A quick primer: how tCPA compares to other bidding strategies
Google Ads offers a few fundamentally different approaches to setting bids:
- Manual CPC / Enhanced CPC — you set the bid on each keyword or ad group yourself; Enhanced CPC layers on a modest automatic adjustment up or down in the moment of the auction. It's the most transparent form of control, but it doesn't scale: a human can't account for the thousands of contextual auction signals (device, time, location, user history, and so on) the way automation can.
- Maximize Clicks — optimizes for click volume within a set budget, with no regard for conversions at all. Good for reach goals, not for managing the price of a result.
- Maximize Conversions / Maximize Conversion Value — the system maximizes the number of conversions (or their total value) within a set budget, but with no cap on cost per conversion — meaning it can let CPA drift upward if it judges that trade-off worthwhile for more volume.
- tCPA (Target CPA) — the same conversion-maximizing logic, but with an explicit constraint: the system aims to keep bids around your stated cost per conversion. Of all these strategies, this is the only one that gives you a direct lever over the price of the result, not just the volume.
- tROAS (Target ROAS) — the same idea as tCPA, but expressed as return on ad spend instead of cost per conversion (for e-commerce and other scenarios where conversions carry different values).
The key thing that sets tCPA apart from both manual bidding and Maximize Conversions is that it gives you predictable control over your unit economics: you set the acceptable cost of acquisition up front, and the strategy is required to work within it (as far as the auction allows) — rather than simply "trying to find as many conversions as possible" or requiring you to constantly recalibrate bids by hand.
Straight recommendation: if a campaign has enough conversion volume (a reliable ~15–30+ per month) and the conversion data is trustworthy (no tracking issues, duplicates, or attribution problems) — you should switch to tCPA. This isn't a "weigh the pros and cons" situation; staying on manual bidding is the exception, not the safer default. Manual bid management almost always loses to automation when it comes to reading real-time auction signals, and its main practical advantage — full transparency and control — doesn't outweigh that gap once you have enough data. The only real reasons to stay on manual bidding are insufficient conversion volume to learn from, or unreliable conversion data. Outside of those, moving to tCPA (not Maximize Conversions without a target) is what actually gives your business control over the price of the result.
What to check before switching
- Do you have enough conversion history under manual bidding (typically a reliable 15–30+ conversions over the last 30 days)? That's the minimum base smart bidding needs to learn predictably.
- How stable are your current results under manual bids? If CPA is already bouncing around unpredictably with no clear tie to your bid changes, that may point to data instability that will carry over into tCPA too.
- Is your conversion tracking clean and reliable (no obvious tracking gaps, duplicates, or attribution errors)? Switching to smart bidding on a CPA target only makes sense once the underlying conversion data is trustworthy — otherwise the system will be learning from a distorted signal.
- Is the account/campaign in a transitional period (a recent site redesign, a new offer, a season change)? It's usually better to ride out sharp external changes on familiar manual bidding first, then move to automation once things settle.
Possible approaches to the switch itself (not a one-size-fits-all path)
- The standard path is a direct switch to tCPA, with the target set at (or just above) your actual CPA at the time of the switch, so you're not stacking two constraints at once (a new strategy plus an immediately tight target).
- Some agencies use an in-between step: switch to Maximize Conversions (no target) first, let the strategy learn how to allocate bids in principle, and only then add a target through tCPA — more cautious, but slower to get you the price control you're after.
- If you're unsure, you can test the switch through a Campaign Experiment (comparing manual bidding against tCPA on a slice of traffic) rather than moving the whole campaign at once — this lowers the risk if the timing turns out to be premature.