Quick answer: Usually not by itself. One successful run may show that an outcome is possible, but it may not show that the effect is repeatable, attributable to the variable being tested or stable enough to support the next investment decision. Where an outcome cannot be known or determined in advance by a competent professional on the basis of current knowledge, information or experience, repeating a result under appropriately controlled conditions may be part of resolving that uncertainty.
A successful run is evidence, not yet a general rule
Technical teams remember the run that worked.
The prototype finally stayed stable.
The model reached the expected performance.
The material survived the load case.
The process produced the desired output without the failure seen earlier.
At that point, momentum naturally shifts towards the next question:
Can we scale it?
But there is an intermediate question that is easy to skip:
Would we expect the same result again?
A single successful result can be valuable. It can support a hypothesis, rule out certain explanations or show that a proposed mechanism is technically possible.
What it cannot automatically establish is whether the effect is reliable.
The Australian Government's current guidance on core R&D activities describes experimentation as a scientific process designed to test a hypothesis. It specifically notes that experiments should be repeated enough to distinguish a real effect from chance and designed to reduce bias. The same guidance emphasises comparing outcomes and evaluating the relationships between parameters that are varied, held constant, tested or measured. Business.gov.au
That does not mean every commercial experiment requires a large statistical study.
It means the evidence needs to be strong enough for the conclusion the team wants to draw.
If the next decision is inexpensive and reversible, one encouraging result may justify another development step.
If the next decision involves production tooling, a field deployment, a substantial software migration or a material capital commitment, the evidentiary standard may reasonably be higher.
The important question is not:
Did it work once?
It is:
What does this result allow us to conclude with confidence?
Repeat the uncertainty, not just the procedure
Repeating an experiment only adds value if it tests the part of the result that remains uncertain.
Running exactly the same procedure again can sometimes be useful.
But in many projects, the real question is whether the result survives normal variation.
Suppose a new manufacturing method produces an acceptable component on the first successful run.
The next uncertainty may not be whether the operator can reproduce the same setup perfectly.
It may be whether the outcome remains stable when raw-material properties vary within the expected range.
For software, a promising result on one dataset may need to be tested against another representative data slice.
For an engineering system, performance may need to be checked across realistic load conditions rather than repeated at one favourable operating point.
For a formulation, the relevant question may be whether the effect persists across normal variation in inputs, storage conditions or processing.
This is where teams should separate several different ideas.
Repeatability asks whether substantially the same conditions produce a consistent result.
Robustness asks whether the result remains stable when relevant conditions vary within the environment the system is expected to encounter.
Scale-up asks whether the technical behaviour persists when volume, geometry, throughput, equipment or operating context changes.
Those are different sources of technical uncertainty and may require different hypotheses or experimental designs.
A repeatable bench result does not prove scale-up.
A successful scale trial does not necessarily prove robustness across production variability.
And a one-off success establishes even less.
Current government guidance makes a similar distinction for scale-up: applying knowledge already established at smaller scale will generally not qualify as core R&D merely because the work is larger, while a scale-up activity may involve a new technical uncertainty where the outcome still cannot be determined in advance and the activity follows the required hypothesis-driven systematic progression. Business.gov.au
The decision to scale should therefore follow the uncertainty.
If the known mechanism is sufficiently established, the next activity may simply be engineering implementation.
If repeatability remains unresolved, scaling may multiply the uncertainty rather than resolve it.
Decide in advance what another run needs to prove
Teams often repeat tests reactively.
A result looks promising, so they run it again.
If the second result is different, they run a third.
Soon the project has several runs but no clear rule for what the sequence means.
A stronger approach is to decide before the repeat what uncertainty the additional evidence is meant to reduce.
The team might ask:
• Are we checking whether the first result was anomalous?
• Are we testing sensitivity to a specific input?
• Are we checking whether operator, batch, dataset or environmental variation changes the outcome?
• Are we trying to estimate an operating range?
• What observation would make us less confident in the original hypothesis?
• What result would justify moving to a larger or more expensive stage?
The repeat should be designed to inform a decision. That decision does not have to be binary.
That decision does not have to be binary.
The result might show that the mechanism works but only inside a narrower operating window than expected.
It might show that the average performance is acceptable but variability is too high.
It might reveal that the original success depended on an uncontrolled condition.
Each of those findings can be more valuable than simply obtaining another positive result.
This is also why contemporaneous records matter.
Australian Government record-keeping guidance states that records should be created at the time the activity is conducted. Examples of records that may demonstrate the systematic progression include experimental plans, records of trial runs and test reports showing results and evaluation. Business.gov.au
For repeated experiments, the record should make it possible to see what stayed the same and what changed.
If the first run used one material batch and the second used another, that difference may matter.
If the software version changed between tests, that matters.
If the measurement method was recalibrated, that matters.
Without that context, several similar-looking results may not actually be comparable.
Make repeatability a business decision gate
Repeatability is not only a scientific concern.
It is an investment-control tool.
A founder deciding whether to commit to production, a CTO deciding whether to integrate a new architecture and a CFO deciding whether to fund another development stage all face the same underlying question:
How much uncertainty are we carrying into the next commitment?
A useful go/no-go gate can therefore distinguish between three outcomes.
The evidence may be strong enough to proceed.
The result may be promising but still require another bounded investigation.
Or the variation may be large enough that scaling would increase exposure before the technical mechanism is understood.
The R&DTI Manufacturing sector guide provides a practical example in which experimental protocols, experimental results, manufacturing reports, run sheets and quality-control records were used to document experimental work conducted in a production environment. Business.gov.au
That same discipline is useful even where no R&DTI question arises.
The point is not to keep experimenting indefinitely.
It is to avoid making a large commercial decision on evidence that only supports a much narrower technical conclusion.
One successful test may be enough to say:
This can happen.
Repeatable evidence may allow the stronger statement:
This happens consistently under the conditions we have tested.
Evidence across the relevant operating conditions may then support a further conclusion:
We understand the conditions under which it happens well enough to make the next decision.
That can provide a more useful technical basis for deciding whether the project is ready to move towards implementation or whether further experimental work is still justified.
Conclusion
A successful experiment can show that an outcome is possible without establishing that it is repeatable, robust or ready for scale-up.
Further experimentation should therefore be driven by the uncertainty that remains — whether that concerns repeatability, sensitivity to relevant variation, operating limits or changes introduced by scale.
The objective is not to repeat a successful test until the team feels confident. It is to gather the evidence needed for the next technical and investment decision.
Sources & Further Reading
Australian Government — Conducting core R&D activities for the R&D Tax Incentive
Australian Government — Record keeping for the R&D Tax Incentive
Australian Government — Manufacturing sector guide for the R&D Tax Incentive
Federal Register of Legislation — Income Tax Assessment Act 1997, Division 355
Australian Government — Get help from a research service provider
Talk to Ignition Research if your team has one promising technical result but is unsure whether it is strong enough to justify scale-up, deployment or another significant investment step. As a Registered Research Service Provider, Ignition Research assists businesses with research question definition, experimental design, technical investigations and contemporaneous supporting records within its registered research fields. Business.gov.au
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.
R&DTI note: Repeating an experiment does not, by itself, extend or create a core R&D activity. Further experimental work still needs to address an outcome that cannot be known or determined in advance and satisfy the applicable core R&D requirements.
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.
