Target CPA or Target ROAS: Which Google Ads Bidding Strategy to Choose

Target CPA and target ROAS are two goals that tell Smart Bidding in Google Ads how to spend your budget: the first controls the average cost of one conversion, the second the sales value from every unit of spend. Below we explain how they differ, when to choose which, how much data you need and how to set the number so the campaign does not stall.

How target CPA differs from target ROAS

Target CPA (cost per acquisition) sets the average conversion cost the algorithm should hold, and target ROAS (return on ad spend) sets the ratio of conversion value to cost. The first knows only the number of events, the second also takes their amounts into account. That is the only real difference, but it has big consequences.

In the newer Google Ads interface both strategies appear as optional goals inside the maximise strategies: target CPA is maximise conversions with a cost target, and target ROAS is maximise conversion value with a return target. Older guides use the former names, but the mechanism is the same.

When to choose target CPA

Choose target CPA when conversions have similar value or the value cannot be assigned reliably, and holding a steady acquisition cost is the priority. This covers most service businesses: a quote request for a renovation firm or an enquiry about a training course costs about the same, no matter how much the customer eventually pays.

• Lead generation: contacts are not priced at the moment of enquiry, so amounts in Google Ads would be guesses anyway.
• A shop with a narrow price range: if every product costs 90-110, basket value changes little.
• A new account: at the start it is easier to set a maximum cost per enquiry than a return on a value the account has not yet seen.

One condition: an enquiry has to be a good measure of success. If part of your leads are spam and wrong-number calls, the algorithm will multiply them. In that case combine target CPA with offline conversion import, so the campaign learns from qualified leads, not form fills.

When to choose target ROAS

Choose target ROAS when conversion values genuinely differ and you pass them to Google Ads, because then the algorithm can pay more for clicks likely to bring a large basket and less for those with a small one. This is typical in e-commerce with a wide range, where an order worth 50 and one worth 800 should not cost the same.

The ROAS goal is calculated from margin. If margin after cost of goods is 30%, break-even is 1 / 0.30, about 333%: every unit of ad spend must bring 3.33 in revenue to break even. A goal above that threshold is where profit begins. We lay out the full calculation, together with maximum CPC, in the text on break-even ROAS.

• Basket value: the tag has to pass the transaction amount, ideally without shipping costs if you calculate from product margin.
• Returns and cancellations: revenue from orders that come back inflates ROAS, so correct the value or set the goal with headroom.
• Varied margins: when categories have very different margins, one goal for the whole account hides that part of the range works at a loss.

We write about how ROAS can look good with weak results in the article on inflated ROAS in Performance Max. The same rule applies to every campaign: a metric from the panel does not replace a profit calculation.

How many conversions you need to run either strategy

According to Google's help, a strategy with a goal needs at least 15 conversions in 30 days, and for target ROAS it recommends starting with maximise conversions or target CPA and switching once there is enough data. That is a technical minimum, not the level at which the strategy works steadily: the more conversions, the smaller the swings.

The first learning phase after a strategy change usually lasts one to two weeks, and during that time you should not change the goal or budget. Tweaking too often resets learning and turns the campaign into a series of unfinished experiments. If you run an account with low volume, merge similar campaigns instead of splitting a dozen conversions across five separate strategies.

How to set the goal value without choking the campaign

Set the first goal close to what the campaign already achieves, then change it gradually, because an overly ambitious value limits reach and the campaign stops spending its budget. An algorithm that cannot pay for a click what the auction requires simply does not take part in it.

• Start from the 30-day average: if the campaign has a CPA of 120, set 120, not 80.
• Change by 10-15% every two weeks: bigger jumps wipe out what was learned.
• Watch spend: a drop in budget delivery after tightening the goal means the goal is too hard.
• Judge over a 30-day window: daily swings are normal with Smart Bidding.
• Mind conversion delay: with a purchase cycle of two weeks, the latest days look worse than they will once sales are added.

When you choose between these strategies for Search and Performance Max campaigns, also read how Performance Max differs from Search, because the campaign type decides how much control you have over where the budget goes.

The most common mistakes when choosing a bidding strategy

• A goal from thin air: a CPA or ROAS figure copied from an industry article instead of your own margin and close rate.
• ROAS without values: all conversions carry the same default value, so target ROAS works like a more expensive CPA.
• Optimising for the wrong signal: counting a click on a phone number or a thank-you page view without a submitted form as a conversion.
• Frequent changes: goal and budget adjusted every few days, before the learning phase ends.
• Too many campaigns: conversions spread across a dozen small campaigns leave none of them enough data to learn.

Strategy selection and goals are part of the settings we deploy when running Google Ads campaigns. As a digital marketing agency we start from the client's profitability calculation, and only then type the number into the panel.