What is Portfolio bid strategy?

A portfolio bid strategy is a shared strategy object in Google Ads. Instead of each campaign carrying its own target, several campaigns — or ad groups, or keywords — are attached to one strategy with one target, and the bidder works to meet that target across the group.

The practical difference is what "meeting the target" means. Under per-campaign strategies, each campaign is held to the number on its own. Under a portfolio, the number applies to the aggregate, and the bidder is free to run one campaign above it and another below.

That freedom is the entire feature. Everything that follows is a consequence of it.

Trading efficiency between campaigns. Some campaigns are structurally cheaper than others — a branded campaign against a prospecting one, an established market against a new one. Held to the same target separately, the cheap one is constrained below what it could deliver and the expensive one throttles itself trying to reach a number it cannot. Pooled, the bidder buys all the cheap volume it can and uses the headroom to fund the dearer campaign up to the point where the blended target still holds.

Pooled learning. The model works across the whole portfolio, so campaigns that individually fall short of the volume threshold can together clear it. For an account with many small campaigns permanently unable to leave learning, this is frequently the strongest argument for the arrangement, and it is a structural fix rather than a tuning one.

One number to manage. A single target derived from the business, applied once, instead of a target per campaign that drifts apart as different people adjust different campaigns over a year.

The cost is interpretability, and it bites in a specific way.

A campaign inside a portfolio is not being held to the target individually. Its cost per result is therefore not a verdict on the campaign; it is one term in an average that the bidder is deliberately managing. A campaign running well above target may be doing exactly what the strategy intends.

This breaks the most common review habit in advertising, which is to sort campaigns by cost per result and act on the worst. Inside a portfolio that sort is meaningless, and acting on it — pausing the expensive campaign — removes the volume the cheap campaign's headroom was funding and leaves the portfolio smaller without being more efficient.

The reporting has to move up a level with the strategy. The portfolio is the unit of judgement; the campaign is a component.

The test is whether they share a ceiling.

A ceiling comes from contribution margin and the rate at which a result becomes a customer. Two campaigns selling the same thing to the same kind of customer share it. Two campaigns selling products with different margins, or reaching markets where a customer is worth different amounts, do not.

Averaging across campaigns with genuinely different ceilings produces a predictable error in both directions at once: the campaign whose true ceiling is higher is constrained below it, and the campaign whose true ceiling is lower is funded above it. The blended target is met and both halves are wrong.

The failure is quiet, because the portfolio's aggregate number looks fine. It shows up in the accounts rather than in the platform.

Two useful groupings and two poor ones.

Good: several prospecting campaigns across similar audiences for the same product line. Same ceiling, different structural costs, exactly the case the arrangement was built for.

Good: a set of small geographic campaigns, each too thin to learn on its own, all selling the same thing at the same margin.

Poor: brand defence pooled with prospecting. Branded clicks convert cheaply and reliably, so the portfolio will fund them enthusiastically and the target will be met largely by traffic that was going to arrive anyway. The prospecting half gets whatever is left.

Poor: high-margin and low-margin product lines together. The bidder cannot see margin; it sees conversions or revenue, and it will optimise a number that does not reflect what either line contributes.

Derive the target once, from the business, and apply it to a group that shares the economics behind it.

Review at the portfolio level. Set up the reporting that way before switching, because the habit of reading campaigns individually is strong and the arrangement makes it misleading.

Watch the composition. A portfolio whose volume drifts toward its cheapest campaign over months is meeting its target on an increasingly narrow base. That is worth noticing even though nothing in the aggregate figure complains about it.

Expect a learning period. Attaching campaigns to a shared strategy is a material change, and the model is rebuilding against a different problem.

Do not use it to rescue a campaign. A campaign that cannot meet its target alone will not be fixed by being averaged with better ones; it will be funded by them. Sometimes that is the intent, and it should be a decision somebody made rather than a side effect they discovered.

Shared bidding and shared budgets are separate mechanisms and they are easy to confuse, because both involve the platform allocating across several things.

A shared budget decides how much money each campaign gets. A portfolio strategy decides what each campaign is willing to pay per auction, and lets the target be met on average rather than individually.

They can be used together and they solve different problems. A portfolio without a shared budget means each campaign has its own money and a common view of what a result may cost. A shared budget without a portfolio means the money moves but each campaign is still held to its own target.

The combination is powerful and the least diagnosable arrangement in an account, because two allocators are now operating on the same campaigns and a change in either can explain almost any pattern. Accounts that use both should expect to review at the group level entirely and should resist reading anything into an individual campaign's figures.

Three things change in ways worth anticipating.

Individual campaign costs diverge. This is the strategy working and it is the part that alarms people.

Volume concentrates somewhat. Campaigns with cheaper results get more of the delivery, because they let the portfolio hold its target while buying more.

The aggregate steadies. The blended figure should become less volatile than the individual campaigns were, since the bidder is now managing the average directly rather than several numbers that happened to average.

If the aggregate does not steady after the learning period, the usual cause is that the campaigns in the portfolio do not belong together.