Do you know the three parts of R&D work?

Many developers only meet R&D as a section in the sprint review email, and never learn how their day-to-day work becomes part of an R&D claim. When work is not recognised as R&D at the time, the evidence is thin or missing when the report is written.

Every piece of R&D work fits into one of three parts.

1. Core activity

The question the team is trying to answer, and how it will know whether it succeeded. See Do you define your core activities for R&D?.

2. Experiments

An experiment is an attempt to do the thing the core activity set out to do. Experiments can be broken down in many ways.

Example core activity: Establish whether AI can identify tasks discussed in video calls.

• Experiment 1: Give the same call transcript to several LLMs and compare each result with the tasks you expected
• Experiment 2: Repeat the test with a non-English call, such as one recorded by an overseas office, and compare the results

Each experiment needs a measurable outcome. See Do you make your R&D experiment outcomes measurable?.

3. Supporting activities

Supporting activities are the work that makes the core activity possible. It is not the experiment itself, but it exists to support it.

Common examples:

• Research - how others run similar experiments, and what your benchmark should be
• Spikes - see Do you know when to use the term SPIKE (Agile) or PoC (proof of concept)?
• Documentation - on how to use the feature you are building or testing, and on the outcomes
• Testing - manual testing and unit tests that help show whether the core activity worked

How the parts fit together

Figure: Each core activity has experiments and supporting activities underneath it

When you create a PBI, link it to the right core activity and say whether it is an experiment or a supporting activity. See Do you link commits, branches, and PRs to a PBI?.