain-language help for operations and IT leaders.

// FAQ

Custom software cost: common questions

Why do software quotes vary so much?

Vendors assume different scopes, quality bars, testing and support. A written brief that every vendor quotes against, with assumptions and exclusions listed in each answer, makes estimates comparable.

Why don’t you publish prices?

Because any number before we understand your system would be a guess. Your first call is with an engineer who can tell you what a sensible first piece of work looks like.

How can we lower the cost of custom software?

Start with one high-value workflow, improve what you already have where possible, and reuse existing systems and data. Add the next piece once people are using the first.

Should we share our budget with vendors?

If you can, share a range. It lets each vendor propose a first phase that fits, instead of quoting projects of different sizes. If you’d rather not, ask every vendor to quote the same phase-one scope from the same brief.

Is it cheaper to improve our existing system than to replace it?

Often, when the system still matches how you work and its data is sound, because improving the worst parts first is a smaller project than rebuilding everything. If the rules have changed underneath it or nobody can change the code safely, rebuilding part of it may be the better spend. The rewrite vs refactor scorecard helps you weigh it.

// Bring us in

Tell us about the software your business runs on.

A new build, a system that’s hard to use, a stuck project or a prototype that needs to go live. An engineer, not a salesperson, will get back to you.

202-88-MAGIC · or email info@effortlesspresence.com