The scientific experiment in the R&D Tax Incentive

What is an R&D Tax Incentive experiment? One significant reservation that potential applicants have with the R&D Tax Incentive is the need for the claimed R&D to be based on the principles of scientific experimentation.

Many people who read that are put off because they think that their R&D may not measure up to the sort of rigour one normally associates with scientific experiments.

Forget about R&D stereotypes such as scientists in white lab coats in a laboratory. AusIndustry is not expecting an academic or scientific level of rigour in the experiments that applicants conduct as part of their claimed R&D. The key phrase is ‘principles of scientific experimentation’. AusIndustry just expects you to follow the general principles of scientific experimentation.

Refer to the ‘refreshed’ AusIndustry guide for more information on how AusIndustry view the principles of scientific experimentation.

So, what is an experiment in the R&D Tax Incentive? Well, this article explores the principles that AusIndustry is expecting you to follow.

A practical guide to experimentation in the R&D Tax Incentive

What AusIndustry is getting at is that you have conducted your experiments in a planned and systematic manner. What does that mean?

It means that:

  • You have a theory or hypothesis that you want to confirm, relating to a knowledge gap.
  • You design a (controlled) experiment that is capable of proving or disproving your hypothesis.
  • You conduct that experiment according to a plan you wrote down before doing the experiment.
  • You use the results from the experiment to prove or disprove your hypothesis.

The important thing is that you have a theory/proposition/hypothesis that can be confirmed (or not) by the experiment. An experiment which simply aims to collect data which it is hoped may tell you something about your problem, is not a valid experiment in the eyes of AusIndustry. An experiment without a plan is also not a valid experiment, even if you did learn something.

The plan doesn’t need to be complicated or extensive – you just need to demonstrate that you thought about the experiment before you performed it. The word ‘experiment’ also causes angst. Just think of ‘testing’, prototyping, evaluation where you weren’t confident in the outcome.

What is a hypothesis?

Anyone who has investigated the R&D Tax Incentive to a reasonable degree will have come up with a conundrum. If AusIndustry expects you to have a technical gap – a problem that you can’t solve, how are you supposed to come up with a hypothesis that you can prove in an experiment?

Well, how would you normally tackle your problem? You do enough investigation to come up with an idea or at least a partial solution, often one with high technical risk. You may collect some data and analyse it, looking for a pattern.

AusIndustry recently released a ‘Refreshed Guide to Interpretation‘ of the R&D Tax Incentive. The main reason for the update was an attempt  by AusIndustry top make the guide more readable, but there were some actual content changes, and one of those related to how a hypothesis was defined:

an idea or proposed explanation for how you could achieve a particular result and why that result may or may not be achievable“.

The “may or may not be achievable” clause relates to the technical risk inherent in the experiment. This technical risk is related to the knowledge gap you must have in order to have eligible R&D.

This is what separates an experiment in the R&DTI from routine testing. If you had confidence in the result of an ‘experiment’, then there was no technical risk and no knowledge gap. It wasn’t an experiment – it was a test you did to confirm a known result. Sorry, but if the test was not successful because of a ‘bug’ or mistake in your experimental setup – that doesn’t make it a valid experiment – you just made a mistake. There was no knowledge gap.

This is why AusIndustry focuses so much on establishing that there was a knowledge gap – you need to prove that you did not know the answer up-front. You can only do this if you can show the searching that you did in the public domain, without finding the knowledge you needed. See Prior art searching for R&D Tax Incentive.

Regardless, a hypothesis must be specific and measurable. AusIndustry is not actually interested in your hypothesis – just that you have one, and that it is capable of being proven or disproven by your experiment. In fact, the actual hypothesis is not often stated in the AusIndustry R&D Tax Incentive application., as they can be highly technical in nature.

However, you do need to write down the object of your experiment in your supporting documentation as AusIndustry may ask to see it in a subsequent review. You may not call it a ‘hypothesis’ but the object of the experiment must be documented, it must be specific and it must be testable.

The hypothesis doesn’t need to be a complete answer to your technical gap. It is common for a technical gap to require a series of experiments – each experiment has a hypothesis for part of the overall solution. With each experiment, you build up your level of knowledge.

These hypotheses may come from your own experience, or you may find partial solutions on the Internet. This is partly what AusIndustry refers to as a ‘systematic progression’ of work.

The hypothesis is typically expressed in terms of the outcome of the experiment – I hypothesize that my new algorithm will achieve an accuracy of ‘X’, or achieve a response time of ‘Y’. Achieving that result may require a series of experiments where the configuration or input variables of the experiment are adjusted to achieve the desired result.

Failures in Experiments

Unsuccessful experiments are not bad things. In fact, failed experiments can help show that the technical gap was difficult to address – if the failure was due to missing knowledge. As noted above, debugging is not experimentation! The exception is where you have a design flaw in your system or process due to a knowledge gap. AusIndustry does expect you to document what you learnt from the failed experiments – how did this learning help frame future experiments (which may have been successful).

Even if the series of experiments was never successful, but you gained new knowledge from those experiments and documented that knowledge – that could still be eligible R&D.

Documenting your experiments

The documentation level only needs to be sufficient to back up what is claimed in the R&D Tax Incentive application. However, the supporting documentation (by definition) will be written before the R&D Tax Incentive application is drafted, so you do need to discuss what you plan to claim with your R&DTI consultant during your R&D so that you will have enough documentation to cover the claims that will be made in the later R&D Tax Incentive application.

The level of documentation expected also depends upon the company’s size (and the size of the R&D claim). Small companies are not expected to have detailed documentation although you still need to show that the claimed R&D did actually take place