
How to evaluate a software development company in Sydney: a buyer's checklist
Most shortlists for a software development company in Sydney start on a directory. SelectedFirms lets you filter Sydney agencies by hourly rate, tech stack, industry, reviews and services. Goodfirms lists 172 Sydney software companies it has reviewed, with an hourly rate and team size for each. Those filters help you cut a long list down to a short one. They won’t tell you how a firm scopes work, or what happens the first time production breaks on a Friday afternoon.
We have built custom software in Sydney since 2001. Most of it is business systems and integrations that clients run for years, and we keep supporting them long after launch (PicNet’s provider comparison). This checklist covers the questions we think a buyer should put to any vendor, us included. It is part of our Buying IT Services series for IT managers and executives choosing suppliers.
Treat rankings as a starting point
Directories sell placement. SelectedFirms says that certain agencies may be featured through sponsored placements, and Goodfirms labels listings from agencies that pay for premium placement as “Sponsored”. Rankings written by vendors are marketing. The 2026 ranking published by Apptunix puts Apptunix second, just behind Atlassian, which it describes as a builder of platforms that other developers use.
A Sydney address on a listing tells you where to reach the firm, so ask separately where its engineers sit. Our comparison of Sydney providers records this for each firm. The answers range from teams based entirely in Sydney to firms with large delivery teams in Nepal or India (PicNet). Either model can work, but you should know which one you are buying before you sign.
1. Ask how discovery and scoping work
Before anyone writes code, a competent firm wants to understand the problem: who uses the system, what it connects to, what data it holds and what “done” looks like. Ask what discovery costs and what you receive when it ends.
A good discovery output is a document you could hand to a different vendor. It should contain:
- a written scope that says what is in and what is out
- the systems you need to integrate with, and who owns each one
- an architecture sketch covering hosting and where your data is stored
- the risks and assumptions, written down
- an estimate given as a range, with the reasoning behind it
A firm that gives you a fixed price for a large system after one meeting has either padded the number or not yet understood the job.
2. Ask what happens when requirements change
On any project of real size, the requirements will change. Users see the first working screens and realise the process they described is different from the one they follow. Ask the firm how it deals with that.
A fixed-scope, fixed-price contract suits a small job you understand well. On larger work, the vendor protects its margin with change requests, and your meetings turn into arguments about whether a feature was in scope. Iterative delivery gives you a working release every few weeks and a backlog you can reorder, so you can change priorities without renegotiating the contract. It also needs effort from your side, because someone has to review each release and make decisions quickly.
Plenty of buyers settle on a mix: fixed-price discovery, then time-and-materials delivery against a budget cap that is reviewed every sprint. Ask the vendor to walk you through an anonymised change request from a past project and explain how it was priced.
3. Ask who owns the code and how handover works
Ownership follows the contract. Check that the agreement assigns the intellectual property in the code written for you once you pay. The source should also sit from day one in a repository you control, such as your own GitHub or Azure DevOps organisation, with the vendor added as users. If the contract says nothing about IP, you are relying on legal defaults you probably haven’t checked.
Ask what the firm reuses from earlier work. Most firms bring their own libraries, so you need a licence to use them that survives the end of the relationship. Open-source components carry their own licences, and a careful vendor will list them for you.
Ask about AI coding tools as well. Many firms now use them, and we use them where the client allows it. You should know which tools touch your code and whether your code or data leaves your environment.
Then ask what handover looks like if you later move the work in-house or to another vendor. A good answer covers:
- infrastructure defined as code that someone else can deploy
- a README that gets a new engineer running locally in a day
- documentation for each integration, plus a record of who holds which credentials
- an agreed period during which the original team answers questions
If the vendor can’t describe handover clearly, assume you will be locked in.
4. Ask about testing, deployment and support after go-live
Business systems run for years, so the support arrangement matters as much as the build. Ask how the firm tests. Automated unit and integration tests that run on every change are a different standard from a manual click-through the day before release. A pipeline from a current project will show you more than a slide will. Deployments should be automated and repeatable, and there should be a tested way to roll back.
For support, find out who answers when production fails out of hours, and whether that person helped build the system. One vendor-written guide suggests looking for at least eight weeks of post-launch maintenance. For a system you expect to run for years, eight weeks is a warranty period. Ask what support looks like in year three and who does the framework and security upgrades.
5. Ask to meet the engineers
Ask to talk to the people who will write your code before you sign. An account manager can explain the process. An engineer can tell you how they would handle your hardest integration and what worries them about the project, and their answers will show you more about the firm than a capability deck.
Check continuity too. Find out whether the people in the pitch will work on the project, and what happens if one of them leaves halfway through. At PicNet every manager is an engineer, so the people in a first meeting with us are engineers. A firm that won’t put an engineer in the room before you sign is showing you who you will be dealing with afterwards.
Questions for the first meeting
- What does discovery cost, and what document do we get at the end?
- How do you price and approve a change in scope?
- Will the code live in our repository from day one, and does the contract assign the IP to us?
- Which AI coding tools do you use, and does our code leave our environment?
- Where do your engineers sit, and which of them will work on our project?
- How do you test and deploy, and can we see a pipeline?
- Who supports the system after go-live, and what will that cost in year three?
Red flags to walk away from
- A fixed quote for a large system after a single meeting.
- The code stays in the vendor’s repository until final payment, or the contract says nothing about IP.
- No engineer is available to talk to before the contract is signed.
- Testing comes down to “our QA team checks it before release”, with no automated tests.
- Support is only a warranty period, with no plan for the years after it.
- Every reference is a launch, and none comes from a client whose system the firm still supports. As our comparison page puts it, a long support record says more than a portfolio of launches.
PicNet builds custom software and production AI systems for Australian organisations. If you are working through a shortlist, talk to us about what a first project could look like.