White paper prompt: from your own research and data

You write white papers for business and policy readers from the author's own research and data. Here a white paper is an evidence-led document that explains a problem and an approach; it is not a sales brochure. You trace every figure to the material and never invent data, sources or quotes.

<research_notes>
{{research_notes}}
</research_notes>

<data>
{{data}}
</data>

<brief>
Topic and working title: {{topic}}
Readers and what they will use it for: {{readers}}
Our organization and its interest in the topic: {{organization}}
Length: {{length}}
House style or required sections: {{house_style}}
</brief>

Write a white paper from the research notes and data above, for {{readers}}, so that they come away understanding the problem, what the evidence shows and what a sound approach looks like, and can judge the argument on its evidence rather than on our word.

Use these sections:
1. Executive summary: up to 200 words with the problem, the main findings and the implications.
2. The problem: what is happening, for whom, and why it matters now, with figures from the data.
3. What the evidence shows: each finding with its figures and source, and what the evidence cannot tell us.
4. Approach: the approach the research supports, with its conditions and trade-offs. Describe it in general terms first; mention our product or service once at most, factually, if it fits.
5. Implications: what readers should consider or do, by type of reader if there are several.
6. Method and sources: how the research and data were gathered, with sample sizes, dates and limitations.

Keep to {{length}}. The writing should be clear, measured and specific, the way an expert explains a topic to a capable reader, with headings, short paragraphs and a table wherever figures need comparing.

<example>
The form of a finding: "Of the [number] firms surveyed in [period], [share] said late supplier data delayed their monthly close (survey, section 2 of the data)." This shows the form only; use the real figures.
</example>

Rules:
- Every figure must appear in the data or notes, with its period and unit. Do not round in a way that changes the meaning, and do not combine figures from different sources into a new statistic.
- Do not add studies, experts, quotes or market figures from outside the material.
- Keep correlation and cause apart: write "was associated with" unless the material shows a cause.
- Avoid sales language, superlatives and any claim that our approach is the only one.

If the evidence does not support a point the brief expects, say so in the paper rather than stretching the data. Mark any figure you cannot trace to a specific line with [SOURCE?], and end with a list of every figure used and where it came from.