Bidding Strategies

Target CPA vs Portfolio Bid Strategies: Target or Scope

In short: These sit on different axes. Target CPA is the cost-per-conversion field inside Maximize Conversions. A portfolio is an account-level container that applies one strategy, target included, across several campaigns whose conversion data then pools. A portfolio can hold a target CPA, so the real question is whether that number should be per campaign or blended across a group.

By the AdFlint research team · Fact-checked against current Google and Meta platform behavior · Last reviewed July 2026

Portfolio Bid Strategies

Google bid strategies stored at account level and shared across multiple campaigns, ad groups, or keywords working toward one common goal.

Rather than each campaign optimizing alone, a portfolio pools their conversion data and balances spend against the shared target, with optional bid limits and spend targets. It helps when individual campaigns are too small to learn on their own, or when you manage a group of campaigns against one blended CPA. The catch is losing per-campaign control, because the strategy can starve one campaign to fund another.

Full definition

Target CPA

Bids automatically toward a chosen average cost per conversion, letting individual conversions land above or below that number.

It now lives as a target field inside Maximize Conversions rather than a standalone strategy. Google bids up on auctions it expects to convert near your target and backs off elsewhere, so a target far below recent actual CPA simply throttles volume. Set it close to what the account already achieves, then move it in increments and let the learning period settle before judging the outcome.

Full definition

Side by side.

The differences that actually change what happens in your account.

 Portfolio Bid StrategiesTarget CPA
What it isA shared, account-level container for one bid strategy.A cost-per-conversion target field inside Maximize Conversions.
Where it is setShared library, then attached to campaigns.On the campaign, or on the portfolio if you use one.
How the target appliesAs a blended average across every attached campaign.To whatever it is set on, normally a single campaign.
Conversion dataPools across the group, which is the whole mechanical benefit.Only the campaign's own history unless pooled.
Extra controls exposedBid limits and a spend target across the group.The target amount itself.
Main riskOne campaign starves another to hold the blended number.Target set below achievable CPA, which throttles delivery.

What actually separates them.

01

A portfolio is not a different algorithm; attach a target CPA to it and every auction is priced by the same Smart Bidding model, just against a pooled target.

02

A campaign-level target CPA is honored per campaign, while a portfolio target is honored as a group average, so any single campaign inside it can run far above or below.

03

Pooling conversion data is the only real reason to reach for a portfolio, and it matters only when campaigns are individually too thin to learn on their own.

04

Portfolio strategies expose minimum and maximum bid limits that campaign-level Smart Bidding does not, and those limits apply to every attached campaign.

05

A portfolio shares bidding, not money; unless you also attach a shared budget, each campaign keeps its own daily amount while the strategy decides how hard to compete.

Which one should you use?

Use Portfolio Bid Strategies when

  • Several campaigns each produce too few conversions to train a model on alone.
  • You are judged on one blended cost per acquisition across a group.
  • Near-identical campaigns exist only because of geography, language, or account structure.
  • You need bid ceilings or floors that campaign-level strategies do not expose.

Use Target CPA when

  • The campaign already converts enough to learn from without borrowing data.
  • Its economics differ from neighboring campaigns, so one blended number would be wrong.
  • You need clean per-campaign diagnostics when cost per conversion moves.
  • You are testing a target change and want the effect isolated to one campaign.

Common questions.

Is a portfolio bid strategy better than a campaign-level Target CPA?

Neither is better; they answer different questions. The portfolio decides which campaigns share one target and one pool of conversion data. The target decides what that number is. If a campaign already converts enough to learn on its own and has its own economics, pooling it mostly imports someone else's target.

Can campaigns inside a portfolio have different target CPAs?

No. A portfolio holds one target and every attached campaign optimizes toward it as a blended average. If a campaign needs its own number, give it a campaign-level strategy or create a second portfolio. Splitting one portfolio into two grouped by economics is usually cleaner than a target that fits neither half.

Does moving campaigns into a portfolio restart learning?

Treat it as a bid strategy change, which means a fresh learning period on the affected campaigns. Expect delivery and cost per conversion to move while the pooled model settles, and do not judge the change until the learning period has passed plus a full conversion window on top of it.

Or stop choosing between them.

AdFlint picks the setting, writes the ads, and keeps optimizing inside the Google and Meta accounts you already own.

Try AdFlint freeSee how it works

Related comparisons

All comparisons