Manual CPC vs Target CPA
Manual CPC wins for new accounts, low-conversion-volume verticals, and operators who actively manage. Target CPA wins for stable, high-volume accounts where 30+ conversions/month/campaign exist and the algorithm has enough signal.
Pawel "Vilo" Wilk — Google Ads specialist. The verdicts below reflect decisions I make on real accounts every week.
- Your campaign averages under 30 conversions/month — smart bidding lacks signal
- You're a new account or a fresh campaign without 14+ days of conversion history
- You're running tests and need predictable per-keyword bid behavior
- You're actively managing the account day-to-day and trust your judgment over the algorithm
- Your campaign produces 30+ conversions/month consistently
- You've had stable conversion tracking for 14+ days minimum
- You're willing to give the algorithm 2-4 weeks of learning before judging
- Your account-management capacity is limited and you'd rather have the algorithm optimize 24/7
Manual
Manual CPC
You set the maximum cost-per-click bid for every keyword. Google never adjusts bids automatically (Enhanced CPC adjusts within a small window).
tCPA
Target CPA
You set a target cost-per-acquisition. Google's smart-bidding algorithm adjusts bids in real time aiming to hit the target average across conversions.
Side-by-side comparison
| Dimension | Manual | tCPA | Winner |
|---|---|---|---|
| Per-keyword bid control | Full — you set every bid | None — algorithm controls all bids | Manual |
| Cost predictability | High — bid caps mean predictable max CPC | Low — CPCs spike or drop based on smart-bidding signals | Manual |
| Learning phase required | None — runs immediately | 2-4 weeks of accumulated conversion data before steady state | Manual |
| Minimum conversion volume | Works at any volume | 30+ conversions/month/campaign minimum for reliable performance | Manual |
| Optimization at scale | Manual review every few days; manager attention required | Algorithm runs 24/7 across all signals | tCPA |
| Real-time signal use (device, audience, time) | Limited — Enhanced CPC adjusts within a small range | Full — every signal weighted in real-time bid | tCPA |
| Reaction to market changes | Slow — requires manual bid updates | Fast — algorithm adjusts within hours | tCPA |
| Performance during conversion-tracking issues | Stable — bid logic doesn't depend on conversion data | Catastrophic — algorithm goes blind, bids erratically | Manual |
| Brand vs non-brand handling | Manually controlled — set low bids on brand, high on non-brand | Algorithm often over-bids on brand because it converts highest | Manual |
| Suitability for testing new ad copy | Clean test results — bid stays constant | Confounded — algorithm shifts bids around the test | Manual |
| Suitability for stable, mature campaigns | Workable but operator-time-heavy | Strong — algorithm captures signals operator can't see | tCPA |
| Required management attention | High — daily/weekly bid review | Low — set target, review weekly, intervene only on drift | tCPA |
Want this decision made on YOUR account, with real data?
Perfoads audits which campaign types and bid strategies are leaking spend in your account specifically — $49/mo per Google Ads account, cancel anytime.
When Manual CPC wins
- 1New campaign with no conversion history yet — start manual to gather data, switch to tCPA later.
- 2Low-volume account (under 30 conversions/month/campaign) where smart bidding lacks signal.
- 3Account with broken or recently-changed conversion tracking — never run smart bidding on bad data.
- 4Active testing of ad copy or landing pages where bid noise would confound test results.
- 5Brand-isolated campaigns where you want to keep brand CPCs low while non-brand campaigns absorb the smart-bidding lift.
When Target CPA wins
- 1Established e-commerce or lead-gen account with 60+ conversions/month/campaign and clean tracking.
- 2Account with high signal-to-noise ratio in conversion data (high-quality tracked conversions, no spam).
- 3Operator with limited day-to-day management bandwidth — the algorithm running 24/7 outperforms a part-time manager.
- 4Campaigns running across multiple geographies, devices, or audiences where manual signal-weighting is impractical.
- 5Mature campaigns past the 4-week learning period with stable target CPA performance — switching back to manual usually loses the gains.
PerfoAds Read: the verdict
The most common bid-strategy mistake I find on audits is switching too early or too late between Manual CPC and Target CPA. Switch too early — before 14 days of stable conversion tracking and at least 30 conversions of history — and Target CPA bids erratically because it lacks signal. Switch too late — leaving Manual CPC running on a mature 200-conversion-per-month campaign — and you leave 15-30% performance lift on the table by not letting the algorithm capture device, audience, time-of-day signals you can't manually weight at scale.
Conversion tracking integrity is the silent prerequisite that almost no Google Ads consultant emphasizes enough. Smart bidding (any flavor — tCPA, tROAS, Maximize Conversions) is only as good as its signal. If your conversion action fires on irrelevant events (form views instead of submissions, navigation events, micro-conversions you don't actually care about), tCPA optimizes toward whatever produces those events most cheaply. The result: a tCPA bid strategy that perfectly hits the target while delivering garbage leads. I audit accounts where the fix isn't "switch bid strategies," it's "fix the conversion action and re-run for 30 days."
The brand-vs-non-brand tCPA trap deserves explicit mention. Branded queries convert at 25-50%; non-brand at 3-8%. If you run tCPA across both in one campaign, the algorithm sees branded queries as cheap conversions and over-bids on them — at premium CPCs you'd never choose manually. The fix is structural: separate brand and non-brand into different campaigns, run brand on Manual CPC at low bids (your brand bid floor), and run non-brand on tCPA where the algorithm has room to actually work.
Practical recommendation for most accounts: start Manual CPC for the first 30 days of any new campaign to establish baseline performance and accumulate conversion data. Switch to Target CPA only after you have 30+ conversions, stable tracking, and at least one full week of consistent performance. Set the initial tCPA target at 110-120% of historical CPA — give the algorithm room to learn. Plan to revisit the target every 30 days; if performance degrades, switch back to Manual rather than chase the algorithm with target adjustments.
Frequently asked
When should I switch from Manual CPC to Target CPA?+
After your campaign has at least 30 conversions and 14 days of stable tracking history. Set the initial tCPA target at 110-120% of historical CPA so the algorithm has room to learn. Don't switch in the middle of a major change (new ads, new keywords, new tracking) — give one variable at a time room to settle.
Why does my Target CPA campaign perform worse than Manual CPC?+
Three usual causes. First, conversion tracking issues — tCPA bids on whatever signal you give it, so bad signal produces bad bids. Second, insufficient conversion volume (under 30/month). Third, brand and non-brand mixed in one campaign causing tCPA to over-bid on branded queries. Audit those three before declaring tCPA broken.
Is Maximize Conversions different from Target CPA?+
Yes. Maximize Conversions tries to spend the entire daily budget while maximizing conversion count, with no CPA target. Target CPA tries to hit an average CPA and won't over-bid even if budget remains. Maximize Conversions is reasonable for new campaigns gathering data; Target CPA is the steady-state strategy once you know your target CPA.
Should I use Target ROAS instead of Target CPA for e-commerce?+
Yes, almost always — provided you import conversions with revenue values. Target ROAS optimizes against revenue, which is what you actually care about. Target CPA optimizes against conversion count, which can mean optimizing for low-AOV purchases over high-AOV ones. The catch: Target ROAS requires reliable revenue tracking via the Merchant Center feed or e-commerce platform integration.
How long does Target CPA need to learn before performance stabilizes?+
Typically 2-4 weeks for steady-state performance. The first week is unpredictable (algorithm exploring); week 2-3 shows performance trending toward target; week 4 is usually stable. Don't judge tCPA in week 1 — many accounts panic-switch back to Manual during the learning phase and never see the post-learning performance.
Make this decision against YOUR account
Perfoads runs the comparison logic against your real Google Ads data — campaign type performance, bid-strategy efficiency, lead-quality signals — and surfaces the verdict specific to your account in 5 minutes.
Start the audit — $49/mo per accountAdditional accounts $29/mo each (up to 10) · Cancel anytime · 30-day guarantee