Understanding the core concepts behind automated and manual bids
Paid‑search platforms offer both automated and manual bidding models. Automated models such as target CPA (tCPA) and target ROAS (tROAS) rely on machine learning to adjust bids in real time, while manual CPC gives the advertiser full control over each keyword’s maximum cost per click. Knowing the data requirements, conversion goals and budget constraints for each model is essential before any switch is made.
How target CPA (tCPA) works in practice
tCPA tells the system to aim for an average cost per acquisition that matches the value you set. The algorithm evaluates dozens of signals – device, location, time of day and user intent – and raises or lowers bids to stay within the target. The system does not guarantee that every click will cost exactly the target amount; instead it optimises across the entire campaign to meet the average.
Key requirements for tCPA include a reliable conversion tracking setup and a sufficient volume of conversion data. Google recommends at least 30 conversions in the past 30 days for the algorithm to learn effectively. Without that data the system may over‑bid or under‑bid, leading to wasted spend or missed opportunities.
When tCPA delivers the strongest performance
tCPA shines when the primary business goal is to acquire customers at a predictable cost. E‑commerce stores with a clear profit margin per order, subscription services with a known acquisition cost, or lead‑generation businesses with a fixed cost per qualified lead all benefit from the predictability of tCPA. The model also works well when the campaign targets a single conversion action, such as a purchase or sign‑up, rather than multiple disparate actions.
Because the algorithm focuses on the cost per acquisition, it may sacrifice some overall volume in favour of staying within the target. Marketers should monitor both the average CPA and the total number of conversions to ensure that the reduced cost does not overly limit growth.
Decoding target ROAS (tROAS) and its optimisation logic
tROAS extends the principle of tCPA by targeting a specific return on ad spend ratio rather than a fixed cost per acquisition. Advertisers set a target percentage – for example 400 % – and the system aims to generate revenue that is four times the spend. The algorithm evaluates the estimated revenue of each click, using historical conversion values and real‑time signals, and adjusts bids accordingly.
Accurate revenue tracking is a prerequisite for tROAS. This means implementing enhanced e‑commerce tracking, passing transaction values to the ad platform, and ensuring that refunds or cancellations are reflected in the data. Inaccurate revenue data will mislead the algorithm and can cause sub‑optimal bid adjustments.
Ideal scenarios for applying tROAS
tROAS is most effective when the advertiser’s objective is to maximise profit rather than simply acquire users. Brands with high average order values, tiered pricing models or variable profit margins benefit from the revenue‑centric approach. The model also supports campaigns that aim to drive multiple conversion types with different values, as long as each conversion’s monetary value is correctly reported.
One limitation of tROAS is its sensitivity to revenue volatility. Seasonal spikes, promotions or sudden price changes can distort the algorithm’s predictions. In such periods, marketers may need to pause the automated strategy or adjust the target ROAS to reflect the new reality.
Manual CPC: retaining full control over each bid
Manual CPC places bid decisions entirely in the advertiser’s hands. Each keyword or ad group receives a maximum cost per click that the platform will not exceed. This approach is useful when the marketer has deep knowledge of keyword performance, competitive landscape or margin constraints that the algorithm may not capture.
Because bids are static, manual CPC requires regular monitoring and optimisation. Changes in competition, user behaviour or landing‑page relevance can quickly render a previously successful bid too high or too low. The workload can be significant for large accounts with many keywords.
When manual CPC is the preferred choice
Manual CPC is advantageous in highly specialised niches where conversion data is scarce or where the advertiser wants to test granular bid variations. It is also suitable for brand‑protected campaigns that require tight budget control or for advertisers who need to align bids with offline inventory limits.
Another common use case is the early stage of a new product launch. Before enough conversion data accumulates to feed an automated model, manual CPC allows the team to gather performance signals while maintaining cost discipline.
Transitioning between strategies: a step‑by‑step framework
Moving from manual CPC to an automated model should be treated as a controlled experiment. First, audit the existing conversion tracking to ensure all relevant actions are captured and that value data is accurate. Second, identify the primary business goal – cost per acquisition or return on spend – and choose the corresponding automated strategy.
Next, create a duplicate of the high‑performing campaign, switch the bidding type, and set a conservative target that reflects historical performance. Monitor the key metrics for at least two weeks, allowing the algorithm time to learn. If performance diverges significantly, pause the automated campaign and revert to manual CPC while refining the data inputs.
Hybrid approaches and advanced considerations
Some advertisers combine strategies within a single account. For example, they may run tCPA campaigns for high‑margin products and manual CPC for low‑margin items where precise control is essential. Platforms also offer portfolio bidding, allowing multiple campaigns to share a common target CPA or ROAS, which can smooth out data sparsity across smaller campaigns.
Advanced users may incorporate bid adjustments based on audience segments, device types or ad schedules, even when using automated bidding. These adjustments act as additional signals for the algorithm, fine‑tuning performance without overriding the core optimisation goal.
Key decision criteria to choose the right model
Before selecting a bidding strategy, ask the following questions: Do you have at least 30 conversions in the past month for the targeted action? Is the monetary value of each conversion reliably tracked? What is the primary KPI – cost per acquisition or profit margin? How much time can you allocate to ongoing bid management? The answers will point you toward tCPA, tROAS or manual CPC.
Remember that no single model is universally superior. Successful advertisers treat bidding strategy as a lever that can be adjusted as business goals evolve, data quality improves and market conditions shift.
Practical tips for ongoing optimisation
Regularly review the “search term” report to identify new keywords that may need separate bids. Refresh conversion values when product pricing changes. Use the platform’s bid simulation tools to understand how different targets would affect spend and volume. Finally, keep an eye on the “learning” status – campaigns that are still in the learning phase may exhibit volatile performance and should not be judged prematurely.
By aligning the chosen bidding model with clear business objectives, reliable data and a disciplined optimisation routine, marketers can turn the complex world of paid‑search bidding into a predictable engine for growth.
Leave a Reply