The question I hear most often, usually after a few minutes of polite small talk, is some version of this: just tell me roughly, what does something like this cost?
Not a sales pitch. Not a feature list. A number. Four grand? Forty? A hundred? They are sitting on a quote, or thinking about getting one, and they have no idea whether the right answer is in the low thousands or somewhere that would make the finance person wince.
I get why that makes people nervous. Nobody gives a straight answer online. Agencies hide behind "it depends" until you have sat through three discovery calls. Freelancers sometimes toss out a number that is either too cheap to be true or scary enough that you put the whole idea back in the drawer. If you finish this piece knowing what actually moves the price, and whether you are being taken for a ride, it has done its job.
The question everyone actually asks
You are not asking me to build the thing yet. You are trying to work out if the idea is even in the right ballpark for your business. Maybe you have ten staff and a spreadsheet that three people maintain differently. Maybe you have forty staff and four systems that do not talk to each other. Either way, you want a sanity check before you invest time in a proper scope conversation.
Fair enough. Here is the straight answer I can give without guessing: custom software in New Zealand, built properly for a business your size, usually lands somewhere between about $4,000 and $60,000 + GST for the first useful version. That is a wide band on purpose. The rest of this article is about what pushes you toward either end.
Why nobody gives a straight price on day one
Scope is genuinely unknown at the start. I can hear the problem in twenty minutes. I cannot see every edge case in your business until we have walked through it properly.
If I give you a fixed number before that, one of two things happens. I pad the quote so hard that you walk away, or I under-quote and someone eats the difference later. Usually the developer eats it, which means corners get cut, or the project stalls when the budget runs out. Neither is good for you.
That is not evasiveness. It is honesty. The goal of a first conversation is not to nail a price to the wall. It is to work out whether custom is the right move at all, and what shape the work might take. Once we have that, I can give you a number you can trust.
What actually moves the number
These are the levers that matter. Not in theory. In every quote I have written in the last fifteen years.
How much it has to do, and how much is custom thinking. A login screen and a CRUD table is not the same job as a workflow where three departments hand work off to each other with exceptions on Fridays. Standard parts (accounts, password resets, basic admin) are well understood. The bespoke rules in your business are where the hours go.
Integrations. Wiring into Xero, Stripe, a booking system, a warehouse export, Mailchimp, a carrier API. Each connection is real work: authentication, error handling, mapping fields that do not quite line up, and the week you discover their sandbox behaves nothing like production. One integration might be a day. Three integrations with messy legacy data might be a month.
Data migration. Getting the mess out of the old spreadsheets, the Access database, or the previous developer's export and into the new thing cleanly. This is often the slow part. Owners assume the new system is the hard bit. Frequently the hard bit is untangling what they already have.
Who builds it. One person who owns it end to end versus an agency with account managers, project managers, designers, and developers in different time zones. Layers cost money. Sometimes they are worth it. Often, for a business with ten to fifty staff, they are not.
Whether it is a one-off or something that keeps growing. A tool that replaces one painful spreadsheet is a different conversation to a platform you plan to extend for five years. The second one needs more thought upfront on structure, permissions, and how you will change things without breaking production.
Rough ranges, by type of build
These are bands from my own work, not industry averages pulled off a blog. Your project might sit outside them. That is fine. The point is to give you something concrete to react to.
A small internal tool that replaces one painful spreadsheet. Think: a single team, one workflow, maybe one export to accounting. Often $4,000 to $9,000 + GST. Maybe fifteen to forty-five hours of build, depending on how tangled the spreadsheet logic actually is.
A proper internal system. Dashboards, reports, approvals, multiple user roles, the kind of thing that runs a part of the business day to day. Often $11,000 to $34,000 + GST. This is where Glovebox lives: multi-tenant rental ops, fleet, bookings, bonds, operator dashboards. Not a weekend project.
A customer-facing web app or platform. Accounts, payments, notifications, the public-facing side as well as admin. Often $19,000 to $60,000+ + GST for a first version you would actually put customers on. Freelegs is a marketplace in this territory: listings, claims, driver checks, payments, replacing Facebook groups and phone tag.
An integration job on its own. You already have a system. You need Xero talking to it properly, or Stripe wired in, or a warehouse feed that stops someone copying CSV files by hand. Often $1,500 to $8,000 + GST, unless the data on either side is a disaster.
A real example on the smaller end. I scoped a staged web platform for a local workshop at around 65 hours and $4,900 + GST for the first slice: marketing site, shop, and a v1 of the workshop tuning file portal on one Laravel app. That is not the finished forever version of the business. It is a sensible first cut you can use and grow. Hourly, that works out around $75/hr on that quote. Ad-hoc work outside a scoped project is billed differently (currently $115 + GST/hr on my client terms).
Honest ranges beat a single scary number every time.
Fixed price versus hourly
Owners want certainty. I want to build something that actually fits the problem. Those two wants pull in different directions.
Fixed price works when the scope is tight enough to write down without hand-waving. I can tell you what is in, what is out, and what happens if you change your mind mid-build. I reach for fixed price when we have done the thinking together and the boundaries are clear. If the scope shifts, we talk before the budget shifts.
Hourly works when the problem is exploratory, or when you know the first version is really a discovery phase. Spike the risky integration first. See what the data looks like. Then decide whether to keep going. Hourly needs trust on both sides. You are not signing a blank cheque. I am not padding hours because I under-quoted.
On most projects I quote a fixed price for a defined slice, show working software weekly, and treat changes as a conversation before they become a surprise invoice. That is the model described on the web apps & platforms page, and it is how I prefer to work when we can get the scope honest enough.
What you are actually paying for
Not lines of code. You are paying for someone who has built this before spotting the expensive mistake before you commit to it. The decision to model bookings one way versus another. The call to use Stripe the way Stripe expects, not the way a tutorial from 2019 suggested. The boring work of backups, logging, and deployment that keeps you out of trouble at 9pm on a Sunday.
The other big one: you own it. Code, data, hosting. Yours. Not seats rented on someone else's SaaS forever, with your business logic trapped inside their roadmap. I will happily tell you when Shopify or WordPress or an off-the-shelf tool is the right call. When custom is right, you should end up with something you control.
The costs people forget
Hosting. A modest Laravel app on managed hosting might be $25 to $115/month, depending on traffic and redundancy. Not free, not frightening.
Maintenance. Security updates, dependency bumps, the odd server patch. Small retainer or reactive hours. If nobody looks after it, it slowly rots.
The changes you will want in six months. Businesses move. New reporting line. New integration. A workflow that made sense in March and does not in September. Budget for evolution, not just launch day.
I would rather tell you this now than have you feel ambushed later.
When you should not build custom
Sometimes the right answer is not custom software.
If Shopify or WordPress does the job, use it. If a booking SaaS covers ninety percent of what you need at a sensible monthly fee, start there. If the problem is process, not software, no amount of Laravel will fix it.
I have turned work down when the maths did not stack up. I have recommended off-the-shelf tools when that was genuinely better for the client. That is not noble. It is practical. You trust the rest of what I say more when I am willing to tell you not to hire me.
This article completes a trio with when to replace a spreadsheet with custom software (whether to do it) and build versus buy: when an off-the-shelf booking system stops being enough (whether to rent or build the booking tool).
How I quote
First, a free chat. Call, video, or coffee in Christchurch if that suits. I want to hear what you are trying to do, what is in the way, and what "done" looks like.
Then a written scope: what I would build, what it costs, and what it costs if you leave it as-is. Fixed price when that is sensible. Hourly when we need to explore first. Either way you know upfront before we start.
If that sounds useful, get in touch. No hard sell. Just a straight conversation about whether I can help, and what it would roughly look like if I can.