In short: a portfolio is a single target that governs several campaigns and makes them learn together: it is how you give Smart Bidding the volume no single campaign gathers on its own. The rule that avoids disaster is one portfolio per economic goal, with brand always kept apart. Bid limits are an emergency brake, not a management tool.
Smart Bidding needs volume, and many accounts have that volume spread across campaigns that, on their own, do not reach it. Google's solution is the portfolio bid strategy: a single strategy with one target governing several campaigns and learning from all of their data. Used well, it gives quality Smart Bidding to mid-sized accounts; used badly, it mixes things that should not be mixed.
What is a portfolio bid strategy?
A bid strategy (target CPA, target ROAS, maximise conversions/value, impression share) created in the shared library and applied to N campaigns. They all share the target and the model: one campaign's conversions teach the others. Each campaign keeps its own budget (unless they also share a budget — see below).
When should you use a portfolio?
| Situation | Portfolio? |
|---|---|
| Several campaigns with the same economics (the same acceptable CPA/ROAS) and few conversions each | Yes: the textbook case |
| Campaigns by area or by product with the same economics | Yes |
| One large campaign and several small ones of the same type | Yes: the small ones inherit the learning |
| Campaigns with different economics (brand vs generic, high-margin vs low-margin product) | No: the portfolio will put the same target on things that are worth different amounts |
| Campaigns on different networks (Search and Display) | No, unless the target is identical, and even then carefully |
| A single campaign with enough volume | Not needed (although it does no harm) |
Rule of thumb: one portfolio per set of economics. Brand on its own, generics together, ecommerce by margin band.
Should you set bid limits in a portfolio?
Portfolios let you (in the strategies with CPC) set a maximum and minimum bid per click. It is an emergency brake to avoid absurd CPCs in auctions where the algorithm "gets carried away", or to guarantee presence. Use them sparingly: a maximum limit that is too low stops the system bidding hard on the auctions that convert best, which are precisely the expensive ones. If you need tight limits, the problem is usually the target, not the bid.
When should you use a shared budget?
Separately from the portfolio, several campaigns can share one budget (also in the shared library). Google spreads the daily money between them according to where it performs best. The combinations:
- Portfolio + shared budget: maximum flexibility — one target and one pot for a group of homogeneous campaigns. Google decides which campaign spends each day.
- Portfolio without a shared budget: the same target, but each campaign has its own ceiling. Useful when you want to guarantee spend for each area or product.
- Shared budget without a portfolio: unusual; the campaigns compete for the pot on different criteria.
Which mistakes are made with portfolios?
- One portfolio for the whole account: it mixes brand, generic, Display and Shopping under the same target. Brand subsidises the rest and nobody knows what is performing.
- Changing the portfolio target with one campaign in mind: it affects them all. If one campaign needs a different target, take it out of the portfolio.
- Adding new campaigns with no data to a mature portfolio "so they learn": fine, but keep an eye on them not distorting the average CPA while they get going.
- Bid limits inherited from the manual era that strangle the strategy.
- Pooling campaigns with very different conversion cycles: the portfolio measures performance over a single window.
Example portfolio structure
| Portfolio | Campaigns | Strategy |
|---|---|---|
| Brand | Brand Search | Target impression share, 90% absolute top |
| Generic acquisition | Generic Search by service/area | Target CPA (one, set from the real figure) |
| High-margin ecommerce | Shopping/PMax for high-margin categories | High target ROAS |
| Low-margin ecommerce | Low-margin categories | An even higher target ROAS (or excluded) |
| Remarketing | Display/Demand Gen remarketing | Its own target CPA (usually lower) |
Five portfolios, five targets, each one making economic sense. As you grow they subdivide; as you shrink they merge.
💡 Ninja trick: the Suite's Smart Bidding (SBNS) works at campaign level and at portfolio level: it reads the account's portfolio strategies, moves their targets (tCPA/tROAS) and their campaigns' budgets in gradual steps towards the business target, and respects the fact that a campaign in a portfolio is not judged on its own. If your account has well-built portfolios — one per set of economics — the script has the structure it needs.
What you should remember
- Portfolio = one target for several campaigns that learn together: Smart Bidding for accounts with no volume per campaign.
- One portfolio per set of economics; brand always on its own.
- Bid limits: an emergency brake, not a management tool.
- A shared budget for homogeneous campaigns; separate budgets when you want to guarantee spend.
- The portfolio's target belongs to all of them: if one campaign needs a different one, out it goes.