Build the roadmap around five layers: the question, the evidence you need, the method that produces it, the validation that tests it and the contribution that becomes a chapter, paper or decision. Add thesis milestones after the reasoning path is clear.
Start with the question
Write one sentence naming the system, comparison and result you will measure. A broad topic is not a roadmap anchor. A measurable question gives each later stage a job.
Map evidence before dates
For every claim, name the observation that could support it. Then list the sample, dataset, instrument or corpus that produces that observation. This prevents a detailed calendar from hiding an inaccessible data source.
State the question and the gap.
Name the evidence needed for each aim.
Assign a method, baseline or control.
Define what would count as support or failure.
Map the result to a chapter, paper or decision.
Turn the roadmap into milestones
Use milestones for decisions, not routine activity. “Complete experiments” is weak. “Validate the measurement protocol against the baseline sample” marks a decision that changes the next stage.
| Layer | Planning question |
|---|---|
| Question | What must the thesis establish? |
| Evidence | What makes the claim testable? |
| Method | What turns input into a result? |
| Validation | What could show the claim is wrong? |
| Contribution | Where will the result live? |
Keep it readable
- Label arrows with the evidence moving between stages.
- Show one contingency branch for a failed aim.
- Use the same wording in the roadmap, proposal and thesis outline.
- End with a contribution, not “write thesis”.
Related resources
See the Technical Roadmap Pack, the Research Architecture Pack and how to draw a research roadmap.