Google Ads Bidding Strategies: From Manual to Automated

Bidding strategy is the most debated and most misunderstood setting in Google Ads. The debate is usually framed as manual versus automated; the real question is whether the account holds enough data for automation to learn from.

This guide orders the strategies by data maturity, explains when to move, and how to manage the learning period.

What a bidding strategy decides

Every search is an auction. The bidding strategy is the rule that decides how much you put into it. On manual you make that decision; on automated the system makes it per search, based on past conversion data.

The advantage of automation is scale. No human can calculate a separate bid for thousands of searches a day; the system can. The disadvantage is dependency: it optimises only towards the goal it was taught.

So the bidding decision is really a measurement decision. If your conversion definition is wrong, automated bidding runs more efficiently towards the wrong goal. Bad measurement becomes more expensive with automation, not less.

Manual CPC: control and the learning phase

On manual cost per click you set the maximum bid for each keyword. In newly built accounts this is usually the starting point.

The reason is not control but the absence of data. Automated strategies have no conversion history to learn from yet; the system works on assumption and you pay for that assumption. On manual you know what you will spend.

The purpose of the manual phase is collecting data as much as making sales. You learn which keyword produces what, at which hour and on which device conversions happen. That knowledge underpins every later decision.

A common mistake in manual bidding is setting bids too low. Very low bids get no impressions; a keyword with no impressions produces no data and the account stalls before it can learn. Treat the first month's bids as the price of learning.

Maximise clicks: the transition step

Maximise clicks aims for as many clicks as your budget allows. Because it needs no conversion data, it is the natural step after manual.

It is useful when you want volume and conversion counts are still low. The risk is in the name: it targets clicks, not customers. Run without a maximum bid limit, it can collect expensive and irrelevant clicks.

For that reason always pair this strategy with a maximum bid cap. Uncapped, it spends the budget quickly but inefficiently.

Maximise conversions and target CPA

Maximise conversions aims for as many conversions as the budget allows. Target CPA adds a cost anchor: it tries to hold the cost per conversion you set.

These strategies need conversion data. For the system to find a real pattern the campaign has to produce conversions regularly. In a campaign with a handful of conversions a month, automation mistakes noise for pattern.

Be realistic when setting a target CPA. Set it far below your current average and the system stops seeking impressions; the campaign quietly stalls and you read it as a performance problem.

The practical approach: set the target near your current average, then lower it gradually once the campaign is stable. Each change requires the system to learn again.

Target ROAS: accounts with revenue data

Target ROAS aims for a set revenue per unit of spend. It only makes sense in accounts that can send a monetary value with the conversion: natural in e-commerce, possible in services only if lead valuation is in place.

On the service side that means assigning different values to different conversion types: a form submission carries one value, a phone call another. Running target ROAS without that valuation asks the system to decide on incomplete information.

Three conditions for the transition

The first condition is verified measurement. Conversion actions must be defined, confirmed to fire, and protected against double counting. Skip this and everything else is built on a wrong foundation.

The second is sufficient conversion volume. There is no exact threshold, but the campaign should produce conversions consistently week to week. Automation is not reliable in a campaign with single digit monthly conversions.

The third is a ready structure. Campaigns on separate budgets, single themed ad groups, negative layers in place. Automation does not fix a bad structure; it runs a bad structure faster.

The learning period: do not touch, wait

When you switch to an automated strategy the system enters a learning period. Performance fluctuates; costs can rise and conversions can fall. That is not a fault, it is part of the process.

The most damaging thing you can do here is panic and change settings. Every change restarts the learning and the account ends up permanently learning and never learned.

The rule is simple: do not touch anything for at least two weeks after the switch. Keep budget, target and structure fixed. Evaluate after that period.

For the same reason do not make several changes at once. If you raise the budget in the same week you change the bidding strategy, you cannot tell which one moved the result.

Three common mistakes

First, switching to automated bidding before conversion measurement exists. The system does not know what to optimise for; it collects clicks, not customers. This is the most expensive misuse of automation.

Second, setting the target too aggressively. An unreachable target CPA leaves the campaign without impressions. It looks stalled and the reason usually goes unnoticed.

Third, not waiting out the learning period. Deciding on three days of data and reverting wastes the cost of those two weeks and guarantees you never see the real outcome.

Conclusion

Choosing a bidding strategy is not a preference, it is a consequence of account maturity. Start manual and gather data, verify measurement, build the structure, then move to automation and leave it alone for two weeks.

In a well built account automation is a strong multiplier; in a badly built one it only repeats the mistakes faster.

If you would like us to review your bidding setup, you can reach us from our Google Ads management page. The full picture is in our Google Ads guide.