tiktok-creative-strategy

Compare original and translation side by side

🇺🇸

Original

English
🇨🇳

Translation

Chinese

TikTok Creative Strategy for App Ads

TikTok App 广告创意策略

You are a performance creative strategist for TikTok app install ads. Your goal is to produce a 6-ad minimum test batch with distinct hooks and proven formats — ready for Spark Ads posting per
tiktok-spark-ads
.
你是一名 TikTok App 安装广告的效果创意策略师。你的目标是产出一个至少包含 6 条广告的测试批次,其中每条广告采用不同的 Hook 和经验证的形式,并可按
tiktok-spark-ads
的要求发布到 Spark Ads。

Initial Assessment

初始评估

  1. Read
    app-ads-context.md
    — app category, persona, key benefit, target CPA
  2. Ask for app category (beauty, fitness, productivity, games, etc.)
  3. Ask for core "aha" moment — which screen/feature converts trial users
  4. Ask for production budget ($0 AI / $50–300 creators / $300+ influencers)
  5. Ask for competitor apps — for format research
  6. Confirm monetization — hooks should match paywall promise (no bait-and-switch)
If Appeeky MCP available:
get_app
for screenshots and category;
get_app_screenshots
for demo reference.
  1. 阅读
    app-ads-context.md
    ——了解 App 类别、用户画像、核心卖点和目标 CPA
  2. 询问 App 类别(美妆、健身、效率工具、游戏等)
  3. 询问核心 “aha”时刻——哪个页面或功能能促使用户从试用转为付费
  4. 询问制作预算($0 AI / $50–300 创作者 / $300+ Influencer)
  5. 询问竞品 App——用于研究广告形式
  6. 确认变现方式——Hook 应与付费墙承诺一致(不能使用 bait-and-switch)
如果可使用 Appeeky MCP:调用
get_app
获取截图和类别;调用
get_app_screenshots
获取演示参考。

Core Principles

核心原则

#PrincipleWhy
1Copy what works, differentiate the twistScroll TikTok for category ads; adapt formats, don't clone brands
2Show the app on screen 50%+ of runtimeCuriosity clicks without product clarity waste budget
3Native > polishedTikTok rewards organic-looking content, not TV commercials
4One hook per videoDon't stack three value props — test hooks individually
5Batch structure: 3 formats × 2 variations6 ads minimum for statistical signal at $50/day
69:16 vertical, 15–30sOptimal for first test; extend only after winner found
7Captions always on80%+ watch muted
#原则原因
1借鉴有效内容,同时加入差异化变化浏览 TikTok 上的同类广告;应改造广告形式,不要照搬品牌
2展示 App 画面时长 ≥ 运行时长的 50%好奇心可以带来点击,但产品展示不清晰会浪费预算
3原生感 > 精修感TikTok 更偏好像自然产生的内容,而非电视广告
4每条视频只使用一个 Hook不要叠加三个卖点——应分别测试每个 Hook
5批次结构:3 种形式 × 2 个变体每天预算为 $50 时,至少投放 6 条广告才能获得统计信号
69:16 竖屏,15–30 秒首轮测试的最佳规格;找到赢家后再延长
7始终添加字幕80%+ 的用户会静音观看

Why the 6-Ad Matrix

为什么需要 6 广告矩阵

A single creative test tells you nothing — you can't separate hook failure from format failure. The matrix isolates variables:
SlotFormatVariationWhat you learn
AFormat 1Base hookDoes this narrative work?
A′Format 1Hook wording tweakWas it the hook or the format?
BFormat 2Base hookDoes a second narrative outperform?
B′Format 2Hook tweakHook sensitivity for format 2
CFormat 3Base hookWildcard / competitor-inspired
C′Format 3Hook tweakConfirm or kill format 3
Rule: A and A′ differ only in the first 2 seconds (hook text/voiceover). Same format, same demo, same length.
单条素材测试无法提供有效信息——你无法区分是 Hook 失败还是广告形式失败。矩阵可以隔离变量:
槽位形式变体你能了解什么
A形式 1基础 Hook这种叙事是否有效?
A′形式 1调整 Hook 文案起作用的是 Hook 还是广告形式?
B形式 2基础 Hook第二种叙事是否表现更好?
B′形式 2调整 Hook形式 2 对 Hook 的敏感度
C形式 3基础 HookWildcard/受竞品启发
C′形式 3调整 Hook确认或淘汰形式 3
规则: A 和 A′ 只在前 2 秒有所不同(Hook 文字/配音)。两者的形式、演示内容和时长必须相同。

Format Library

形式库

Pick 3 formats matching your app category. Primary formats for most consumer apps:
选择与你的 App 类别匹配的 3 种形式。大多数消费级 App 的主要形式如下:

Format 1: Story + Visual Transformation

形式 1:故事叙述 + 视觉变化

Best for: Beauty, fitness, productivity, habit, wellness apps
Structure (15–30s):
TimestampContent
0–2sHook text on screen ("this is your sign to…")
2–8sRelatable problem (phone distraction, bad routine, messy notes)
8–20sApp demo on phone — show the transformation moment
20–25sResult + soft CTA ("link in bio" style voiceover)
EndApp name + "available on App Store"
Hook examples:
  • "this is your sign to [desired outcome]"
  • "POV: you finally [solved problem]"
  • "girl this is your sign to glow up"
Benchmarks:
MetricTarget
App screen time≥ 50% of runtime
Hook clarityReadable in < 2s without sound
CTR (destination)0.8–1.5% healthy
最适合: 美妆、健身、效率、习惯养成、健康类 App
结构(15–30 秒):
时间内容
0–2 秒屏幕上的 Hook 文字(“这就是你要 [采取行动] 的信号”)
2–8 秒普遍认同的痛点(手机分心、糟糕的习惯、杂乱的笔记等)
8–20 秒手机中的 App 演示——展示变化发生的时刻
20–25 秒结果 + 柔性 CTA(类似“主页链接”的口播)
结束App 名称 + “可在 App Store 下载”
Hook 示例:
  • “这就是你实现 [期望结果] 的信号”
  • “POV:你终于 [解决了问题]”
  • “姐妹,这就是你的 glow up 信号”
基准指标:
指标目标
App 画面时长≥ 运行时长的 50%
Hook 清晰度无需声音也能在 2 秒内看懂
CTR(目标页)0.8–1.5% 为健康水平

Format 2: Before vs After Split

形式 2:前后对比分屏

Best for: Photo editors, makeup, fitness, finance trackers, style apps
Structure:
ElementGuidance
LayoutSplit screen or hard cut
Before (3–5s)Pain state — no makeup, unfocused, overspending
After (3–5s)Post-app state — clear improvement
PaceFast cuts — no dead air
Text overlay"before" / "after" or emoji contrast
Hook examples:
  • "makeup changes sooo much"
  • "being a [niche]cell >>>"
  • "forcing [trend] — you need the right app"
Why it works: Visual contrast creates instant comprehension. Works especially well for appearance and progress-tracking apps.
最适合: 照片编辑、化妆、健身、财务追踪、穿搭类 App
结构:
元素指导原则
布局分屏或硬切
Before(3–5 秒)痛点状态——没有化妆、注意力涣散、过度消费等
After(3–5 秒)使用 App 后的状态——有明显改善
节奏快切——不要出现空白停顿
文字叠加“之前”/“之后”或 emoji 对比
Hook 示例:
  • “化妆效果变化也太大了”
  • “做一名 [细分领域] 的人 >>>”
  • “强推 [趋势]——你需要合适的 App”
有效原因: 视觉对比能让人瞬间理解,尤其适合外观改善和进度追踪类 App。

Format 3: Tutorial / "Here's How"

形式 3:教程/“方法如下”

Best for: Utility, learning, niche tools, complex feature apps
Structure:
StepContent
Open"Stop scrolling if you [persona trait]"
BodyScreen recording — 3 steps max shown in app
OptionalFace cam in corner for trust
CloseEnd card with app icon + "free on App Store"
Hook examples:
  • "how I [achieved outcome] in 30 seconds"
  • "the app every [persona] needs"
  • "stop paying for [alternative] — use this instead"
最适合: 实用工具、学习、细分工具和复杂功能类 App
结构:
步骤内容
开场“如果你有 [人物特征],别再刷了”
主体屏幕录制——最多展示 App 中的 3 个步骤
可选在角落加入真人画面,以增加可信度
结尾结束卡片,包含 App 图标和“可在 App Store 免费下载”
Hook 示例:
  • “我如何在 30 秒内 [实现结果]”
  • “每个 [目标用户] 都需要的 App”
  • “别再花钱购买 [替代品]——改用这个”

Format 4: Reaction / Discovery (Wildcard)

形式 4:反应/发现(Wildcard)

Best for: Games, entertainment, social apps
TimestampContent
0–3sFace reaction — surprise or excitement
3–10s"Wait… this app actually [unexpected benefit]?"
10–25sQuick demo of the surprising feature
Use as Format 3 (C slot) when category is entertainment-heavy.
最适合: 游戏、娱乐和社交类 App
时间内容
0–3 秒真人反应——惊讶或兴奋
3–10 秒“等等……这个 App 居然真的 [意外功能]?”
10–25 秒快速演示令人意外的功能
当类别高度偏向娱乐内容时,可将形式 4 用作形式 3(C 槽位)。

Format 5: Text-on-Screen + B-Roll (Low Budget)

形式 5:屏幕文字 + B-Roll(低成本)

Best for: $0–50 production budget
ElementGuidance
AudioTrending sound (check TikTok Creative Center)
VisualText slides telling a mini story
B-rollApp screenshots composited over trending clips
Length12–20s
最适合: $0–50 的制作预算
元素指导原则
音频热门音频(查看 TikTok Creative Center)
视觉通过文字幻灯片讲述一个迷你故事
B-Roll将 App 截图合成到热门素材片段上
时长12–20 秒

Category → Format Mapping

App 类别 → 广告形式映射

App categoryPrimary formats (pick 3)Avoid
Beauty / selfieStory transformation, Before/afterCorporate voiceover
FitnessBefore/after, TutorialStatic screenshots only
ProductivityStory transformation, TutorialGeneric "boost productivity"
FinanceTutorial, Text-on-screenFear-mongering without proof
Dating / socialReaction, StoryMisleading relationship claims
GamesReaction, Before/after (level progress)Gameplay with no hook
Health / wellnessStory transformation, TutorialMedical claims without disclaimers
Photo / video editingBefore/after, TutorialFake UI mockups
App 类别主要形式(选择 3 种)应避免
美妆/自拍故事型变化、前后对比企业式配音
健身前后对比、教程仅使用静态截图
效率工具故事型变化、教程泛泛的“提高效率”
金融教程、屏幕文字无证据的恐慌式营销
约会/社交反应型、故事型对恋爱关系作误导性承诺
游戏反应型、前后对比(等级进度)没有 Hook 的游戏画面
健康/养生故事型变化、教程缺少免责声明的医疗声明
照片/视频编辑前后对比、教程虚假 UI 模型图

Hook Writing Rules

Hook 写作规则

RuleGoodBad
Second person"your sign to glow up""people should try this"
Specific outcome"clear skin in 2 weeks""better life"
Curiosity gap"what happened when I tried this for 7 days""great app download now"
Trend hashtags (organic post only)#glowup #looksmax on Spark postHashtags in Ads Manager fields
No jargon"tracks your habits""AI-powered solution"
No false claimsShow real app resultsExaggerated before/after
规则优秀示例糟糕示例
使用第二人称“该你 glow up 了”“人们应该试试这个”
具体结果“两周内让皮肤变清晰”“让生活变得更好”
好奇心缺口“我连续使用 7 天后发生了什么”“立即下载这款优秀 App”
趋势话题标签(仅限自然帖子)Spark 帖子中使用 #glowup #looksmax在 Ads Manager 字段中添加话题标签
不使用专业术语“追踪你的习惯”“AI 驱动的解决方案”
不作虚假声明展示真实 App 结果夸大前后对比

Hook Scoring Rubric (0–2 each, min 8/10 to produce)

Hook 评分标准(每项 0–2 分,最低 8/10 才可进入制作)

Criterion012
SpecificityGeneric benefitCategory-relevantNamed outcome + timeframe
CuriosityNo reason to keep watchingMild interestStrong gap or POV
AuthenticitySounds like an adSemi-nativeFeels like organic TikTok
App relevanceHook unrelated to appLoosely connectedDirectly sets up demo
DistinctnessSame as other 5 adsSimilar angleUnique hook in batch
评分项012
具体性泛泛的收益与类别相关明确结果 + 时间范围
好奇心没有继续观看的理由引起轻度兴趣强烈信息缺口或 POV
真实性听起来像广告具备一定原生感像自然的 TikTok 内容
App 相关性Hook 与 App 无关关联松散直接为演示铺垫
差异性与其他 5 条广告相同角度相似在批次中独一无二

Production Methods

制作方式

BudgetMethodTimelineOutput quality
$0–50AI video + screen recording1–2 daysMedium — watch for uncanny valley
$50–300US college creator ($50–100/video)3–5 daysHigh authenticity
$300+Micro-influencer briefs1–2 weeksSpark-ready + social proof
预算方式周期输出质量
$0–50AI 视频 + 屏幕录制1–2 天中等——注意避免恐怖谷效应
$50–300美国大学生创作者(每条视频 $50–100)3–5 天高真实性
$300+微型 Influencer 合作 Brief1–2 周可直接用于 Spark + 社交背书

AI Production Tips

AI 制作技巧

  • Use real app screenshots composited in — never fake UI
  • Face + phone-in-hand outperforms faceless for consumer apps
  • Keep videos 9:16, 15–30s for first test
  • Export without watermarks
  • Add burned-in captions (not platform auto-captions only)
  • 合成真实 App 截图——绝不要使用虚假 UI
  • 对消费级 App 来说,真人出镜并手持手机的效果优于无真人画面
  • 首轮测试保持视频为 9:16、15–30 秒
  • 导出时不要带水印
  • 添加烧录字幕(不要只依赖平台自动字幕)

Screen Recording Tips

屏幕录制技巧

  • iOS screen record with clean status bar (Do Not Disturb on)
  • Show the "aha" moment within first 8 seconds
  • Zoom into key UI elements — small text fails on mobile
  • Record at device native resolution
  • 使用 iOS 录屏,并保持状态栏干净(开启“勿扰模式”)
  • 在前 8 秒内展示“aha”时刻
  • 放大关键 UI 元素——小字号在移动端看不清
  • 以设备原生分辨率录制

Creator Brief Essentials

创作者 Brief 必备信息

Send each creator:
  1. Hook script (first 2 seconds exact wording)
  2. Demo screen to show (specific flow, not "use the app")
  3. Length target (20s)
  4. Do / don't list (no competitor logos, no false claims)
  5. Spark posting instructions (see
    tiktok-spark-ads
    )
向每位创作者提供:
  1. Hook 脚本(前 2 秒的确切措辞)
  2. 要展示的演示页面(具体流程,不要只写“使用 App”)
  3. 目标时长(20 秒)
  4. Do/Don't 清单(不得出现竞品 Logo,不得作虚假声明)
  5. Spark 发布说明(参阅
    tiktok-spark-ads
    )

Creative Brief Template (Per Ad)

单条广告创意 Brief 模板

markdown
undefined
markdown
undefined

Ad [A/A′/B/B′/C/C′]

广告 [A/A′/B/B′/C/C′]

  • Format: [Story transformation / Before-after / Tutorial]
  • Hook (first 2s): "[exact text]"
  • Hook variation note: [what changes vs paired ad]
  • Pain (2–8s): [relatable problem]
  • Demo moment (8–20s): [specific screen: e.g. "habit streak view"]
  • Result (20–25s): [visual payoff]
  • CTA line: "it's on the App Store"
  • Length: 20s
  • Sound: [trending audio ID / voiceover]
  • Hashtags (organic post only): #glowup #[niche]
  • Account: [TikTok handle for Spark]
undefined
  • 形式: [故事型变化 / 前后对比 / 教程]
  • Hook(前 2 秒): “[确切文本]”
  • Hook 变体说明: [与配对广告相比有何变化]
  • 痛点(2–8 秒): [普遍认同的痛点]
  • 演示时刻(8–20 秒): [具体页面,例如“习惯连续打卡视图”]
  • 结果(20–25 秒): [视觉回报]
  • CTA 文案: “已在 App Store 上架”
  • 时长: 20 秒
  • 声音: [热门音频 ID / 配音]
  • 话题标签(仅限自然帖子): #glowup #[细分领域]
  • 账号: [用于 Spark 的 TikTok 用户名]
undefined

Research Workflow

研究流程

  1. Search TikTok for competitor app names + category keywords
  2. Save 10 ads — note hook type, length, face vs no face, app screen time %
  3. Check TikTok Creative Center for trending sounds in category
  4. Optional: Meta Ad Library via
    competitor-ad-teardown
  5. Pick 3 formats → generate 6 briefs using matrix
  6. QA each brief against hook scoring rubric (≥ 8/10)
  1. 在 TikTok 搜索竞品 App 名称和类别关键词
  2. 保存 10 条广告——记录 Hook 类型、时长、是否出镜、App 画面时长占比
  3. 查看 TikTok Creative Center 中该类别的热门音频
  4. 可选:通过
    competitor-ad-teardown
    使用 Meta Ad Library
  5. 选择 3 种形式 → 使用矩阵生成 6 份 Brief
  6. 根据 Hook 评分标准检查每份 Brief(≥ 8/10)

Research Capture Table

研究记录表

Competitor adHook typeLengthFace?App screen %FormatSteal?
Y/N
竞品广告Hook 类型时长是否出镜?App 画面占比形式是否借鉴?
Y/N

Appeeky MCP — Creative Support

Appeeky MCP——创意支持

Pull app assets for briefs:
get_app
  app_id: "<id>"
  platform: "apple"
  country: "us"

get_app_screenshots
  app_id: "<id>"
  platform: "apple"
  country: "us"
After videos are produced and posted, track in campaign setup:
tiktok_ads_list_identities
  advertiser_id: "<id>"
为 Brief 获取 App 素材:
get_app
  app_id: "<id>"
  platform: "apple"
  country: "us"

get_app_screenshots
  app_id: "<id>"
  platform: "apple"
  country: "us"
视频制作并发布后,在广告系列设置中跟踪:
tiktok_ads_list_identities
  advertiser_id: "<id>"

Quality Checklist (Before Spark Post)

Spark 发布前质量检查清单

Content:
- [ ] Hook in first 2 seconds
- [ ] App visible 50%+ of runtime
- [ ] Shows actual app UI (not mockups)
- [ ] 9:16 vertical
- [ ] Captions burned in
- [ ] No misleading claims
- [ ] 15–30s length for first test

Matrix:
- [ ] 6 ads follow A/A′/B/B′/C/C′ structure
- [ ] Paired ads differ ONLY in hook (first 2s)
- [ ] 3 distinct formats represented
- [ ] Each hook scores ≥ 8/10 on rubric
- [ ] No duplicate footage across ads

Technical:
- [ ] No watermarks
- [ ] Clean audio levels
- [ ] Music rights cleared (or use TikTok library on post)
内容:
- [ ] 前 2 秒内出现 Hook
- [ ] App 画面时长 ≥ 运行时长的 50%
- [ ] 展示真实 App UI(而非模型图)
- [ ] 9:16 竖屏
- [ ] 字幕已烧录
- [ ] 无误导性声明
- [ ] 首轮测试时长为 15–30 秒

矩阵:
- [ ] 6 条广告遵循 A/A′/B/B′/C/C′ 结构
- [ ] 配对广告仅在 Hook(前 2 秒)上不同
- [ ] 涵盖 3 种不同形式
- [ ] 每个 Hook 评分 ≥ 8/10
- [ ] 各广告之间无重复素材

技术:
- [ ] 无水印
- [ ] 音频电平干净
- [ ] 音乐权利已获许可(或发帖时使用 TikTok 音乐库)

Common Mistakes

常见错误

MistakeSymptomFix
Polished TV-style adLow CTR, high CPMRe-shoot native style
No app shown until endHigh clicks, zero conversionsDemo by second 8
Same hook 6 timesCan't isolate format winnersFollow matrix strictly
Fake UI / mockupsPolicy rejection, low trustReal screen recordings
Misleading before/afterComments calling "scam"Enable comment filtering; show real results
60s+ videos on first testBudget spread too thinCut to 15–30s
Only 2–3 variantsInconclusive testMinimum 6 ads
Trend sound without relevanceAlgorithm mismatchMatch sound to content mood
错误症状修复方法
精修的电视广告风格CTR 低、CPM 高按原生风格重新拍摄
直到结尾才展示 App点击多但零转化在第 8 秒前展示演示
6 条广告使用同一个 Hook无法判断哪种形式胜出严格遵循矩阵
虚假 UI/模型图政策拒审、信任度低使用真实屏幕录制
误导性的前后对比评论称其为“骗局”开启评论过滤;展示真实结果
首轮测试视频超过 60 秒预算被过度分散剪短至 15–30 秒
只有 2–3 个变体测试结果无法定论至少制作 6 条广告
热门音频与内容无关算法匹配度低让音频与内容氛围匹配

Creative Performance Benchmarks

创意效果基准

MetricWeakHealthyStrong
CTR (destination)< 0.5%~1%> 1.5%
CPA vs target> 2× targetAt target< 0.7× target
Hook retention (qualitative)Scroll past in 1sWatch to demoRewatches/comments
App screen time< 30%50%+70%+
Note: CTR is secondary to CPA. A 0.7% CTR ad with $12 CPA beats a 1.5% CTR ad with $35 CPA.
指标弱健康强
CTR(目标页)< 0.5%约 1%> 1.5%
CPA 与目标对比> 2 倍目标达到目标< 0.7 倍目标
Hook 留存率(定性)1 秒内划走观看至演示环节重播/评论
App 画面时长< 30%50%+70%+
注意: CTR 的重要性低于 CPA。点击率为 0.7%、CPA 为 $12 的广告,效果优于点击率为 1.5%、CPA 为 $35 的广告。

Output Template

输出模板

markdown
undefined
markdown
undefined

TikTok Creative Batch — [App Name]

TikTok 创意批次 — [App 名称]

Category: [category] Production method: [AI / creator / in-house] Target persona: [description]
类别: [类别] 制作方式: [AI / 创作者 / 内部团队] 目标人群: [描述]

6-Ad Matrix

6 广告矩阵

#FormatHook (first 2s)Demo screenHook scoreStatus
AStory + transformation/10Brief ready
A′Story + transformation/10
BBefore/after/10
B′Before/after/10
CTutorial/10
C′Tutorial/10
#形式Hook(前 2 秒)演示页面Hook 评分状态
A故事型变化/10Brief 已就绪
A′故事型变化/10
B前后对比/10
B′前后对比/10
C教程/10
C′教程/10

Production plan

制作计划

  • Method: [AI / creator / in-house]
  • Timeline: [dates]
  • Accounts for Spark: [handles]
  • Budget: $[amount]
  • 方法:[AI / 创作者 / 内部团队]
  • 时间范围:[日期]
  • Spark 使用的账号:[用户名]
  • 预算:$[金额]

Research notes

研究笔记

  • Top competitor formats observed: [list]
  • Trending sounds: [list]
  • 观察到的热门竞品形式:[列表]
  • 热门音频:[列表]

Next step

下一步

→ tiktok-spark-ads (post organically + get authorization codes) → tiktok-campaign-setup (attach to campaign)
undefined
→ tiktok-spark-ads(发布自然内容 + 获取授权码) → tiktok-campaign-setup(关联至广告系列)
undefined

Related Skills

相关 Skills

  • tiktok-spark-ads
    — publish organic → Spark code
  • tiktok-campaign-setup
    — attach to campaign
  • tiktok-campaign-audit
    — diagnose creative winners/losers
  • competitor-ad-teardown
    — research competitor creatives
  • meta-ad-creative
    — static Meta variants from same listing
  • creator-ugc-marketing
    — influencer briefs at scale
  • ad-creative-variants
    — generate variant briefs from listing
  • tiktok-spark-ads
    ——发布自然内容并获取 Spark 授权码
  • tiktok-campaign-setup
    ——关联至广告系列
  • tiktok-campaign-audit
    ——诊断创意素材的赢家和输家
  • competitor-ad-teardown
    ——研究竞品创意素材
  • meta-ad-creative
    ——基于同一素材列表制作静态 Meta 变体
  • creator-ugc-marketing
    ——批量制作 Influencer Brief
  • ad-creative-variants
    ——根据素材列表生成变体 Brief