11 min read · Google Ads · Last updated July 2026
Quick answer: Target CPA tells Google to get as many conversions as possible at or below your specified cost-per-acquisition. It requires 30+ conversions in 30 days to work effectively, and your initial target should be set 20-30% above your current average CPA — then lowered gradually as data accumulates.
Introduction
Target CPA is the most commonly used smart bidding strategy for lead generation campaigns — and the most commonly misconfigured. The two most expensive mistakes: setting it too low too soon (Google can’t find enough qualifying auctions and impressions collapse), and switching to it before you have enough conversion data (it optimises toward noise instead of signal).
Get the setup right and Target CPA consistently outperforms manual bidding by 15-30% CPA reduction once the campaign matures.
What you’ll learn:
– How Target CPA works under the hood
– The exact data requirements before you switch
– How to set your initial target and lower it safely
– What to do when CPA is running above target
Table of Contents
- How Target CPA Works
- Data Requirements Before Switching
- Setting Your Initial Target
- The Learning Period
- Optimising After Learning
- When CPA Exceeds Your Target
- Target CPA for Different Business Types
- FAQ
How Target CPA Works
Target CPA uses Google’s machine learning to set bids for every single auction in real time. For each potential impression, Google predicts the probability that a click will result in a conversion — factoring in dozens of signals including time of day, device, location, browser, user’s recent search history, and audience membership.
It then calculates a bid: Max bid = Target CPA × Predicted Conversion Rate
If the predicted CVR for this specific user in this specific context is 5%, and your Target CPA is $50, Google bids $2.50. If CVR is 20%, Google bids $10. This happens billions of times per day.
Google’s algorithm averages performance over time — some conversions will cost more than your target, some less. The goal is the average CPA equalling your target over a meaningful sample size (typically 30+ conversions per month).
Data Requirements Before Switching
Google recommends 30 conversions in the last 30 days before switching to Target CPA. This is a minimum — 50+ produces more stable results.
Why this matters: with 30 conversions, Google has a meaningful sample of what a converting user looks like — their behaviour patterns, search context, device preferences. With 5 conversions, the sample is too small to generalise.
Before switching, also verify:
– Conversion tracking is recording genuine commercial actions (not page views or soft events)
– Your conversion window matches your typical sales timeline
– There are no duplicate conversions inflating your count
If you switch before reaching 30 conversions, the algorithm will optimise toward whatever signal it can find — which may not reflect your actual best customers.
Setting Your Initial Target
Never set your initial Target CPA at your ideal CPA. Set it at 20-30% above your current actual average CPA.
Example: Your account currently averages $75 CPA under Maximize Conversions. Set your initial Target CPA at $90-100. Once performance is stable for 4 weeks, lower to $80-85. Then to $70-75 after another stable month.
This “staircase” approach gives Google flexibility to find conversions while you gradually tighten the constraint.
Setting it too low: Google can’t find enough conversions at that price. Impressions drop. Budget stops spending. You panic and either abandon the strategy or raise the target beyond what you wanted.
Setting it too high: You’re overpaying but volume is good. Use this as your starting point, not your resting point.
The Learning Period
Every time you switch to or change Target CPA, a learning period begins. During learning, Google tests different bid levels and observes conversion outcomes. Performance can be erratic — sometimes worse than your previous baseline.
Learning period characteristics:
– Typically lasts 1-2 weeks or 50 conversions, whichever comes first
– Performance volatility is expected and normal
– Making major changes (new keywords, major budget changes, campaign structure) resets learning
– The “Learning” label appears in campaign status
What triggers a reset:
– Changing the Target CPA target by more than 20%
– Significantly changing campaign structure
– Changing bidding strategy
– Major budget changes (>50%)
– Adding or removing many keywords
Protect the learning period. Don’t make changes based on 5-day performance windows. Evaluate at 2-3 weeks minimum.
Optimising After Learning
Once learning completes and you have 4+ weeks of data under Target CPA:
Check actual CPA vs target. If average CPA is within 10-15% of your target (above or below), the strategy is calibrated well. If it’s more than 20% above target consistently, investigate why.
Lowering the target safely: Reduce by 10-15% at a time. Wait 3-4 weeks. Repeat. Never drop more than 20% in one change — it triggers another significant learning period.
Scaling while maintaining CPA: Increase daily budget. Target CPA aims to maintain your CPA regardless of budget level — but practical limits exist. If you’re already showing in most relevant auctions, more budget won’t produce proportionally more conversions.
Adding new keywords: Add to ad groups, don’t change campaign bidding. New keywords take time to accumulate history. Expect slightly elevated CPA for 2-3 weeks when adding significant new keyword sets.
When CPA Exceeds Your Target
Campaign is spending but CPA is higher than target. Diagnosis framework:
1. Target is too aggressive. If you recently lowered the target significantly, raise it back 15-20%. Let the algorithm re-stabilise.
2. Budget is too low. Target CPA works best when daily budget is at least 5-10x the target CPA. If your daily budget is $30 and Target CPA is $50, you’re budgeting for less than one conversion per day — the algorithm can’t optimise effectively.
3. Conversion quality issue. Are the “conversions” you’re tracking actually valuable? Soft conversions (page views, session duration) will produce poor CPA because Google optimises for lots of low-value events.
4. Seasonality or competition change. CPCs spike in peak seasons. Your Target CPA may need temporary adjustment during high-competition periods.
5. Negative keyword gaps. Irrelevant traffic inflating cost without conversions. Review Search Terms report for recent anomalies.
6. Landing page regression. Did your landing page change recently? Conversion rate drops directly translate to CPA increases.
Target CPA for Different Business Types
E-commerce with low AOV: Target CPA works if you have high conversion volume (100+/month). For products under $50, CPA targets are tight — margins don’t allow much. Use Target ROAS instead.
High-value services (law, finance, construction): Target CPA is ideal. Even 30 conversions/month is sufficient, and the value of each conversion justifies higher CPAs.
B2B with long sales cycles: Import offline conversions from CRM. Without this, Target CPA optimises for lead form submits — not actual sales — which can produce high volume but poor quality.
Local services (trades, healthcare): Target CPA works well when combined with call tracking. Use 60-second minimum call duration to filter quality calls.
FAQ
Q: Can I use Target CPA with broad match keywords?
A: Yes, and it’s a common advanced strategy. Broad match gives Google more signal diversity; Target CPA provides the guardrail. But this requires strong negative keywords and sufficient conversion history (50+ per month). Start with exact/phrase before adding broad.
Q: My Target CPA campaign stopped spending. What happened?
A: Target is likely too low. Google predicts most available clicks will cost more than your target CPA. Raise the target 20-30% until spending resumes, then lower gradually.
Q: Should each campaign have its own Target CPA or use portfolio bidding?
A: Campaign-level Target CPA works well when each campaign has 30+ monthly conversions. For campaigns with fewer conversions, combine them in a portfolio bid strategy so the aggregate data meets the threshold.
Q: How long before I see Target CPA outperform Manual CPC?
A: 6-12 weeks typically. First 2 weeks: learning period. Weeks 2-6: calibration and stabilisation. Weeks 6-12: full optimisation. Judge fairly against a proper baseline.
Q: Can I set different CPAs for different times of day?
A: Not directly with Target CPA — that’s a manual bidding feature. Target CPA automatically adjusts bids by time, day, device, and audience signals. If certain hours perform badly, use ad scheduling to restrict hours.
Conclusion
Target CPA is one of Google’s most powerful bidding strategies — when fed with the right data and given appropriate targets. The discipline required is patience: don’t set targets too aggressively, don’t make changes during learning, and don’t judge performance after 5 days.
Build the data foundation first, set generous initial targets, and gradually tighten. Accounts that follow this protocol consistently achieve better CPAs than those that try to shortcut the process.
Work With Ignited Nepal on Google Ads
Our Google Ads team manages campaigns across Australia, UAE, Nepal, USA, and UK — focused on verified conversions, not just clicks.
Written by the Ignited Nepal paid acquisition team. ignitednepal.com