Budgets & Bidding Strategies

Manual CPC vs Smart Bidding: the fundamental choice

Manual CPC means you set the maximum you're willing to pay per click yourself, and adjust it based on judgement and reporting. Smart Bidding means you hand that decision to Google's machine learning, which sets bids automatically in each individual auction based on the likelihood of that specific search leading to a conversion, using far more real-time signals — device, location, time, browsing history — than a human could realistically factor in manually.

In most accounts today, Smart Bidding outperforms manual bidding once it has enough conversion data to learn from — but that "once it has enough data" condition is the part beginners skip, and it's the difference between Smart Bidding working well and Smart Bidding quietly wasting budget. A genuinely new account, with no conversion history at all, is often better served starting on a conservative manual bid or Maximize Conversions with a capped budget, purely to build up the history that every other strategy depends on.

The Smart Bidding menu: Maximize Conversions, Target CPA, Target ROAS

• Maximize Conversions — spends your full daily budget trying to get as many conversions as possible, with no target cost per conversion. A sensible starting point for a new campaign with limited data, since it's the least constrained strategy.
• Target CPA (cost per acquisition) — aims to hit an average cost per conversion you specify, spending more or less per click depending on how likely each auction is to convert. Needs a realistic target based on your actual historical data, not a wish.
• Target ROAS (return on ad spend) — aims for a target revenue-to-spend ratio, and requires conversion value tracking, not just a conversion count, since it optimises toward revenue rather than volume. Most relevant for ecommerce or accounts where deal size varies meaningfully.

A common, sensible progression is to launch on Maximize Conversions, watch the average cost per conversion over several weeks once volume is stable, and only then introduce a Target CPA set close to that observed average — rather than guessing a target on day one with no data to base it on. Moving to Target ROAS follows the same logic, but only once conversion value is being tracked accurately enough to trust.

The learning phase: how much data Smart Bidding needs

Whenever you turn on a Smart Bidding strategy, or make a significant change to one, the campaign enters a "Learning" status while the algorithm recalibrates. Performance during this window is often less stable, and sometimes worse than before the change — this is expected, not a sign of failure. Smart Bidding needs a genuine base of historical conversions to learn from; a campaign with only a handful of conversions a month gives the algorithm very little to work with, and results will stay noisy until that history builds up.

The practical implication: give a new bidding strategy real time, typically several weeks, and a real accumulation of conversions before judging it or switching again. Constantly toggling strategies resets the learning process and is one of the most common self-inflicted problems in small accounts — an advertiser who panics after three quiet days and switches strategy again is often just restarting the same learning window from scratch, over and over, without ever letting it complete once.

Daily budget math: translating budget into real volume

A useful gut-check before launching: divide your realistic monthly budget by an estimate of your average CPC for your keywords, visible in Keyword Planner, to get a rough sense of monthly click volume, then apply a realistic estimated conversion rate to see how many leads that might produce. If the resulting click volume is very low — a handful of clicks a day — expect the account to take longer to build the conversion history Smart Bidding needs, and consider narrowing keyword focus rather than spreading a small budget across too many keywords or campaigns at once.

It's also worth sanity-checking a daily budget against your average CPC directly: a daily budget that can only cover two or three clicks at your keywords' typical cost per click will struggle to gather any meaningful signal, no matter which bidding strategy is selected. Narrowing to your highest-intent keywords first, rather than spreading thin across a large list, usually produces a more useful early data set.

When to change strategy (and when to leave it alone)

Change bidding strategy when you have a clear, data-backed reason: a Target CPA that is mathematically unreachable given your actual conversion rate and average order value, or a campaign that has matured well past the learning phase and consistently under- or over-delivers against a realistic target. Do not change strategy based on a single bad day, a short dip, or impatience in the first week or two after launch — that's usually just the learning phase working as intended.

• Has the campaign had enough time — generally several weeks — since the last major change?
• Has it accumulated a meaningful, steady flow of conversions, not just a handful?
• Is conversion tracking (Chapter 7) accurate enough that you'd trust it to let an algorithm act on it?
• Is the target you're setting realistic given your actual historical CPA or ROAS, not just a hopeful number?