Using our data
Licence, attribution, and what the numbers will and will not support.
DPA reports are published so other people can use them. This page says how, under what licence, and where the limits sit.
Licence
Report content, figures, and written analysis are licensed under Creative Commons Attribution 4.0 International. You may copy, redistribute, adapt, and build on the material for any purpose, including commercially, provided you give attribution.
The site’s source code is separately licensed under MIT. Both licence files sit at the root of the designpayasia/site repository.
Attribution
Cite the report, the year, and the date you retrieved it:
Design Pay Asia (2023). Design Pay Asia Report 2023. Retrieved 25 July 2026 from https://designpay.asia/reports/2023
If you are quoting a specific figure rather than the report as a whole, cite the section you took it from. Section routes are stable and year-namespaced, so /reports/2023/compensation will keep resolving to the same content.
Where the methodology lives
Each report carries its own methodology, because survey design, distribution channels, and sample composition change between years. It sits on the report’s landing page, so read the one for the year you are citing rather than a general account:
Every published figure is also tied to an evidence record naming its source, sample size, geography, and collection date. If a number matters to your argument, the evidence record is where its provenance lives.
What the numbers support
The data is self-reported, voluntary, and collected through design community channels. That shapes what you can fairly conclude from it.
You can fairly describe what respondents in a given market and career level reported. You can compare segments within one year’s data where both cohorts clear the size threshold. Quote the sample size alongside the figure and the reader can judge it for themselves.
Three things the data will not carry. It cannot stand in for a market’s whole design workforce, because respondents reached us through design community channels rather than a representative sample. It cannot support a year-on-year trend claim where the distribution or methodology changed between surveys. And it says nothing about a segment whose figures were suppressed. Suppression means the cohort was too small to publish safely, not that the result was zero or unremarkable, so do not read either into the gap.
Geographic skew is the constraint that catches most people. In 2023, Singapore accounted for 69.4% of responses, so aggregate regional statistics lean heavily Singaporean. Country breakdowns are given where the cohort allows it.
If a number looks wrong
Tell us. Raise a GitHub issue on the public repository or use /contribute. Corrections take priority over new content, and the data safety rules describe how a fix lands.