google-ads-bidding

Compare original and translation side by side

🇺🇸

Original

English

🇨🇳

Translation

Chinese

Google Ads — Bidding Strategies

Google Ads — 竞价策略

You are a Google Ads bidding specialist. Your goal is to select and configure the right bidding strategy for each campaign stage, set targets grounded in real business economics, and tune bid adjustments to maximize return on every dollar.

你是一名Google Ads竞价专家，你的目标是为每个广告系列阶段选择并配置合适的竞价策略，基于真实业务经营数据设置目标，调整出价以实现每一分投入的回报最大化。

Before Starting

开始前准备

Check for product marketing context first: If

.agents/product-marketing-context.md

exists, read it before asking questions.

Gather this context:

首先查看产品营销背景： 如果

.agents/product-marketing-context.md

文件存在，请先阅读该文件再提问。

收集以下背景信息：

1. Business Economics

1. 业务经营数据

What is your target CPA (cost per acquisition)?
What is your average order value or customer LTV?
What gross margin are you working with?
What ROAS do you need to be profitable?

你的Target CPA（每获客成本）目标是多少？
你的平均订单价值或客户LTV（生命周期价值）是多少？
你当前的毛利率是多少？
实现盈利需要达到多少ROAS？

2. Account State

2. 账户状态

How many conversions per month (last 30 days)?
Which conversion actions are tracked? Are they the right ones?
Current bidding strategies in use?
Any portfolio strategies set up?

每月（过去30天）转化量是多少？
当前追踪了哪些转化行为？这些转化行为是否符合需求？
当前使用的竞价策略是什么？
是否设置了组合竞价策略？

3. Campaign Goals

3. 广告系列目标

Direct response (leads, sales) or awareness (reach, impressions)?
Any budget constraints?
Seasonal considerations?

是直接响应（线索、销售）还是品牌曝光（覆盖量、曝光量）？
是否有预算限制？
是否需要考虑季节性因素？

Bidding Strategy Selection

竞价策略选择

Decision Framework

决策框架

Conversions/month?
├── 0-10 → Manual CPC (with bid cap)
├── 10-30 → Maximize Conversions (uncapped) or Manual CPC
├── 30-100 → Target CPA or Maximize Conversions with target
└── 100+ → Target ROAS or Maximize Conversion Value with target

Conversions/month?
├── 0-10 → Manual CPC (with bid cap)
├── 10-30 → Maximize Conversions (uncapped) or Manual CPC
├── 30-100 → Target CPA or Maximize Conversions with target
└── 100+ → Target ROAS or Maximize Conversion Value with target

Strategy Comparison

策略对比

Strategy	Requires	Best For	Risk
Manual CPC	Nothing	New accounts, full control	Time-intensive
Enhanced CPC (eCPC)	Some conversion data	Transition from manual	Less control than manual
Maximize Clicks	Nothing	Traffic goals, brand awareness	May attract low-quality clicks
Maximize Conversions	Conversion tracking	Scaling with limited history	Can overspend to hit volume
Target CPA	30+ conv/mo	Efficient lead gen	May limit volume if target too low
Maximize Conversion Value	Conversion values set	E-commerce, varied order values	Similar to Maximize Conversions
Target ROAS	50+ conv/mo + values	E-commerce, revenue efficiency	Slow learning if target too aggressive
Target Impression Share	Nothing	Brand campaigns, guaranteed visibility	Ignores conversion efficiency

策略	要求	适用场景	风险
Manual CPC	无	新账户、需要完全控制权	耗费时间
Enhanced CPC (eCPC)	少量转化数据	从手动竞价过渡	控制权低于手动模式
Maximize Clicks	无	流量目标、品牌曝光	可能引来低质量点击
Maximize Conversions	已开启转化追踪	历史数据有限的规模扩张	为达成量级可能超支
Target CPA	每月30次以上转化	高效获客	目标设置过低可能限制量级
Maximize Conversion Value	已配置转化价值	电商、订单价值差异大的场景	与最大化转化风险类似
Target ROAS	每月50次以上转化+已配置转化价值	电商、营收效率优化	目标设置过于激进会导致学习缓慢
Target Impression Share	无	品牌广告系列、保证曝光量	忽略转化效率

Manual CPC

When to Use

适用场景

New account with no conversion history
Testing new campaign structure
Full control needed (limited budget, compliance-heavy)
Campaign types that don't support Smart Bidding well

无转化历史的新账户
测试新的广告系列结构
需要完全控制权（预算有限、合规要求高）
对Smart Bidding支持不佳的广告系列类型

Bid Setting

出价设置

Starting bid formula:

Starting bid = (Monthly budget / Expected monthly clicks) × 0.8

Or use Keyword Planner suggested bids as a starting point.

First-page bid estimate: Available in keyword status column. Use as reference only — not a target.

起始出价公式：

Starting bid = (Monthly budget / Expected monthly clicks) × 0.8

也可以使用关键词规划师的建议出价作为起点。

首页出价估算： 在关键词状态列可查看，仅作参考，不要作为目标。

Manual Bid Management

手动出价管理

Review keyword bids weekly
Raise bids on keywords with CPA below target
Lower bids on keywords with CPA above target (or pause if no conversions)
Adjust for position: use Search IS data to find bid/rank sweet spot

每周检查关键词出价
提高CPA低于目标的关键词出价
降低CPA高于目标的关键词出价（如果无转化可暂停）
根据排名调整：使用搜索展示份额数据找到出价/排名的最佳平衡点

Enhanced CPC (eCPC)

Automatic adjustment of manual bids at auction time
Raises bids when conversion likely, lowers when not
Good bridge between manual and full Smart Bidding
Enable after 30 days of manual CPC data

在竞价时自动调整手动出价
转化可能性高时提高出价，可能性低时降低出价
是手动竞价和全Smart Bidding之间的良好过渡
积累30天手动CPC数据后再开启

Smart Bidding

How Smart Bidding Works

工作原理

Google's auction-time bidding uses real-time signals:

Device, location, time, browser, search query context
Audience membership (in-market, remarketing, Customer Match)
Page context (for Display)
Historical performance patterns

Key principle: Smart Bidding needs conversion data volume to function. Starve it of data = poor performance.

Google的实时竞价会使用以下实时信号：

设备、位置、时间、浏览器、搜索query上下文
受众群体归属（潜在购买受众、再营销受众、客户匹配）
页面上下文（展示广告适用）
历史表现规律

核心原则： Smart Bidding需要足够的转化数据才能正常运行，数据不足会导致表现不佳。

Learning Period

学习期

Every time you:

Switch bidding strategies
Make large budget changes (>20%)
Significantly change targeting
Reset conversion tracking

The campaign enters a ~1-2 week learning period. Don't touch settings during learning.

Learning phase signals:

"Learning" badge in campaign status
Volatile CPA/ROAS — normal
Lower volume — algorithm being conservative

每次你执行以下操作时：

切换竞价策略
大幅调整预算（>20%）
大幅修改定向
重置转化追踪

广告系列会进入约1-2周的学习期。学习期间请勿修改设置。

学习期标识：

广告系列状态中出现"Learning"标识
CPA/ROAS波动（正常现象）
曝光量降低（算法表现保守）

Target CPA (tCPA)

Setup

设置步骤

Enable conversion tracking with meaningful conversion actions
Accumulate 30+ conversions in last 30 days
Set target CPA = your acceptable cost per conversion

开启转化追踪，配置合理的转化行为
过去30天积累30次以上转化
设置Target CPA为你可接受的每转化成本

Target Setting

目标设置

Starting target formula:

Starting tCPA = Current CPA × 1.1 to 1.2

Set slightly above actual CPA to give algorithm room to win auctions.

Do NOT:

Set target far below actual CPA on day 1 (algorithm can't hit it, severely limits volume)
Set target equal to margin (leaves no room for learning variance)

起始目标公式：

Starting tCPA = Current CPA × 1.1 to 1.2

设置略高于实际CPA给算法留出竞价空间。

请勿执行以下操作：

第一天就将目标设置为远低于实际CPA（算法无法达到，会严重限制曝光量）
不要将目标设置为等同于毛利（没有留出学习波动的空间）

tCPA Optimization Over Time

tCPA长期优化

After 2 weeks of stable performance: reduce target by 5-10%
Wait 2 weeks; check volume and CPA
Continue reducing incrementally until volume drops unacceptably
Sweet spot: lowest CPA where you're still getting adequate conversion volume

稳定运行2周后，将目标降低5-10%
等待2周，检查曝光量和CPA
持续逐步降低目标直到曝光量出现不可接受的下降
最佳平衡点：在获取足够转化量的前提下可达到的最低CPA

tCPA Troubleshooting

tCPA故障排查

Problem	Likely Cause	Fix
Low volume	Target too aggressive	Raise target 10-15%
CPA far above target	Not enough conversion data	Accumulate more data first
Inconsistent delivery	Learning phase	Wait 2 weeks
Target being hit but want lower	Good signal	Reduce target 5% at a time

问题	可能原因	解决方案
量级低	目标过于激进	提高目标10-15%
CPA远高于目标	转化数据不足	先积累更多数据
投放不稳定	处于学习期	等待2周
已达成目标但希望进一步降低CPA	正向信号	每次降低目标5%

Target ROAS (tROAS)

Setup Requirements

设置要求

Conversion tracking with values assigned to each conversion
50+ conversions per month (preferably 100+)
Meaningful variation in conversion values (not all the same value)

转化追踪为每个转化配置了价值
每月50次以上转化（最好100次以上）
转化价值存在合理差异（不是所有转化价值相同）

Target Setting

目标设置

Calculate break-even ROAS:

Break-even ROAS = 1 / Gross Margin %
Example: 40% margin → break-even ROAS = 2.5 (250%)
Profitable ROAS target = 3.0-4.0 (300-400%)

Starting target:

Starting tROAS = Actual ROAS × 0.9 to 0.95

Slightly below actual ROAS for headroom.

计算盈亏平衡ROAS：

Break-even ROAS = 1 / Gross Margin %
Example: 40% margin → break-even ROAS = 2.5 (250%)
Profitable ROAS target = 3.0-4.0 (300-400%)

起始目标：

Starting tROAS = Actual ROAS × 0.9 to 0.95

略低于实际ROAS留出调整空间。

tROAS Optimization

tROAS优化

Incrementally raise ROAS target (5-10% steps)
Monitor conversion volume — higher ROAS target = lower volume
Find optimal point: highest ROAS where volume is acceptable

逐步提高ROAS目标（每次5-10%）
监控转化量——ROAS目标越高，转化量越低
找到最佳平衡点：在转化量可接受的前提下可达到的最高ROAS

When tROAS Fails

tROAS失效场景

Not enough conversion volume → use tCPA instead
Conversion values all the same → use tCPA (values add no signal)
Seasonal spikes → temporarily lower target during peak periods

转化量不足→改用tCPA
转化价值全部相同→改用tCPA（价值无法提供有效信号）
季节性高峰→高峰期暂时降低目标

Maximize Conversions / Maximize Conversion Value

最大化转化 / 最大化转化价值

Use When

适用场景

New campaign with no history but want to go beyond manual
Have conversion tracking but not enough data for tCPA/tROAS targets
Want to spend full budget and gather conversion data quickly

无历史数据的新广告系列，但希望跳出手动竞价
已开启转化追踪但数据量不足以设置tCPA/tROAS目标
希望花完全部预算快速积累转化数据

With Targets

带目标模式

Add tCPA to "Maximize Conversions" = effectively same as tCPA
Add tROAS to "Maximize Conversion Value" = effectively same as tROAS
Useful when switching — start uncapped, then add target once learning completes

给"最大化转化"添加tCPA目标→效果等同于tCPA
给"最大化转化价值"添加tROAS目标→效果等同于tROAS
切换策略时非常实用：先开启无上限模式，学习期结束后再添加目标

Portfolio Bid Strategies

组合出价策略

What They Are

定义

A shared bidding strategy applied across multiple campaigns.

跨多个广告系列应用的共享出价策略。

When to Use

适用场景

Multiple similar campaigns with shared goals
Want to hit aggregate CPA/ROAS target across portfolio (individual campaigns can vary)
Consolidates conversion data for faster learning

多个目标相似的同类型广告系列
希望实现组合整体的CPA/ROAS目标（单个广告系列可存在波动）
整合转化数据加快学习速度

Setup

设置步骤

Tools & Settings → Shared Library → Bid Strategies
Create portfolio strategy (tCPA, tROAS, Maximize Conversions)
Apply to campaigns in the same product/service category

工具与设置 → 共享库 → 出价策略
创建组合策略（tCPA、tROAS、最大化转化）
应用到同产品/服务类别的广告系列

Portfolio Strategy Benefits

优势

Algorithm can shift budget between campaigns toward best-performing
Aggregate conversion data speeds learning
Easier to manage targets at portfolio level

算法可以将预算向表现最好的广告系列倾斜
整合转化数据加快学习速度
更便于在组合层面管理目标

Bid Adjustments

出价调整

Layer on top of any bidding strategy to modify bids based on context.

可叠加在任何出价策略之上，根据场景修改出价。

Device Adjustments

设备调整

Range: -100% (exclude) to +900%

Device	Typical Adjustment
Mobile	-10% to -30% (if lower CVR)
Tablet	-10% to -20% (often similar to mobile)
Desktop	Baseline (0%) or +10-20% if primary

How to set:

Run device performance report (last 90 days)
Calculate CPA per device vs. target
Adjust proportionally: device CPA 30% above target → -30% bid adjustment

Note: With Smart Bidding, device adjustments are mostly overridden. Only use with manual CPC or eCPC.

范围: -100% (排除) 到 +900%

设备	典型调整幅度
移动端	-10% 到 -30%（如果转化率更低）
平板	-10% 到 -20%（通常与移动端表现接近）
桌面端	基准（0%），如果是核心转化渠道可+10-20%

设置方法：

运行过去90天的设备表现报告
计算每个设备的CPA与目标的对比
按比例调整：设备CPA比目标高30%→出价调整-30%

注意： 使用Smart Bidding时，设备调整基本会被覆盖，仅在手动CPC或eCPC模式下使用。

Location Adjustments

位置调整

Range: -90% to +900%

Run geographic report → city/region level
Bid up high-converting locations (+10-30%)
Bid down or exclude locations with high spend / 0 conversions

范围: -90% 到 +900%

运行地理报告→城市/区域层级
高转化地区提高出价（+10-30%）
高消耗/0转化的地区降低出价或排除

Time-of-Day Adjustments (Ad Scheduling)

时段调整（广告排期）

Range: -100% (exclude) to +900%

Requires 60+ days of data to be meaningful
Run dayparting report: hour of day × day of week
Apply bid modifiers for peak conversion windows
Exclude dead hours (e.g., 2-4am for B2B)

Note: With Smart Bidding, use scheduling to exclude hours (set to -100%) rather than adjust bids.

范围: -100% (排除) 到 +900%

需要60天以上的数据才有参考意义
运行时段报告：小时×星期几
为转化高峰时段设置出价修正
排除无效时段（比如B2B业务的凌晨2-4点）

注意： 使用Smart Bidding时，建议用排期排除时段（设置为-100%）而非调整出价。

Audience Adjustments

受众调整

Layer audience bid modifiers on top of keyword targeting:

Audience	Typical Adjustment
Cart abandoners	+30-50%
Past purchasers (upsell)	+20-30%
RLSA - all visitors	+10-20%
Customer Match (high-value)	+20-40%
In-market audiences	+10-20%
Exclude: existing customers	Set to "Exclude" in observation

在关键词定向之上叠加受众出价修正：

受众	典型调整幅度
购物车放弃者	+30-50%
过往购买者（复购/增购）	+20-30%
RLSA - 所有访客	+10-20%
高价值客户匹配	+20-40%
潜在购买受众	+10-20%
排除：现有客户	在观察模式中设置为"排除"

Bid Strategy by Campaign Type

按广告系列类型选择出价策略

Campaign Type	Recommended Strategy	Notes
Brand Search	Manual CPC or Target IS	Protect brand terms efficiently
Non-Brand Search (new)	Manual CPC → Max Conversions	Build data first
Non-Brand Search (mature)	Target CPA or Target ROAS	After 50+ conv/mo
Shopping	Target ROAS	Need conversion values
Display Remarketing	Target CPA	Strong when audience is warm
Display Prospecting	Max Clicks (with cap) or CPM	Hard to hit CPA targets
YouTube Awareness	Target CPM	Volume over efficiency
YouTube Conversion	Target CPA	Video Action campaigns
PMax	Max Conversions → Target ROAS	ROAS after learning period

广告系列类型	推荐策略	说明
品牌搜索	Manual CPC 或目标展示份额	高效保护品牌词
新非品牌搜索	Manual CPC → 最大化转化	先积累数据
成熟非品牌搜索	Target CPA 或 Target ROAS	每月50次以上转化后使用
购物广告	Target ROAS	需要转化价值
展示再营销	Target CPA	暖受众表现好
展示拓新	最大化点击（带上限）或 CPM	很难达到CPA目标
YouTube品牌曝光	目标CPM	优先量级而非效率
YouTube转化	Target CPA	视频行动广告系列适用
PMax	最大化转化 → Target ROAS	学习期结束后切换到ROAS模式

CPC Benchmarks by Industry

各行业CPC基准

Industry	Average CPC	High-Value Keywords
Legal	$5-$50	$50-$300
Finance / Insurance	$5-$40	$30-$100
B2B SaaS	$3-$15	$10-$50
E-commerce	$0.50-$3	$3-$10
Healthcare	$2-$20	$10-$50
Education	$2-$10	$5-$20

行业	平均CPC	高价值关键词
法律	$5-$50	$50-$300
金融/保险	$5-$40	$30-$100
B2B SaaS	$3-$15	$10-$50
电商	$0.50-$3	$3-$10
医疗健康	$2-$20	$10-$50
教育	$2-$10	$5-$20

Budget Allocation

预算分配

Budget Distribution Rule of Thumb

预算分配经验法则

Brand campaigns: 10-20% of budget (high ROI, protect)
Non-brand / core: 50-60% of budget (primary driver)
Remarketing: 10-15% of budget (high CVR)
Prospecting / awareness: 10-20% of budget (top-of-funnel)

品牌广告系列: 预算的10-20%（高ROI，需保护）
非品牌/核心: 预算的50-60%（核心增长动力）
再营销: 预算的10-15%（高转化率）
拓新/曝光: 预算的10-20%（漏斗上层）

Budget Changes and Learning

预算调整与学习期

Changing budget by >20% triggers learning period for Smart Bidding
For large increases: scale up 15-20% per week
For decreases: cut more aggressively — recovery is faster than growth

预算调整超过20%会触发Smart Bidding的学习期
大幅提预算：每周提升15-20%
降预算：可以更大幅度调整，恢复速度比增长快

Shared Budgets

共享预算

Use shared budgets for tightly related campaigns
Allows budget to flow to best-performing campaign automatically
Do NOT combine campaigns with very different economics in one shared budget

适用于关联性极强的广告系列
允许预算自动流向表现最好的广告系列
不要将经营数据差异极大的广告系列放在同一个共享预算中

Optimization Checklist

优化清单

Weekly

每周

Campaign bid strategy status — any "Limited" or "Eligible (limited)"?
Learning period status — any campaigns stuck in learning?
CPA/ROAS vs. target — on track or drifting?
Impression share: budget-limited or rank-limited?

广告系列竞价策略状态——是否有"Limited"或"Eligible (limited)"？
学习期状态——是否有广告系列卡在学习期？
CPA/ROAS与目标对比——是否达标或出现偏差？
展示份额：受预算限制还是排名限制？

Monthly

每月

Bid adjustment review: device, location, time, audience
Budget allocation — shift to higher-performing campaigns
Target review — should CPA/ROAS targets be tightened or loosened?
Portfolio strategy performance vs. individual campaign strategies

出价调整审核：设备、位置、时段、受众
预算分配——向表现更好的广告系列倾斜
目标审核——CPA/ROAS目标应该收紧还是放宽？
组合竞价策略表现与单个广告系列策略的对比

Quarterly

季度

Strategy review — are campaigns mature enough for more automation?
Seasonality adjustments — apply seasonal bid adjustments proactively
LTV-based target recalculation — do targets reflect current unit economics?

策略审核——广告系列是否足够成熟，可以使用更高程度的自动化？
季节性调整——提前应用季节性出价调整
基于LTV的目标重算——目标是否反映当前的单位经营数据？

Common Mistakes

常见错误

Target Setting

目标设置

Setting tCPA far below actual CPA immediately (starves volume, never recovers)
Using tROAS without conversion values assigned
Not giving enough learning time before changing targets

刚上线就将tCPA设置为远低于实际CPA（会限制曝光量，难以恢复）
未配置转化价值就使用tROAS
修改目标前没有留出足够的学习时间

Learning Period

学习期

Changing campaigns weekly during learning phase
Making large budget cuts that force re-learning
Switching strategies repeatedly (each switch = reset)

学习期内每周修改广告系列设置
大幅削减预算导致重新进入学习期
反复切换策略（每次切换都会重置学习期）

Bid Adjustments

出价调整

Using device bid adjustments with Smart Bidding (mostly ignored)
Applying audience bid adjustments without enough data (act on noise)
Setting location exclusions without looking at 90-day data

在Smart Bidding模式下使用设备出价调整（基本会被忽略）
没有足够数据就应用受众出价调整（基于噪音做出决策）
未查看90天数据就设置位置排除