Maximize Conversions vs Target CPA: the right starting point for your bidding strategy

One of the most consequential decisions in Google Ads is choosing the right bidding strategy at the right moment. Yet this is precisely where many advertisers go wrong: they launch directly with Target CPA, see disappointing results, and conclude that Smart Bidding does not work. The reality is more nuanced. Maximize Conversions and Target CPA are not competing strategies. They are sequential steps in a well-structured Smart Bidding journey. The difference is not about which strategy is superior, but about which strategy is appropriate given the available data and the campaign objective at a specific point in time.

What is Maximize Conversions?

Maximize Conversions is a Smart Bidding strategy in which Google Ads automatically sets bids with the goal of generating the maximum number of conversions within a given daily budget. There is no explicit CPA target: the algorithm focuses purely on volume. Google uses signals such as device type, time of day, location, search query, and user behaviour to determine the optimal bid in each individual auction.

This strategy is particularly valuable in situations where conversion data is limited or entirely new. Smart Bidding needs historical conversion patterns to steer effectively. Without that data, Target CPA acts like a blindfold: the system tries to hit a cost target without the context to do so reliably. Maximize Conversions does not have this problem. It simply collects as many conversions as possible within the available budget, building exactly the data that is later needed for tCPA steering.

• No minimum conversion threshold required to get started
• Full utilisation of the daily budget
• Ideal for new campaigns, new accounts or new product categories
• Also works well for seasonal campaign re-launches
• Accelerates the Smart Bidding algorithm's learning period
• Provides insight into the realistically achievable CPA level

A concrete example: suppose ToetsJeKennis.nl launches a new Google Ads campaign for their online exam preparation courses. The account has no historical conversion data for this campaign type. By starting with Maximize Conversions, ToetsJeKennis.nl allows the Smart Bidding algorithm to learn freely: which keywords lead to purchases? At what times does the target audience convert best? Only after enough data has been collected does it make sense to set a Target CPA. This prevents the system from being constrained too early and suppressing its own performance.

What is Target CPA?

Maximize Conversions vs Target CPA

When each strategy works best

Maximize Conversions

• New campaign or account (<30 conversions/month)
• No historical data available
• Goal: accelerate the learning period
• Budget is fully utilized
• Ideal starting point for Smart Bidding
• No CPA control, pure volume focus

Target CPARecommended after learning period

• At least 30-50 conversions per month available
• Stable historical conversion data
• Goal: efficient cost-per-conversion control
• Budget may go unspent if CPA target is too low
• Ideal after a successful learning period
• AI actively steers toward desired cost per conversion

Target CPA (tCPA) is a Smart Bidding strategy where you set a desired cost per conversion and Google Ads automatically adjusts bids to reach that goal as closely as possible. The algorithm will deliberately skip certain auctions when the expected CPA is too high, and bid more aggressively when the probability of conversion is high at a low cost. The goal is to keep the average cost per conversion across all auctions as close to the set target as possible.

Target CPA is the ideal strategy for accounts with sufficient conversion history. Google recommends having at least 30 conversions per month before switching, but in practice 50 or more conversions per month provides a more stable and reliable foundation for tCPA steering. With too little data, the system risks overreacting: small fluctuations in conversion patterns then lead to large bidding adjustments that destabilise campaign performance.

For a lead generation client like Clima-Active.nl, active in the installation of air conditioning and heat pumps, Target CPA is a powerful instrument. Once enough quote requests have been recorded, the algorithm can precisely learn which type of search query, at which time of day and via which device leads to high-quality leads. Bidding is then fully automated and aligned with the desired cost per quote request.

The learning period: the most underestimated factor in Smart Bidding

Both Maximize Conversions and Target CPA use a learning period. During this phase, the Smart Bidding algorithm analyses incoming conversion data and updates its predictive models. Campaigns in the learning period often temporarily perform less consistently, which advertisers mistakenly interpret as a sign that the strategy is not working.

The learning period is typically triggered by significant changes in a campaign, such as changing the bidding strategy, altering budgets, adding new ad groups, or adjusting conversion settings. It is therefore important not to make large changes during the learning period, as each adjustment partially resets the learning curve.

• Average learning period: 1 to 6 weeks, depending on conversion volume
• Higher conversion volume means a faster learning period
• Avoid major budget or structural changes during the learning period
• Adjusting Target CPA by more than 15-20% at once triggers a new learning period
• Combine Maximize Conversions with a competitive budget to collect data faster

This is exactly where many advertisers make mistakes. They see the learning period status and intervene by changing the bidding strategy, adjusting budgets or restructuring ad groups. This sabotages the learning process and unnecessarily prolongs the unstable phase. Patience and discipline are more important than action during this period.

From Maximize Conversions to Target CPA

Drop-off per Smart Bidding maturity phase (percent-mode)

100% Campaign live (Maximize Conversions)

Starting point: all new campaigns begin here

68% Learning period complete (sufficient data)

About 68% reach the data threshold within 6 weeks

41% Switch to Target CPA

41% successfully switch after AI recommendation

28% Stable CPA within target

28% achieve stable CPA within 3 months

28% achieve a stable Target CPA within 3 months of campaign launch

When to switch from Maximize Conversions to Target CPA

The transition from Maximize Conversions to Target CPA is not a fixed date but a data-driven decision point. The right moment to switch is when the campaign consistently achieves 30 to 50 conversions per month and the historical CPA data is stable enough to set a realistic tCPA target. A common mistake is setting a tCPA significantly lower than the historically achieved CPA. This forces the algorithm to bid too conservatively, leaving campaign budget unused and causing conversion volume to collapse.

A healthy rule of thumb: set the initial Target CPA at the level of the average historical CPA from the Maximize Conversions phase, or no more than 10 to 15% lower. The tCPA can then be gradually reduced in small steps, as long as conversion volume remains stable. This is sometimes called the "step-down method" and is a proven approach for improving efficiency without triggering a new learning period.

For E-4motion.com, the webshop for new electric folding bikes, this pattern is clearly recognisable. When a new bike model is introduced, the campaign starts with Maximize Conversions to quickly learn which search terms and audiences are most relevant. Once the conversion data shows a stable pattern, the campaign switches to Target CPA with an initial tCPA target that realistically reflects the historical data. This protects the campaign budget and steers toward a healthy return.

How AdBrains automates this with AI

At AdBrains, we have developed proprietary AI technology that monitors and automatically manages the entire bidding strategy cycle. This goes far beyond manually monitoring conversion figures and occasionally adjusting a tCPA value. Our AI works continuously and makes decisions based on multiple signals simultaneously.

For new campaigns, the process begins with our Keyword Incubator: new keywords are first safely tested in a separate incubator campaign, linked to a Maximize Conversions strategy. This protects existing production campaigns from the instability of the learning period, while simultaneously collecting controlled data. Once a keyword or ad group reaches the data threshold, our AI automatically promotes it to the production campaign.

Our automatic tCPA/tROAS optimisation monitors conversion volume per campaign on a daily basis. The system detects when a campaign is ready to switch from Maximize Conversions to Target CPA, and then proposes an initial tCPA based on the average historical CPA of the past four weeks, including a configurable safety margin. This ensures a smooth transition without unnecessary learning period resets.

A critical module is our multi-agent verification system: every bidding strategy adjustment is independently reviewed by four AI agents before the change is implemented. This prevents a single data error or anomaly from triggering an incorrect strategy switch. If one of the agents deviates from the recommendation, the decision is deferred and flagged for review. This level of built-in control is impossible to replicate with manual campaign management.

Additionally, our strategy-switch system automatically returns campaigns to Maximize Conversions when conversion volume drops too sharply, for example due to seasonal influences or budget changes. Once volume recovers, the system reactivates the switch to Target CPA at the right moment and with the correct tCPA value. For lead generation clients such as Clima-Active.nl and LeroyBrouwer.nl, this ensures a consistent quality of incoming leads regardless of seasonal fluctuations in demand. Combined with our server-side signal enrichment via our own sGTM infrastructure, this creates a self-reinforcing system: better data leads to better bidding decisions, which in turn generates better results for both Maximize Conversions and Target CPA campaigns.

Smart Bidding: impact of the right starting point

Average benchmarks for Google Ads campaigns in 2026

+34%

more conversions when correctly starting with Maximize Conversions

-22%

lower CPA after switching to Target CPA at the right moment

6 wk

average learning period before a switch is justified

50+

conversions/month recommended threshold for Target CPA

Common mistakes when choosing a bidding strategy

The theory is clear, but practice shows that bidding strategy errors are widespread. The most common pitfalls are:

• Starting directly with Target CPA without sufficient conversion data: the algorithm lacks a reference point and bids either too conservatively or too aggressively.
• Setting a tCPA that is too low at the switch: the system cannot meet the target and leaves budget unused, causing conversion volume to collapse.
• Intervening during the learning period: every significant change resets the learning period, preventing campaigns from ever reaching full maturity.
• Reducing tCPA too aggressively in one step: never change the tCPA by more than 15-20% at a time, as this triggers a new learning period.
• Keeping Maximize Conversions too long without a CPA target: once sufficient data is available but no switch is made, budget may be spent inefficiently.
• Measuring the wrong conversion actions: if the conversions being tracked are not representative of real business value, Smart Bidding steers on the wrong signal.

The quality of conversion tracking is the foundation of every Smart Bidding system. Whether you use Maximize Conversions or Target CPA, both strategies are entirely dependent on accurate, timely and relevant conversion signals. Incomplete or delayed conversion data inevitably leads to suboptimal bidding decisions.

FAQ: Maximize Conversions vs Target CPA

Can I start directly with Target CPA for a new campaign?

Technically yes, but it is strongly discouraged. Target CPA requires historical conversion data to steer effectively. Without this data, the algorithm lacks the context to make well-founded bidding decisions. The system will either bid too conservatively, leaving budget unused and generating few conversions, or too aggressively, resulting in a high CPA and inefficient budget use. The recommended starting point is always Maximize Conversions, followed by a data-driven switch to Target CPA once the campaign consistently achieves at least 30 to 50 conversions per month.

How do I know when my campaign is ready for Target CPA?

There are three signals that indicate a campaign is ready for the switch to Target CPA. First, the campaign consistently achieves 30 to 50 conversions per month. Second, the historical CPA has been stable over the past four to eight weeks without major outliers. Third, the learning period status is no longer active and the campaign is running steadily. When all three conditions are met, it is safe to set an initial Target CPA at the level of the average historical CPA from the Maximize Conversions phase.

What is the difference between Target CPA and Target ROAS?

Target CPA steers on cost per conversion: the algorithm tries to pay a fixed amount per conversion, regardless of the value of that conversion. Target ROAS (tROAS) steers on the ratio between revenue and advertising costs, factoring in the expected conversion value in the bidding decision. For lead generation campaigns without variable lead values, Target CPA is the logical choice. For e-commerce with variable order values, such as ToetsJeKennis.nl or E-4motion.com, Target ROAS provides more nuanced steering because it factors in the value of each transaction.

What happens if I set my Target CPA too low?

If the set Target CPA is significantly lower than the historically achieved CPA, the Smart Bidding algorithm will struggle to win enough auctions that meet the cost target. The result is that the campaign budget remains partially or fully unspent, conversion volume drops sharply, and the campaign is effectively throttled. The recommendation is to lower the Target CPA gradually, by no more than 10 to 15% per adjustment, so the algorithm can adapt without going through a full learning period again.