ntwrk/
Scope the 30-day POC
ANSWERS

How to choose an Australian data consultancy

The questions that separate one Australian data consultancy from another are rarely about method. Almost every firm will describe a similar approach, and most of them can do the work. What differs is what you are left holding when they leave. Ask who owns the code, who can run it without the firm in the room, what the handover actually contains, and what happens when the first approach turns out to be wrong.

Most people reading this have bought something before that did not land. A platform nobody uses, a model nobody trusts, a dashboard that went stale in a month. That is common enough to be unremarkable, and it is usually not a failure of judgement. The pitch is designed to be persuasive and the gaps only show up later, by which point the people who made the promises have moved to another account.

These five questions surface the gaps early. None of them is clever. They work because the answers are hard to fake under follow-up.

Five questions to ask a data consultancy, and the answer that should worry you Five numbered questions listed down a vertical spine, each paired with the answer that should worry you. Who will actually do the work, paired with talk of bench strength and blended teams. When the first approach is wrong, paired with the claim it has not come up. What you own when it ends, paired with hedging on whether the work could continue without them. Where is our plan weak, paired with agreement with all of it. A reference who stopped, paired with another happy client offered instead. ASK WORRY IF YOU HEAR 1 Who will actually do the work bench strength, blended teams 2 When the first approach is wrong "that has not come up" 3 What you own when it ends hedging on the switch question 4 Where is our plan weak agreement with all of it 5 A reference who stopped another happy client
None of the five is clever. They work because the answers are hard to fake once you follow up.

Who will actually do the work

Ask for the names of the people who will be on your project, and ask how many days a week each of them will be on it. Then ask which of the people in the room today will still be involved in month three.

The gap between the pitch team and the delivery team is the most common disappointment in professional services, and it is rarely malicious. Good people get pulled onto the next pitch because they are good. A firm that answers this cleanly has thought about it. A firm that talks about bench strength and blended teams is telling you the answer without saying it.

What happens when the first approach is wrong

Every data project meets something unexpected in the data. Duplicate customers, a field that means two different things in two systems, a source that stopped updating in 2023 and nobody noticed. Ask what happened the last time a firm hit one of these, and who paid to fix it.

A firm that says this has not come up is either very new or not being straight with you. What you want is a specific story with an uncomfortable middle. How they handled being wrong last time is the best available guide to how they will handle it with you.

What you own when it ends

There is a difference between owning the output and owning the capability. Owning the output means you have the dashboard. Owning the capability means the code sits in your repository, the model runs in your environment, the documentation is good enough for someone new to follow, and the data stays in a warehouse you control.

The test is simple. If you changed firms next quarter, could the work continue. Ask it as a direct question and listen for hedging. Some arrangements are perfectly reasonable without full ownership, and renting a platform is often the right commercial call. The thing to avoid is discovering the answer at renewal, when you have least room to negotiate.

Can they talk you out of something

Ask where they think your plan is weak, and ask what they would decline to build. A firm that agrees with all of it is selling to you. A firm that has read your brief closely enough to disagree with part of it has done more thinking than the one that has not.

This one also tells you what the next two years will feel like. The relationships that work are the ones where somebody can say the thing you do not want to hear, early, when it is still cheap.

Ask for a reference who stopped

Any firm can produce a happy reference. Ask instead for a client who ended an engagement early or chose not to renew, and ask to speak to them.

Some firms will refuse, which is fair enough. A few will not have one. How a firm describes a relationship that ended tells you more than how it describes one that worked. Defensiveness, blame directed at the client, or a story where nothing was learned are all worth more than another glowing call.

Weighting the answers

Not all five carry the same weight for every buyer. If you are building something you intend to run yourself for years, ownership and handover matter most. If you need a result before a board meeting in eight weeks, staffing and honesty about failure matter more. Decide which two you actually care about before the first meeting, because it is much harder to weigh them fairly once you like the people.

Where this comes from. These are the questions Australian buyers ask in procurement, and the ones they say afterwards they wished they had asked. Published by NTWRK, an Australian data and AI consultancy that would answer all five of these itself. Worth putting the same questions to another firm before taking any of it as neutral. They work regardless of who you hire.