Target CPA (tCPA) asks Google Ads to generate as many conversions as possible at an average cost per action target. Target ROAS (tROAS) asks Google Ads to generate as much conversion value as possible at an average return on ad spend target.
For lead generation, the right choice depends on the answer to one question:
Should the campaign treat each qualified conversion as roughly equal, or should it bid differently based on expected business value?
Use Target CPA when a conversion has a consistent definition and similar economics. Use Target ROAS when lead value varies materially and those values are reported accurately enough to guide bids.
Target CPA vs Target ROAS at a glance
| Decision factor | Target CPA | Target ROAS |
|---|---|---|
| Primary objective | Conversion volume at an average CPA goal | Conversion value at an average ROAS goal |
| Input required | Reliable conversion actions | Reliable conversion actions and differentiated values |
| Best lead-gen fit | Similar value per qualified lead | Meaningfully different lead or deal values |
| Main target | Cost / conversions | Conversion value / cost × 100% |
| Main risk | Cheap low-quality conversions satisfy the target | Bad or inflated values misdirect bidding |
| Better KPI | Cost per qualified outcome | Qualified value, profit value, or expected revenue per ad dollar |
Google Ads restored the separate Target CPA and Target ROAS labels for Search campaigns beginning in June 2026. Accounts still in transition may show the older labels “Maximize Conversions with a target CPA” and “Maximize Conversion Value with a target ROAS.” Google says the underlying bidding behavior did not change; see its current bid-strategy selection guide.
How Target CPA works for lead generation
Target CPA sets auction-time bids based on the likelihood that an interaction will produce a conversion included in the campaign’s bidding goal. The target is an average, not a ceiling for every conversion.
If you set a $150 tCPA, some conversions may cost $90 and others $220. Google attempts to bring the aggregate cost per conversion toward the target while generating as much volume as conditions allow.
That can work well when the conversion is a qualified lead, booked appointment, or another consistently defined stage. It works poorly when the goal contains outcomes with different business importance.
For example:
| Conversion mix at a $100 platform CPA | Count | Qualified outcomes |
|---|---|---|
| Campaign A: mostly valid prospects | 20 | 14 |
| Campaign B: mostly weak calls and forms | 20 | 5 |
Both campaigns report the same platform CPA. Campaign A’s cost per qualified outcome is about $143; Campaign B’s is $400. Target CPA cannot optimize for the difference if both outcomes are sent under the same biddable conversion definition.
How Target ROAS works for lead generation
Target ROAS predicts both the likelihood of conversion and its reported value. It then sets bids to maximize conversion value while trying to reach the average return target.
The formula is:
ROAS = conversion value / ad cost × 100%
If a campaign spends $10,000 and reports $40,000 in valid conversion value:
$40,000 / $10,000 × 100% = 400% ROAS
This is financial ROAS only when the reported values represent revenue. If the account uses proxy values, such as 10 for a weak lead and 100 for a strong lead, the calculation is a value/cost efficiency index, not literal dollars returned.
Google describes Target ROAS as maximizing conversion value while trying to achieve the return you set in its Target ROAS documentation.
When Target CPA is the stronger choice
Choose tCPA as the more defensible starting point when:
- The primary conversion already represents a qualified, economically similar outcome.
- Sales values are unavailable, delayed for months, or updated inconsistently.
- Most customers buy the same service or have a narrow value range.
- The account needs a cost-per-acquisition guardrail more than a value hierarchy.
- The current value model is based on opinion rather than observed close rates and economics.
Target CPA still depends on clean tracking. A precise target on an imprecise conversion goal simply automates the wrong objective more efficiently.
When Target ROAS is the stronger choice
Choose tROAS as a serious candidate when:
- Lead types have different close rates, ticket sizes, margin, or lifetime value.
- Qualified phone calls can be separated from existing customers, spam, and wrong-service calls.
- Values are reported consistently for most eligible conversions.
- Conversion lag fits the campaign’s decision cycle and reporting window.
- You can distinguish actual revenue, expected revenue, and proxy quality values in reporting.
Google’s general conversion-value documentation supports transaction-specific values when individual leads or sales differ. The critical word is specific: a value should describe the outcome that occurred, not a blanket number applied to every submission.
Calculate a defensible Target CPA
Start from the deepest business stage you can measure reliably.
Method 1: from allowable customer acquisition cost
If the business can spend up to $1,200 to acquire a customer and 20% of qualified leads become customers:
Allowable cost per qualified lead = $1,200 × 20% = $240
A tCPA near $240 is an economic ceiling estimate, not an automatic launch target. Compare it with the campaign’s recent CPA and expected volume. A target far below attainable performance can restrict traffic.
Method 2: from gross profit
Suppose:
- Average collected revenue: $8,000
- Gross margin before advertising: 45%
- Gross profit before advertising: $3,600
- Maximum share of that gross profit available for acquisition: 30%
Allowable customer acquisition cost = $3,600 × 30% = $1,080
With a 25% qualified-lead-to-customer rate:
Allowable qualified-lead CPA = $1,080 × 25% = $270
Use real collected revenue, refunds, fulfillment cost, sales cost, and cash-flow constraints where they materially affect the result.
Calculate a defensible Target ROAS
The target depends on what the conversion value represents.
If values are revenue
If the business can spend 20% of revenue on advertising:
Break-even target ROAS = 1 / 20% = 5.0, or 500%
That is only a simplified boundary. Gross margin, overhead, refunds, sales expense, repeat revenue, and desired profit all affect the real target.
If values are expected revenue
For a qualified lead type with:
- 20% close rate
- $8,000 average collected revenue
Expected revenue per qualified lead = 20% × $8,000 = $1,600
For a stronger lead type with a 35% close rate:
35% × $8,000 = $2,800
Reporting $1,600 and $2,800 makes the value difference explicit. Recalculate periodically as close rates and revenue change.
If values are proxy scores
You can use a relative scale when revenue is not available. For example, use 20 for a valid inquiry, 60 for a qualified prospect, and 100 for a booked appointment.
Do not call the resulting percentage revenue ROAS. Track it as conversion value/cost, then validate whether more proxy value corresponds with more accepted and profitable business.
A lead-gen decision matrix
| Measurement condition | Recommended direction |
|---|---|
| Every raw call or form is a primary conversion | Fix conversion goals before choosing tCPA or tROAS |
| One stable qualified-lead action; similar value | Test Target CPA |
| Multiple lead stages with stable relative values | Consider Target ROAS |
| Transaction-specific revenue or expected profit is reliable | Strong Target ROAS candidate |
| Values cover only a small or biased subset | Keep values observational; use a cleaner conversion objective |
| Sales cycle is long and uploads are inconsistent | Improve feedback coverage and delay before testing tROAS |
The phone-call problem: count is not quality
Call-heavy lead generation exposes the main weakness of an unqualified tCPA goal. A call can be:
- A first-time prospect with a supported need
- A booked appointment
- An existing customer asking for support
- A vendor solicitation
- A wrong location or wrong service
- Spam
If they all enter Google Ads as one primary Phone Call conversion, tCPA is rewarded for the whole mix.
CallRail sends post-call data, including the call recording and tracking details, to LeadUp through a webhook. LeadUp then produces its own transcript and summary, qualifies the lead, and assigns a one-to-five-star score.
LeadUp’s CallRail conversion tracking uses a relative proxy value based on caller type, lead status, and lead rating. You can set different caller-type base values for first-time and repeat callers, different lead-status base values for good and bad leads, and a configurable per-star value. You can also choose whether Google Ads receives only good leads or both good and bad leads, keep first-time and repeat caller conversions separate, and manually adjust a pushed value afterward.
How to move from Target CPA to Target ROAS
1. Keep the existing campaign objective stable
Do not switch bidding on the same day you introduce an untested value model. Start sending values while the current strategy continues to bid on the established conversion goal.
2. Validate value coverage
For every eligible conversion, check:
- Was an outcome received?
- Was a value attached?
- Was it attributed to the correct action and campaign?
- Did duplicate stages inflate total value?
- Does a zero mean “no value,” “not reviewed,” or “technical failure”?
Those states should not share one number.
3. Wait through the reporting cycle
Google recommends waiting four weeks or three conversion cycles after newly reporting or materially changing conversion values before adopting Maximize Conversion Value. Use the same caution before relying on those values for tROAS. The goal is a representative history, not a magic volume threshold.
4. Set an attainable initial target
Use recent, delay-adjusted value/cost performance as a reference. Do not set the target from the return the business wishes it had. An aggressive target can reduce auction participation and volume.
5. Test and judge downstream outcomes
Where available, use an experiment. Compare:
- Total and qualified lead volume
- Cost per qualified lead
- Conversion value and value/cost
- Appointment, opportunity, and customer rate
- Expected and collected revenue
- Spend distribution by query, audience, location, device, and campaign
Let the test cover the full conversion delay. Google notes that imported outcomes extend the conversion cycle by the time it takes to report them; review its explanation of how bidding algorithms learn.
Common tCPA and tROAS mistakes
- Using tCPA with a raw-lead goal: the cheapest lead sources can win even when qualification collapses.
- Using tROAS with arbitrary values: bidding amplifies the assumptions in the value model.
- Setting targets from wishful economics: an unattainable CPA or ROAS can suppress useful traffic.
- Comparing incomplete recent data: later-stage outcomes have not all returned yet.
- Treating the target as a guarantee: both targets are averages pursued under changing auction conditions.
- Changing targets every few days: repeated interventions make results difficult to interpret.
- Calling proxy value “revenue”: stakeholders may mistake a quality index for money earned.
- Leaving micro-conversions in the goal: a high volume of easy actions can overwhelm the desired signal.
Frequently asked questions
Is Target ROAS better than Target CPA?
Not universally. tROAS is more expressive when conversion values vary and are reliable. tCPA is simpler and often more robust when a qualified conversion has consistent value.
Can Target ROAS work for service businesses?
Yes. Service businesses can report actual closed revenue, expected lead value, profit-based value, or validated relative values. The measurement must be maintained and returned consistently.
Does Target CPA ignore conversion value?
Target CPA optimizes toward conversion volume at the CPA target. Values can still appear in reporting, but a conversion-value strategy is the appropriate objective when value differences should directly guide bids.
What is the difference between Target ROAS and Maximize Conversion Value?
Both optimize for conversion value. Maximize Conversion Value pursues the most value within budget; Target ROAS adds an average return goal. See our full Maximize Conversions vs Maximize Conversion Value comparison.
Bottom line
Use Target CPA when the account has one trustworthy definition of a qualified conversion and the value of those conversions is reasonably consistent. Use Target ROAS when value varies enough to matter and the account can report that difference accurately.
For lead generation, better bidding starts downstream. Define a qualified outcome, calculate its economics, return the result to Google Ads, and only then choose the target that matches the data.