Smart Bidding Exploration Is Not More Reach. It Is a Target ROAS Cut You Choose, Now in PMax and Shopping
The short version
Smart Bidding Exploration, a Search feature, now has Help documentation for Performance Max and Shopping, and Search Engine Land reported the extension on October 23. Despite the name, it does not add reach. Google's Help page says it lowers your Target ROAS by a tolerance you choose, 5% to 30%, so the bidder can pursue queries your target was screening out. It needs Target ROAS, it needs a campaign that is not limited by budget, and moving a Shopping or PMax campaign into a portfolio strategy switches it off. Google says results should land within about 15% of the lowered target, which is not a promise of more conversions.
What Google announced
Google's Help page, "How Smart Bidding Exploration for Search, Performance Max, and Shopping campaigns works," now covers all three campaign types, and Search Engine Land reported the expansion on October 23. The expansion itself is not new news: Google flagged it as a beta ahead of Google Marketing Live on May 7, and the journey-aware bidding brief noted it in passing. What is new is the documentation, and the documentation is specific.
On Search, it works as before. On Performance Max, it explores additional searches among the Search-channel queries the campaign is already eligible for. On Shopping, and on PMax with a Merchant Center feed, a single toggle covers both Search and Shopping traffic. PMax without a feed explores Search only. Google says it is fully compatible with Search and Shopping campaigns running AI Max, and Search Engine Land adds that Search Ads 360 does not support it for Shopping or retail PMax.
What it actually means
The name suggests growth. The mechanism is a giveback. Google's own example: a 200% Target ROAS with a 10% tolerance produces an effective target of 180%, and you should expect performance within approximately 15% of that 180% over one to two conversion cycles. You are pre-authorizing the bidder to accept worse-than-target returns on exploratory auctions, in exchange for queries it would otherwise skip. Google also says it does not expand broad match keywords or your outer targeting limits. It finds long-tail queries you already qualify for and were missing because the target was strict.
That makes it a fair trade for one kind of account and a bad one for another. If a Target ROAS campaign is leaving volume unspent because the target is tight and the budget is not the constraint, a 5% to 10% tolerance is a controlled way to buy that volume. If the campaign is spending its budget already, it does nothing: Google says exploration may be paused when the budget limit is hit. That is the same logic that makes the bid target adjustment tool matter mostly for budget-limited campaigns, and the same budget question behind channel prioritization.
Who loses control, and of what: you keep the dial, because you choose the tolerance, but the bidder decides where the extra spend goes inside it. The default and the exact location of the setting are not stated on the Help page, so treat it as something to look for in your account rather than assume. The portfolio trap is the one most likely to catch someone: Shopping and PMax campaigns cannot use portfolio strategies, and Google says moving one into a portfolio strategy automatically deactivates Exploration, with no reminder.
Two limits on what you can say to a client. Google's page gives an expectation about performance relative to the effective target, but it makes no promise of more conversions or better ROAS, and Search Engine Land says the same. And the conversion data under the target matters more than usual, because the bidder is chasing a ROAS number built from the value you send. If values are late or wrong, the extra exploration is aimed at the wrong thing; the offline conversion upload gate is the reporting caveat to know first. For the leadership argument about what a tolerance costs and who should approve it, that lives at Uncommon Move in The Tolerance Band. This brief covers the mechanics.
Exploration is not a new audience. It is a smaller promise about ROAS, delivered in exchange for volume.
Signal or noise: high for a narrow group
High if you run Target ROAS Performance Max or Shopping campaigns that are not limited by budget and are not in a portfolio strategy. Low if your campaigns are budget-limited, run on Target CPA, or rely on portfolio strategies. Google's incentive is visible: a lower effective target means more auctions entered and more spend, and the tolerance is yours to set but the appetite for it is Google's.
The Monday-morning playbook
- Build the candidate list. Filter to Target ROAS Performance Max and Shopping campaigns. Mark any that hit their daily budget in the last 30 days; exploration may pause there, so they get nothing.
- Remove the ineligible. Check bid strategy type on every candidate. Portfolio strategies are not supported for Shopping or PMax, and a move into one deactivates Exploration.
- Choose one non-peak test campaign. Write down its current Target ROAS, the tolerance you will try (start near the 5% end), and the resulting effective target. For a 300% target and a 10% tolerance, that is 270%.
- Find the setting. Check the campaign's bidding settings for the toggle. For PMax with a Merchant Center feed, one toggle covers Search and Shopping; without a feed it covers Search only. If you cannot see it, the rollout has not reached the account yet.
- Run it one to two conversion cycles. Compare conversion volume, ROAS against the effective target (not the old one), and spend. Do not stack other target changes in the same window.
- Calendar the Black Friday decision. Decide by November 20 whether any peak-season campaign gets it before Black Friday and Cyber Monday on November 27. My view: a campaign you have not tested should not go in during the final week.
Related: Customer lifecycle goals now include retention modes and percentage multipliers; avoid testing both on the same campaign.
Related: Smart Bidding Learning status now caps at 7 days, and budget changes no longer trigger it.
This rollout is logged on the Google Ads changes and deadline calendar and rated on the signal vs. noise ledger, both reviewed weekly.
The bottom line
Treat Exploration as a target cut with a ceiling, not a growth feature. Test it on one eligible campaign now, read results against the effective target, and keep untested peak-season campaigns out of it until after November 27.
Frequently asked
What tolerance can I set for Smart Bidding Exploration?
Google's Help page says the Target ROAS tolerance can be set between 5% and 30%. The campaign's effective target is your target reduced by the tolerance: a 200% target with a 10% tolerance gives an effective target of 180%, and Google says to expect performance within approximately 15% of that effective target over one to two conversion cycles.
Does Smart Bidding Exploration work on budget-limited campaigns?
Not reliably. Google says campaigns should not be limited by budget, and that if a budget limit is hit, exploration may be temporarily paused. It also requires Target ROAS bidding, so Target CPA campaigns are not eligible.
Does Smart Bidding Exploration work with portfolio bid strategies?
Only on Search campaigns. Google says portfolio strategies are supported for Search but not for Shopping or Performance Max, and that moving a Shopping or Performance Max campaign into a portfolio strategy automatically deactivates Smart Bidding Exploration.
Get the next brief in your inbox
When Google ships something that matters, you'll get the decode: what changed, whether it's signal or noise, and the move worth making this week. Free, no firehose, unsubscribe any time.