{ "id": "latin-square-subject-rotation", "schemaVersion": 2, "title": { "ja": "件名A〜Dを公平に比べる — 顧客層と配信枠の偏りを抑える", "en": "Compare Subject Lines Fairly — Balance Audience Groups and Send Times" }, "summary": { "ja": "顧客層4つと配信枠4つの各組み合わせに件名A〜Dを1つずつ割り当てる。行と列の両方で4案が揃うため、特定の案が特定の層や時刻に偏るのを防ぐ", "en": "Assign subject lines A–D once across every combination of four audience groups and four send windows. Each line appears once in every row and column, preventing audience or timing bias from favoring one variant." }, "capturedAt": "2026-09-25", "categories": [ "presentation" ], "tags": [ "data-viz", "minimal" ], "useCases": [ "comparison", "testing" ], "format": "slide-16x9", "density": "medium", "layout": { "kind": "matrix", "canvas": { "w": 1920, "h": 1080 }, "margin": { "top": 82, "right": 120, "bottom": 82, "left": 120 } }, "type": { "title": { "size": 44, "weight": 700, "tracking": "0.01em" }, "subtitle": { "size": 22, "weight": 400 }, "axisNote": { "size": 20, "weight": 500 }, "corner": { "size": 20, "weight": 500 }, "columnHead": { "size": 22, "weight": 700 }, "rowHead": { "size": 22, "weight": 500 }, "variant": { "size": 48, "weight": 700 }, "cellCount": { "size": 20, "weight": 500 }, "asideHeading": { "size": 24, "weight": 700 }, "balanceMark": { "size": 30, "weight": 700 }, "balanceCopy": { "size": 20, "weight": 400 }, "totalValue": { "size": 42, "weight": 700, "numeric": "tabular" }, "totalLabel": { "size": 20, "weight": 500 }, "conclusion": { "size": 25, "weight": 500 } }, "colors": { "ground": "#FFFFFF", "surface": "#F4F6F9", "ink": "#131820", "muted": "#667183", "accent": "#1F5FA8", "rule": "#DBE1E9", "series": [ "#2F6FB5", "#A85A2E", "#1C8C79", "#4E7A3A" ] }, "rules": [ { "ja": "顧客層を行、配信枠を列に置き、各行・各列へ件名A〜Dを1回ずつ割り当てる。案が特定の層や時刻に偏ると、件名の効果と条件差を分けられない", "en": "Put audience groups in rows and send windows in columns. Assign A–D once in every row and column; otherwise subject-line effects are confounded with the group or time." }, { "ja": "1セルの送信数を揃える。異なる件名の反応率を比べる前に、各案が受け取る母数を同じにする", "en": "Keep the send count equal in every cell so each subject line receives the same audience size before comparing response rates." }, { "ja": "A〜Dの文字を色と一緒に残す。色だけで案を識別させない", "en": "Keep the A–D labels alongside the colours; colour alone must not identify a variant." }, { "ja": "これは割付の例であり、無作為抽出・重複除外・配信到達を保証するものではない", "en": "This illustrates assignment only; it does not guarantee random sampling, deduplication, or successful delivery." } ], "cover": "https://design-vault.pages.dev/_astro/cover.DRHIBuEj.png", "assets": [], "shortlist": false, "status": "published", "relations": { "similar": [] }, "license": "Original", "publishedAt": null, "reproduction": { "spec": { "format": "slide-16x9", "density": "medium", "layout": { "kind": "matrix", "canvas": { "w": 1920, "h": 1080 }, "margin": { "top": 82, "right": 120, "bottom": 82, "left": 120 } }, "type": { "title": { "size": 44, "weight": 700, "tracking": "0.01em" }, "subtitle": { "size": 22, "weight": 400 }, "axisNote": { "size": 20, "weight": 500 }, "corner": { "size": 20, "weight": 500 }, "columnHead": { "size": 22, "weight": 700 }, "rowHead": { "size": 22, "weight": 500 }, "variant": { "size": 48, "weight": 700 }, "cellCount": { "size": 20, "weight": 500 }, "asideHeading": { "size": 24, "weight": 700 }, "balanceMark": { "size": 30, "weight": 700 }, "balanceCopy": { "size": 20, "weight": 400 }, "totalValue": { "size": 42, "weight": 700, "numeric": "tabular" }, "totalLabel": { "size": 20, "weight": 500 }, "conclusion": { "size": 25, "weight": 500 } }, "colors": { "ground": "#FFFFFF", "surface": "#F4F6F9", "ink": "#131820", "muted": "#667183", "accent": "#1F5FA8", "rule": "#DBE1E9", "series": [ "#2F6FB5", "#A85A2E", "#1C8C79", "#4E7A3A" ] }, "rules": [ "Put audience groups in rows and send windows in columns. Assign A–D once in every row and column; otherwise subject-line effects are confounded with the group or time.", "Keep the send count equal in every cell so each subject line receives the same audience size before comparing response rates.", "Keep the A–D labels alongside the colours; colour alone must not identify a variant.", "This illustrates assignment only; it does not guarantee random sampling, deduplication, or successful delivery." ] }, "prompt": { "ja": "以下の仕様でスライドを 1 枚作ってください。\n\n- 版面 1920 × 1080\n- 余白 上 82 右 120 下 82 左 120\n\n## 文字（版面 1920px に対する px）\ntitle 44px / 700 / 字間 0.01em\nsubtitle 22px / 400\naxisNote 20px / 500\ncorner 20px / 500\ncolumnHead 22px / 700\nrowHead 22px / 500\nvariant 48px / 700\ncellCount 20px / 500\nasideHeading 24px / 700\nbalanceMark 30px / 700\nbalanceCopy 20px / 400\ntotalValue 42px / 700 / 等幅数字\ntotalLabel 20px / 500\nconclusion 25px / 500\n\n## 配色\n地 #FFFFFF / 面 #F4F6F9 / 文字 #131820 / 補助 #667183 / 強調 #1F5FA8 / 罫 #DBE1E9 / #2F6FB5 / #A85A2E / #1C8C79 / #4E7A3A\n\n## 外すと崩れる点\n- 顧客層を行、配信枠を列に置き、各行・各列へ件名A〜Dを1回ずつ割り当てる。案が特定の層や時刻に偏ると、件名の効果と条件差を分けられない\n- 1セルの送信数を揃える。異なる件名の反応率を比べる前に、各案が受け取る母数を同じにする\n- A〜Dの文字を色と一緒に残す。色だけで案を識別させない\n- これは割付の例であり、無作為抽出・重複除外・配信到達を保証するものではない\n\n内容は［ここに自分の内容を書く］に置き換えてください。寸法と配色はそのまま守ってください。\n正確に再現する場合は、この HTML をそのまま使ってください: https://design-vault.pages.dev/designs/latin-square-subject-rotation/pattern.html", "en": "Produce one slide to the following specification.\n\n- Canvas 1920 × 1080\n- Margin top 82 right 120 bottom 82 left 120\n\n## Type (px against a 1920px canvas)\ntitle 44px / 700 / tracking 0.01em\nsubtitle 22px / 400\naxisNote 20px / 500\ncorner 20px / 500\ncolumnHead 22px / 700\nrowHead 22px / 500\nvariant 48px / 700\ncellCount 20px / 500\nasideHeading 24px / 700\nbalanceMark 30px / 700\nbalanceCopy 20px / 400\ntotalValue 42px / 700 / tabular numerals\ntotalLabel 20px / 500\nconclusion 25px / 500\n\n## Colour\nGround #FFFFFF / Surface #F4F6F9 / Ink #131820 / Muted #667183 / Accent #1F5FA8 / Rule #DBE1E9 / #2F6FB5 / #A85A2E / #1C8C79 / #4E7A3A\n\n## Constraints — dropping these breaks it\n- Put audience groups in rows and send windows in columns. Assign A–D once in every row and column; otherwise subject-line effects are confounded with the group or time.\n- Keep the send count equal in every cell so each subject line receives the same audience size before comparing response rates.\n- Keep the A–D labels alongside the colours; colour alone must not identify a variant.\n- This illustrates assignment only; it does not guarantee random sampling, deduplication, or successful delivery.\n\nReplace the content with [your content]. Keep the measurements and colours as given.\nFor an exact reproduction, use the verbatim HTML at https://design-vault.pages.dev/designs/latin-square-subject-rotation/pattern.html." }, "code": { "html": "\n\n件名A〜Dを公平に比べる — 顧客層と配信枠の偏りを抑える\n\n\n\n \n \n \n \n 件名A〜Dを公平に比べる — 顧客層と配信枠の偏りを抑える\n 架空の案内メール実験 / 4つの顧客層 × 4つの配信枠 / 各セル200件\n \n\n 配信枠 →\n 顧客層 ↓\n 月曜 9時\n 月曜 13時\n 火曜 9時\n 火曜 13時\n \n\n 初回利用\n 月1〜2回利用\n 週1回以上利用\n 法人管理者\n\n \n A200件\n B200件\n C200件\n D200件\n \n \n B200件\n C200件\n D200件\n A200件\n \n \n C200件\n D200件\n A200件\n B200件\n \n \n D200件\n A200件\n B200件\n C200件\n \n\n \n \n 偏りを割り当てで均す\n 行\n 各顧客層にA〜Dが1回ずつ\n \n 列\n 各配信枠にもA〜Dが1回ずつ\n \n 数\n 各セルの送信数を同じにする\n \n 3,200件\n 合計 / 件名ごとに800件\n\n \n 件名だけを比べるなら、顧客層と配信枠にA〜Dを均等に割り当てる。\n \n\n", "url": "https://design-vault.pages.dev/designs/latin-square-subject-rotation/pattern.html", "fonts": [ "https://design-vault.pages.dev/designs/latin-square-subject-rotation/font.css", "https://design-vault.pages.dev/designs/latin-square-subject-rotation/font-0.woff2", "https://design-vault.pages.dev/designs/latin-square-subject-rotation/font-1.woff2", "https://design-vault.pages.dev/designs/latin-square-subject-rotation/font-2.woff2" ] } }, "urls": { "ja": "https://design-vault.pages.dev/designs/latin-square-subject-rotation", "en": "https://design-vault.pages.dev/en/designs/latin-square-subject-rotation" }, "notesMarkdown": "件名テストは、配信時刻と受け手の違いが反応に影響する。件名Aを朝に新規ユーザーへ、件名Bを\n夕方に常連へ送って比べても、違いが件名によるものか条件によるものか分からない。\n\nこの割付では、4つの顧客層と4つの配信枠を交差させ、16セルすべてにA〜Dを1回ずつ配置した。\n各案は各層・各時刻に一度ずつ現れ、各セル200件なので、件名ごとの送信数も800件で揃う。\n実運用ではセル内の対象を無作為に割り当て、配信失敗や重複を別に確認する。" }