Portfolio bid strategy in Google Ads: when pooling campaigns pays
What a portfolio bid strategy is, how it differs from standard campaign strategies for Shopping, how shared budgets interact, and when a portfolio tROAS rescues small campaigns.
A portfolio bid strategy is one Smart Bidding strategy shared across multiple campaigns, so they pool their conversion data and optimise to one target together. For Shopping accounts split into many small campaigns, portfolios can fix the thin-data problem that makes per-campaign tROAS erratic, at the price of losing per-campaign targets. This reference covers portfolio vs standard, the shared-budget interplay, and when pooling actually helps.
What a portfolio bid strategy is
A standard bid strategy lives inside one campaign: that campaign's tROAS target, learned from that campaign's conversion data alone. A portfolio bid strategy is created at account level (Tools → Budgets and bid strategies) and attached to several campaigns at once. Every attached campaign optimises to the same target, and (this is the part that matters) Smart Bidding learns from their combined conversion history.
For Shopping, portfolios are available for Target ROAS, Maximise Conversion Value, Maximise Clicks and Target Impression Share. In practice, portfolio tROAS is the one merchants reach for.
Portfolio vs standard: the actual trade
| Standard (per-campaign) | Portfolio (shared) | |
|---|---|---|
| Target | One per campaign | One for the whole portfolio |
| Learning data | That campaign's conversions only | Pooled across all attached campaigns |
| Best when | Campaigns have distinct economics and enough data each | Campaigns are small but economically similar |
| Risk | Thin data → erratic bidding per campaign | One target flattens genuinely different campaigns |
| Learning resets | Per campaign | Portfolio-wide, a change affects all attached campaigns |
The trade is precision for stability. A standard strategy respects that your clearance campaign and your hero-product campaign deserve different targets. A portfolio ignores that distinction, in exchange for giving Smart Bidding enough conversions to actually predict with.
The averaging caution from campaign-level tROAS applies double here: a portfolio target is an average across campaigns, not just products. Pool a 60%-margin accessories campaign with a 15%-margin electronics campaign under one 400% target and Google will quietly starve one and overfeed the other. Only pool campaigns whose economics genuinely match, which is exactly what margin-band segmentation via custom labels is for.
When a portfolio tROAS helps small campaigns
Per-campaign tROAS behaves erratically below roughly 30–50 conversions a month. Many Shopping accounts split their catalogue into six or eight campaigns for structural reasons (brand splits, priority tiers, category segmentation) and end up with each campaign individually starved of data even though the account converts fine in aggregate.
That's the portfolio's home ground. Worked example:
- Four category campaigns, each converting 12–15 times a month, individually well below the stable-tROAS floor, each one wobbling.
- Combined: ~55 conversions a month, enough for tROAS to hold a target.
- If all four run at similar margins, attach them to one portfolio tROAS at the blended trailing ROAS, and the bidding stabilises without restructuring anything.
The test before pooling is always the same two questions: is each campaign's data thin (if they all clear 40–50 conversions/month, standard strategies keep more control), and do their economics match (if break-even ROAS differs materially between them, one target will misprice at least one of them).
Shared budgets: related, separate, often confused
Portfolio strategies and shared budgets live on the same settings page and get conflated constantly. They're independent:
- A portfolio strategy pools bidding data and the target.
- A shared budget pools the spend allowance: one daily pot that Google allocates across the attached campaigns wherever it expects the best results, instead of fixed per-campaign caps.
You can use either alone or both together. Together, they hand Google maximum flexibility: one target, one pot, money flowing daily to whichever campaign converts best. That's efficient, and it means you surrender per-campaign spend control. If you need campaign A capped at £30/day for cash-flow or strategic reasons, don't put its budget in a shared pool. The daily-budget mechanics (monthly pacing, 2× overspend days) work the same as individual budgets, just at pool level.
One genuine caution: a shared budget plus a tight portfolio target reproduces the classic conflict, "spend less" and "only take premium auctions" at the same time, across your whole campaign group at once. Loosen one before tightening the other.
Deep dive Portfolio learning, resets, and the blast radius of changes
A portfolio strategy learns as one unit, and that cuts both ways.
The upside is pooled signal. Smart Bidding's conversion-value predictions get sharper with volume. Four thin campaigns on standard tROAS each run their own under-fed model; the same four in a portfolio feed one model that sees every conversion. For accounts whose structure fragments data, this is the cheapest stability upgrade available: no restructuring, no feed changes, one settings move.
The downside is shared blast radius. Change the portfolio target and every attached campaign enters the adjustment period together; there's no such thing as a contained experiment inside a portfolio. Similarly, attaching or detaching a campaign changes the pooled data the model calibrates against, which can wobble the campaigns that stayed put. The learning period rules apply portfolio-wide: small relative target steps (10–20%), one change at a time, two weeks before judging.
Graduation is the pattern to plan for. Portfolios suit a lifecycle stage, not necessarily forever. A campaign that grows into 50+ conversions a month of its own, with margins that no longer match its portfolio siblings, has earned an individual target. Detach it deliberately: move it to a standard tROAS set at its own trailing ROAS (not the portfolio target), expect a learning period, and leave the rest of the portfolio alone while it settles. This mirrors the incubation pattern at product level (pool while thin, graduate when proven), the same logic BidSmart's incubator applies to individual products, with each graduation staged for your approval rather than executed automatically.
What a portfolio can't fix: product-level divergence inside the pooled campaigns. The portfolio sets one average across everything attached; which individual products could sustain a higher or lower target is invisible at that altitude. That's a product-level analysis question, see tROAS by product, and the answer often reshapes which campaigns belong in which portfolio in the first place.
Setting one up: the short version
- Group candidate campaigns by economics, not convenience: similar margins, similar break-even ROAS, similar goals.
- Create the portfolio under Tools → Budgets and bid strategies → Bid strategies, choose Target ROAS, and attach the campaigns.
- Set the target from the group's blended trailing 30-day ROAS, not an aspiration, the same starting-target rule as any tROAS.
- Expect a learning period across all attached campaigns, and hold changes until it settles.
- Review monthly for outgrowers: campaigns with enough data and diverging economics graduate to their own strategy.
Quick reference
| Situation | Recommendation |
|---|---|
| Several campaigns, each under ~30 conversions/month, similar margins | Portfolio tROAS |
| Campaigns with clearly different margins or goals | Separate strategies (standard, or separate portfolios per margin band) |
| One big campaign, plenty of data | Standard tROAS, a portfolio adds nothing |
| Need hard per-campaign spend caps | Portfolio strategy is fine; avoid the shared budget |
| Campaign inside a portfolio now converting 50+/month with distinct economics | Graduate it to its own target |
Frequently asked questions
What is a portfolio bid strategy in Google Ads?
A single automated bid strategy (such as Target ROAS) applied to multiple campaigns at once. The campaigns share one target and, crucially, pool their conversion data, so Smart Bidding learns from the combined history instead of each campaign's slice alone.
Portfolio vs standard bid strategy, which is better for Shopping?
Neither universally. Standard (per-campaign) strategies suit campaigns with enough of their own conversion data (roughly 30–50+ conversions/month) and genuinely different economics. Portfolios suit groups of small campaigns with similar economics whose individual data is too thin for stable tROAS.
Do campaigns in a portfolio have to share a budget?
No. Portfolio strategies and shared budgets are separate features you can combine or use independently. A portfolio pools bidding data; a shared budget pools spend allowance. Using both hands Google maximum freedom to move money to whichever campaign converts best.
Can I set different targets for campaigns inside one portfolio?
No, one portfolio, one target. If two campaigns need different targets because their margins or goals differ, they belong in different portfolios (or on standard strategies). That constraint is the main reason not to pool everything.