Why \"maximise your claim\" advice is a liability, and what conservative claiming looks like

"We will maximise your R&D claim" sounds like a promise to get you more money. For software companies, it is often a promise to get you more risk. An R&D Tax Incentive claim is not a number you win; it is a position you may have to defend, and a bigger indefensible number is worth less than a smaller defensible one.

Here is why the incentive to over-claim exists, what an inflated claim actually costs you, and what conservative claiming looks like in practice.

This is general information, not tax advice. Your registered tax adviser determines and lodges your actual claim.

Where the pressure to over-claim comes from

Most of it is structural, and it starts with how a lot of providers are paid.

A large part of the R&D advisory market works on a percentage of your refund or a success fee: the bigger your claim, the bigger their cut. That arrangement quietly rewards claiming more, not claiming well. It is also exactly the model the ATO scrutinises: the regulator actively pursues R&D "promoters" who work on contingent fees and push aggressive positions.

Even without an adviser, the pull is there. After a hard year of engineering, it feels natural to treat the whole team's effort as R&D. The work was difficult and genuinely clever, so surely it all counts. But "hard and clever" is not the test. "Technically uncertain, resolved by systematic experiment" is.

What an inflated claim actually costs you

The headline figure on a slide is not money in the bank. It is a claim that has to survive review, and software R&D claims are among the most heavily scrutinised the ATO runs.

When an over-stated claim is reviewed and reduced, you do not just lose the excess. You can face repayment of the offset, plus interest and penalties, a time-consuming audit that pulls your team off product, and a worse position next year because you are now on the radar. The single most flagged error is the whole-of-project claim: treating an entire platform build as R&D when only specific elements were genuinely experimental. The ATO has called this out explicitly.

So the real comparison is not "big claim versus small claim". It is "a defensible claim you keep versus a larger one you might have to give back, with interest". On a risk-adjusted basis, the conservative claim is very often worth more.

What conservative claiming actually looks like

Conservative does not mean timid, and it does not mean leaving genuine R&D on the table. It means claiming what you can defend, and being able to prove it. In practice:

Isolate the experiments, do not blanket the project. Identify the specific activities where the technical outcome was genuinely unknown and you worked systematically to resolve it. Claim those. Leave the routine build, configuration, debugging and maintenance out of the core claim.

Apportion honestly. You can only claim the share of each person's time spent on eligible activities, on a reasonable, documented basis. For a maturing product, that share is usually lower than your first instinct, and that is normal and correct: as a product matures, the proportion of genuine R&D falls even as total engineering spend rises.

Separate technical uncertainty from commercial uncertainty. "Will this sell" and "can we hit the deadline" are not R&D. Whether a technical outcome was achievable is.

Tie every claimed activity to contemporaneous evidence. A hypothesis recorded up front, the experiment, the results, the version-control history. If you cannot point to the evidence, you cannot defend the claim, so it should not be in it.

Actively look for reasons to exclude. The healthiest question to ask of each candidate is not "can we fit this in" but "would this survive someone arguing it is routine". If it would not, take it out.

A test worth running yourself

For any activity you are about to claim, ask:

1. Could a competent professional in our field have known the outcome in advance, without experimenting? If yes, it is not core R&D.
2. Did we have a hypothesis before we started, and can we show it?
3. Is the uncertainty technical, or is it really commercial?
4. Can we link this to real evidence in our code history?
5. Are we claiming the specific experiment, or the whole project around it?

If a candidate stumbles on these, the conservative move is to scope it out. That is not lost money; it is removed risk.

Why we built the tool this way

Our whole approach is built on this principle. We charge a flat fee banded by claim size, never a percentage of your refund and nothing contingent on the outcome, so doing the right thing and removing what cannot be defended never costs us anything. And every claim runs through an adversarial skeptic pass that challenges each candidate and scores it (genuine R&D, weak, or not R&D), whose entire job is to take unsupportable claims out before they reach your package.

We bias to precision over recall, because for an R&D claim the false positive, routine work claimed as R&D, is the dangerous error. The goal is not the biggest number. It is the number your adviser can confidently lodge and you can confidently defend.

A claim that cannot be defended is worse than no claim at all.

This article is general information, not tax advice, and not a determination of eligibility. Speak to a registered tax adviser before lodging an R&D Tax Incentive claim.