Google AdsStep 4 of 5 on the chain

Maximize conversion value

The bid strategy that spends the budget to bring back the highest total conversion value rather than the highest count.

Why it matters

It is the strategy value-based bidding exists to feed, and the one that does nothing at all if every conversion carries the same value.

What goes wrong

Switching to it is the moment most advertisers expect the improvement, and the moment nothing happens if the feed behind it is flat. The strategy is not the intervention; the spread in the values is. It is also where every switch should start: with no target attached, until Google is out of learning and the values have been checked against closed deals.

With no target it spends the whole daily budget, every day, toward the highest total value it can find. That is the right behaviour during learning and the reason the budget has to be a real limit rather than a ceiling you assume it will not reach. Since June 2026 Google labels the same strategy with a target as Target ROAS again, where for several years it read as Maximize conversion value with a target ROAS. Older screenshots show the long name. The bidding is identical.

How to check it on your own account

Campaign, Settings, Bidding. Anything that reads Maximize conversions or Target CPA is bidding on the count of leads and ignoring every value you send. Then check the budget is what you actually intend to spend, because this strategy will spend it.

Next to this

Where this comes up

In your trade

The industry pages that work this through for a kind of business.

The rest of this step

Measure this on your own data

The diagnostic reads a CRM export in your browser and reports your volume, your match rate and the spread between your leads against the thresholds in this glossary. Nothing is uploaded and no account is needed.

See what your own leads are worth

Read your closed deals and find out whether your lead values actually vary, and by how much. Nothing is stored, and your file is read in your browser.

Try it on a sample dataset