Quick answer: Yes, operational data can form part of experimental evidence, but collecting production logs does not by itself make ordinary operation an experiment. A research question may arise where an outcome cannot be known or determined in advance by a competent professional on the basis of current knowledge, information or experience, and the live system is deliberately used to test a defined hypothesis. The important distinction is whether the data forms part of a structured hypothesis-led experiment and evaluation, or simply records business as usual.
The 2026–27 Federal Budget announced R&DTI changes that will start from 1 July 2028. Until then, the R&DTI continues to operate under the current rules.
A production dashboard is not automatically an experiment
Modern businesses generate large volumes of technical data without trying to conduct research.
Manufacturing systems retain process histories. Software platforms record latency, errors and resource utilisation. Energy assets produce SCADA data. Food plants capture temperatures, pressures and production records. Connected equipment can generate sensor streams continuously.
Those records can be extremely valuable.
They are not automatically experimental evidence.
Under the current R&DTI rules, a core R&D activity requires more than observation. Section 355-25 of the Income Tax Assessment Act 1997 requires experimental activities whose outcome cannot be known or determined in advance on the basis of current knowledge, information or experience, but can only be determined through a systematic progression of work based on principles of established science, proceeding from hypothesis to experiment, observation and evaluation and leading to logical conclusions. The activities must also be conducted for the purpose of generating new knowledge. Federal Register of Legislation
The distinction becomes clearer with a simple example.
Suppose a manufacturer monitors motor temperature throughout normal production and notices that temperature rises on certain shifts.
That observation may identify a problem.
It does not yet establish an experiment.
Now suppose the engineering team has competing explanations for the behaviour. It hypothesises that a particular control strategy causes instability when the process load changes rapidly. The team deliberately changes that control parameter under defined operating conditions, records the resulting temperature and load behaviour, compares the observations against an appropriate reference condition and evaluates whether the predicted relationship appears.
The same production historian may supply the data.
What changed is the research structure around it.
Australian Government guidance describes observation as measuring and recording information relating to an experiment, and evaluation as analysing those results to understand what occurred and the causal relationships between parameters that were varied, held constant, tested or measured. business.gov.au
The existence of data is therefore not the key question.
The key question is what the data was designed to tell you.
Turn live operation into a bounded technical test
Real operating environments can be useful precisely because they contain conditions that are difficult to reproduce elsewhere.
A software system may behave differently under real concurrency.
A manufacturing process may respond differently to variable feedstock.
An energy system may encounter combinations of demand, weather and equipment state that a simulation does not fully reproduce.
That does not mean the whole operational period becomes R&D.
A better approach is to define a bounded test inside ordinary operation.
Before changing the system, the team should be able to state the technical question, the hypothesis, the intervention and the evidence that would change the next decision.
Consider a production process whose yield becomes unstable at higher throughput.
The business may already know that lower throughput works.
The unresolved question might be whether changing a particular control sequence prevents the instability without introducing a second failure mode.
A structured live trial could specify the relevant operating range, identify what will be changed, define which other conditions need to remain sufficiently comparable, record the measurements that matter and establish how the result will be evaluated.
The surrounding operation can continue as normal; it is the bounded experimental activity that needs to be clearly defined and assessed against the applicable R&DTI requirements.
Current sector guidance provides practical examples of the same distinction. The Australian Government's energy guidance identifies SCADA records, staged trial records, project tracking histories and technical findings as possible evidence of experimental work. Its software guidance similarly identifies logs, dashboards, performance graphs, test plans and architecture records as potential evidence. The agrifood guidance expressly recognises production records, field notes and sensor logs as possible records of hypothesis-led experimental work. business.gov.au
None of those record types proves that R&D occurred.
They become useful when they show the systematic progression of work.
That is an important discipline for a company already collecting large amounts of data. Instead of asking, “Which logs can we use?”, ask:
What experiment generated—or deliberately relied on—these observations?
Preserve the conditions around the data
Operational data has one major weakness: context changes constantly.
A performance graph may show improvement after a software change, but traffic mix also changed.
A manufacturing trial may show lower defect rates, but the material batch was different.
A process modification may appear to reduce energy use, but production volume fell at the same time.
If those factors are not recorded, the data may remain useful operationally while becoming difficult to interpret experimentally.
The research record therefore needs more than the output metric.
It should allow somebody reviewing the project later to reconstruct what system version was operating, what intervention occurred, which important conditions were controlled or merely observed, and what changed during the comparison.
This is especially important when the business already has automated logging.
Automated systems are very good at preserving measurements.
They are often less good at preserving why an engineer changed the system at that moment.
A version-control record might show that an algorithm changed.
A production historian might show what happened afterwards.
Neither necessarily records the hypothesis that connected those events.
Current government record-keeping guidance states that records should be created at the time the activity is conducted. Examples of evidence that may support R&DTI eligibility include background research, experimental plans, records of trial runs and test reports showing results and evaluation. business.gov.au The ATO also emphasises contemporaneous records and identifies as a compliance concern claims that mix eligible R&D with ordinary business activities, or continue to treat activities as R&D after they have transitioned into ordinary operations. Australian Taxation Office
For operational research, a relatively small amount of additional context can make a large difference. As a practical traceability approach, this may include a dated hypothesis, the configuration being tested, the reason the change was introduced, the planned comparison and the evaluation made after the observations were reviewed.
The goal is not to duplicate every machine log in a research report.
It is to preserve the reasoning that makes the relevant section of that log interpretable.
Know when the experiment has become ordinary operation
Live trials also create a boundary problem at the other end of the project.
At some point, the team may have answered the technical question.
The new control method works across the relevant operating range.
The software configuration is stable enough to deploy.
The process window has been characterised sufficiently for routine production.
The business may continue collecting exactly the same data after that point.
That does not mean the research necessarily continues.
The ATO identifies as a compliance concern situations where activities that were being conducted as R&D transition into ordinary business activities, or where registered activities mix eligible R&D with routine operations. Australian Taxation Office
That boundary is worth documenting prospectively.
A technical team can define what evidence would justify leaving the experimental phase and moving into normal operation. Once that decision is made, continued monitoring may become quality control, performance management, maintenance or ordinary engineering rather than further core experimental work. Any separate supporting R&D treatment would need to be assessed on its own facts.
The reverse can also occur.
Routine operation may reveal behaviour that existing knowledge cannot explain. That may give rise to a new technical question and, where the outcome still cannot be determined in advance and hypothesis-led experimentation is required, a new bounded experimental activity.
The data stream can continue throughout.
The purpose of the activity changes.
For founders and CFOs, that distinction helps answer a practical investment question: are we still paying to generate new technical knowledge, or are we now paying to operate and optimise something we understand?
For CTOs and engineering leads, the corresponding question is more technical:
What uncertainty is this next period of data collection actually intended to resolve?
If there is no clear answer, more monitoring may create more data without creating more research knowledge.
Conclusion
Operational data can support experimental research without turning the surrounding business operation into R&D.
What gives the data research meaning is the defined technical uncertainty, hypothesis, intervention, operating context and evaluation that connect the observations to a reasoned conclusion.
When that uncertainty has been resolved, the same data stream may continue to serve the business even though the experimental activity has ended.
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 Taxation Office — Helping you get R&D claims right
Federal Register of Legislation — Income Tax Assessment Act 1997, Division 355
Talk to Ignition Research if your business is already collecting large volumes of production, sensor, software or operational data but it is unclear whether that data is answering a defined technical research question or simply documenting normal operation. 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.
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.
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.
