Quick answer: A failed experiment does not automatically mean the research should stop. The more useful question is whether the result has reduced the technical uncertainty enough to justify a revised hypothesis, a redesigned experiment or a stop decision. Where the remaining outcome still cannot be known or determined in advance by a competent professional on the basis of current knowledge, information or experience, a further experimental step may still be appropriate.
9 September 2026 — this article describes the current rules. The 2026–27 Federal Budget announced R&DTI reforms for income years starting on or after 1 July 2028. Until those changes take effect, the R&DTI continues to be administered under the current legislation.
The first experiment did not work. The target was missed, the preferred mechanism was not supported, or the result was too unstable to rely on. The team now faces a harder question than the original technical problem: should it run another experiment, change direction, or stop investing?
That decision is easy to blur with ordinary project momentum. Engineers can usually think of another configuration to try. Founders may feel that too much has already been spent to stop. A technical team may also have learned enough from the failed run to see a better approach. None of those reasons, by themselves, tells you whether another research cycle is justified.
A more useful approach is to ask what the failed experiment actually changed about the uncertainty.
A Failed Result Is Not Yet a Stop Signal
For research purposes, a failed result is most useful when the experiment was capable of distinguishing between competing explanations. If the result simply says that a prototype did not work, the next step may still be unclear. If the experiment tested a defined hypothesis under known conditions, the failure can narrow the problem.
For the R&D Tax Incentive, business.gov.au describes a systematic progression of work as proceeding from hypothesis to experiment, observation and evaluation, and leading to logical conclusions. Under section 355-25 of the Income Tax Assessment Act 1997, a core R&D activity must also involve an outcome that could not be known or determined in advance on the basis of current knowledge, information or experience, be based on principles of established science, and be conducted for the purpose of generating new knowledge. The framework does not require the preferred hypothesis to succeed. What matters is whether the activity satisfies those requirements and produces a reasoned conclusion from the evidence.
After an unsuccessful trial, the first question should therefore be: what did this result rule out?
• Ruled out a proposed mechanism: If it ruled out a proposed mechanism under the conditions tested, the research question may remain open but the next hypothesis should change.
• Exposed a test method flaw: If the result exposed a flaw in the test method, the experiment may need redesign rather than a new technical solution.
• Exhausted credible alternatives: If the same limitation now appears across credible alternative approaches, the evidence may be moving towards a stop decision.
• Uncontrolled conditions: If the result was uninterpretable because conditions, versions or measurements were uncontrolled, repeating the activity without redesign may simply produce another ambiguous result.
The important distinction is between learning that changes the next decision and activity that merely creates another iteration.
Define What the Next Experiment Must Tell You
Before approving another research cycle, write down the decision it is supposed to enable. This keeps the experiment tied to the remaining uncertainty rather than to the desire to keep the project moving.
Suppose a team is investigating whether a new process can remain stable under a difficult operating condition. The first experiment fails. There are several possible explanations: the proposed mechanism is wrong, one uncontrolled variable caused the instability, the measurement method is inadequate, or the required operating envelope may simply be unattainable with the current approach.
The next experiment should discriminate between those explanations. It should not merely repeat the build with several settings changed at once.
A useful pre-experiment note can state:
• What the previous trial established and what remains unknown
• Which explanation the new hypothesis is testing
• What will be held constant and what will change
• What observation would support or weaken the hypothesis
• What result would cause the team to continue, redesign or stop this line of investigation
This is also where contemporaneous records matter. Government guidance on R&DTI record keeping says records should be created while activities are being conducted and should capture what was done, how the work was performed, when it occurred and who carried it out. For core activities, the records should also evidence the systematic progression of work from hypothesis through experiment, observation, evaluation and conclusion.
For a changing research programme, that means preserving why the next test followed from the previous result rather than reconstructing the sequence retrospectively after the project is complete.
The record does not need to make every early assumption look correct. In fact, a research history in which hypotheses are rejected, refined and replaced may describe the investigation more accurately than a retrospective narrative in which the final approach appears inevitable from the beginning.
Separate a New Hypothesis from a New Project Objective
One of the easiest ways for research to drift is for the technical question to change without anyone noticing.
The original question might have been whether a defined technical performance could be achieved using a proposed mechanism. After several failures, the team may quietly move to a different objective: reduce the cost, make the interface easier to use, integrate a different supplier, or deliver enough functionality for a customer demonstration.
Those can be sensible business decisions, but they are not automatically continuations of the same technical investigation. The team should identify whether a genuine technical uncertainty remains and whether another hypothesis-driven experiment is needed to resolve it.
Australian Government guidance recognises that R&D plans may change as experimental results are evaluated. A change in plan does not, however, mean that every subsequent activity remains part of the same core R&D activity. The company still needs to identify the actual activities conducted and show how they relate to the technical question, the systematic progression of work and the conclusions reached.
Management decision gate:
If the remaining work can now be resolved using current knowledge, information or experience available to a competent professional, implementation may be the appropriate next pathway.
If a specific technical outcome still cannot be known or determined in advance and resolving it requires another hypothesis-driven experiment within a systematic progression of work, further research may be warranted. That does not, by itself, establish R&DTI eligibility; each activity must still satisfy the applicable statutory requirements.
Conclusion
A failed experiment should not automatically trigger either abandonment or another round of testing. Its value lies in how much uncertainty it removes.
The next research cycle should have a narrower question than the one before it, a hypothesis that reflects what has already been learned, and an explicit result that could cause the team to change course. If the team cannot explain what the next experiment would decide, it may be scheduling another iteration rather than designing a research experiment.
Where the R&D Tax Incentive may also be relevant, eligibility remains self-assessed against the program's requirements. The technical record can support that assessment, but it does not determine the outcome. Current R&DTI requirements should be checked against guidance from the Department of Industry, Science and Resources and the ATO, with appropriate tax advice obtained from a registered tax agent where required.
Sources & Further Reading
Talk to Ignition Research if an uncertain technical project has reached a point where an experiment has failed and the next research step is unclear. As a Registered Research Service Provider, Ignition Research assists businesses with research question definition, hypothesis refinement, experimental design, technical investigations and contemporaneous supporting records within its registered research fields.
RSP registration confirms that the provider has met the relevant registration requirements for its registered research fields. It does not constitute government endorsement of a project or guarantee R&DTI eligibility.
We do not determine R&DTI eligibility or provide tax advice. The R&D Tax Incentive is a self-assessment program, and companies remain responsible for assessing their activities, expenditure and circumstances and obtaining appropriate tax advice.
This article provides general information only and is not tax, legal or financial advice.
Working on a project like this?
Tell us what you're building and the part you're not sure will work. A researcher from our Adelaide team (registered Research Service Provider RSP000047) reads every enquiry and replies within two business days. No documents needed.
