Google Ads Promotion Mode is the most useful feature in Google’s June 2026 bidding overhaul, and it is the one most coverage rushed past on the way to the August 17 headline. Announced on June 15, 2026, Promotion Mode lets you schedule a temporary, self-closing change to your ROAS tolerance and inject extra daily budget across a defined peak window — a flash sale, a holiday ramp, a product launch — then reverts on its own when the window closes.
That auto-reverting design is the whole point. Most peak-period spending is still managed by hand: someone bumps a budget before Black Friday and someone else is supposed to remember to pull it back on the following Monday. Promotion Mode builds the end date into the setup, so the boost expires whether or not anyone remembers it. For agencies running peak calendars across a client roster, that single property changes how you plan a sale.
This guide is the operator’s manual for the two peak-and-reach features in the overhaul: Promotion Mode and the expansion of Smart Bidding Exploration to Performance Max without product feeds. We cover the mechanics, a tool-routing decision matrix you will not find as a standalone asset anywhere else, the 5–30% ROAS tolerance band and how to compute its effective floor, and the setup sequence. We only touch the August 17 target change in passing — we already published a full June 2026 bidding and budgeting overhaul playbook on that.
- 01Promotion Mode schedules a self-closing peak boost.It adjusts ROAS tolerance and adds extra daily budget across a defined 3-to-14-day window for Search and Performance Max campaigns, then reverts automatically when the window closes — no manual rollback to remember.
- 02It is not a seasonality adjustment.Seasonality adjustments signal an expected conversion-rate change so Smart Bidding can pre-adjust its bids. Promotion Mode acts directly on the ROAS-tolerance and budget levers. They answer different questions and can run in parallel for the same event.
- 03Smart Bidding Exploration now reaches PMax without feeds.As of June 15, 2026 it is globally available for Performance Max campaigns without product feeds, across all languages. Shopping and PMax-with-feeds are a separate beta — do not assume Shopping has it until you see it live.
- 04Exploration runs inside a 5–30% ROAS tolerance band.You set how far below your Target ROAS the algorithm may bid on unproven queries. A 200% target at 10% tolerance gives a 180% effective floor. It needs unconstrained budgets and removed CPC bid caps to work as intended.
- 05The August 17 target change is a separate decision.Promotion Mode and Smart Bidding Exploration are peak-and-reach levers. The August 17 Bidding Target Optimization change is a backend stabilisation for budget-limited campaigns — covered in depth in our separate overhaul playbook.
01 — What This Post CoversThree changes shipped together — two of them live here.
Google bundled three distinct moves into its June 15, 2026 announcement, and they pull in different directions. The August 17 Bidding Target Optimization change is a stabilisation lever — it pulls budget-limited campaigns that have been over-delivering their stated targets back toward those targets. The other two are growth levers: Promotion Mode pushes harder through a peak, and Smart Bidding Exploration reaches further into unproven queries. This post is about the growth levers.
The split matters because the two camps demand opposite reflexes. The August 17 change asks you to tighten and confirm your targets; Promotion Mode and Smart Bidding Exploration ask you to deliberately loosen them for a window or a query frontier. Reading all three as “Google is automating bidding” and acting on none of them is the failure mode — each needs a separate decision.
Promotion Mode
Schedule a temporary ROAS-tolerance change plus extra daily budget across a 3-to-14-day peak window. The end date is built into setup, so the boost self-closes. Works with daily and campaign total budgets.
Smart Bidding Exploration
Now globally available for Performance Max without product feeds across all languages. Lets the algorithm bid on queries with unproven conversion histories within a ROAS tolerance band you set.
Bidding Target Optimization
Steers budget-limited campaigns that over-deliver their stated tCPA or tROAS back toward the set target. We cover the audit, the risk matrix, and the action window in our dedicated overhaul playbook.
02 — How Promotion Mode WorksA boost with a built-in off-switch.
Promotion Mode is a new beta for Search and Performance Max campaigns that does two things at once across a date range you define: it temporarily changes your ROAS tolerance, and it injects extra daily budget. The defined window runs from a minimum of 3 days to a maximum of 14 days, and the end date is part of the setup flow rather than an afterthought. When the window closes, the campaign reverts to its standing tolerance and budget on its own.
That self-closing behaviour is the design intent. Google’s Ads Product Liaison, Ginny Marvin, framed it directly when the beta surfaced: the very first step is setting the start and end dates, so there is nothing to remember to switch off afterwards.
"With Promotion mode, there's no need to remember to turn it back off. The very first prompt is to set the start and end dates."— Ginny Marvin, Google Ads Product Liaison, via LinkedIn (cited by PPC Land), June 2026
Promotion Mode is compatible with both daily budgets and campaign total budgets — the latter, launched in 2025, lets you set a fixed budget across a span of days rather than a daily rolling cap, and Google reports a vendor-stated average 66% reduction in manual budget adjustments where advertisers adopt it. Pairing a campaign total budget with a Promotion Mode window is a clean way to express “spend this much, more aggressively, between these dates, then stop” without babysitting the account through the peak.
At beta launch, Promotion Mode is confirmed for Search and Performance Max only — not Shopping, not Display. Because it is a beta and the in-product steps are not fully public, treat the mechanics above as the documented behaviour and validate the exact flow in an account before standardising it across clients.
03 — Decision MatrixOne table that routes every peak event.
No published source puts Promotion Mode, seasonality adjustments, and a permanent Target ROAS change side by side with use-case routing, so we built the table below. The tool windows come from Google’s documentation — Promotion Mode runs 3 to 14 days; seasonality adjustments are recommended for short windows, typically no more than about 7 days. The scenario-to-tool routing is ours, derived from those documented windows and from what each lever actually adjusts.
| Demand scenario | Recommended tool | What it adjusts | Window | Self-closing? | Best paired with |
|---|---|---|---|---|---|
| Black Friday / Cyber Monday core (2–3 days) | Promotion Mode | ROAS tolerance + extra daily budget | 3–14 days | Yes — built-in end date | Daily or campaign total budgets |
| Two-week holiday ramp | Promotion Mode | ROAS tolerance + extra daily budget | 3–14 days | Yes — built-in end date | Campaign total budget for the window |
| Product launch week | Promotion Mode | ROAS tolerance + extra daily budget | 3–14 days | Yes — built-in end date | Smart Bidding Exploration for new queries |
| Flash sale under 24 hours | Seasonality adjustment | Expected conversion-rate signal | Short windows (Google rec. ≤ 7 days) | Ends on its set date | Manual budget headroom |
| Anticipated CVR spike, budget unchanged | Seasonality adjustment | Conversion-rate forecast only | Short windows (≤ 7 days) | Ends on its set date | Existing daily budget |
| Sustained, non-temporary demand growth | Lower the Target ROAS (permanent) | The standing ROAS target | Permanent | No — it is a permanent change | Smart Bidding Exploration |
The one row people get wrong is the sub-24-hour flash sale. Promotion Mode’s minimum window is 3 days, so it cannot cleanly express a sale that lasts an afternoon — a seasonality adjustment that signals the conversion-rate spike is the better fit, paired with manual budget headroom. Anything that fits inside 3 to 14 days and warrants both looser efficiency and more budget is squarely Promotion Mode territory.
04 — Which Tool, Which EventReach for Promotion Mode when the event has a shape.
The matrix is the reference; this is the judgement. Promotion Mode earns its place when an event has a defined start and end and you genuinely want to spend more while accepting a looser efficiency bar for that span. When the event is sub-day, or when you only expect conversion rates to move without wanting to change budget, a seasonality adjustment is the cleaner instrument. When demand has stepped up for good, neither temporary tool applies — you change the standing target.
Black Friday core
A defined two-to-three-day window with real intent to push volume and accept a looser ROAS bar. It fits Promotion Mode’s 3-to-14-day range and self-closes the Monday after — no manual rollback.
Under 24 hours
Shorter than Promotion Mode’s 3-day floor. Signal the expected conversion-rate spike with a seasonality adjustment and give the campaign manual budget headroom for the day.
New product push
A 5-to-7-day window where you also want the algorithm reaching unproven queries. Schedule Promotion Mode and pair it with Smart Bidding Exploration so the launch finds new converting demand.
Demand stepped up for good
Not an event — a new baseline. A temporary window would expire and snap you back. Lower the Target ROAS to your new economics instead of scheduling a boost you would only have to repeat.
05 — Exploration Reaches PMaxSmart Bidding Exploration, now without a product feed.
Smart Bidding Exploration lets the algorithm bid on queries with unproven conversion histories — expanding reach beyond established performance patterns without forcing you to slacken your global ROAS target. It is the latest step in the longer arc of AI-driven bidding automation in Performance Max. As of June 15, 2026 it is globally available for all Performance Max campaigns without product feeds, across all languages — previously it was limited to Search since its 2025 launch.
The line that matters for planning: only the PMax-without-feeds rollout is generally available. For Shopping ads — both standard Shopping and Performance Max with product feeds — Smart Bidding Exploration is launching in a separate beta. That Shopping beta is arguably the bigger story for e-commerce advertisers, because feed-driven Shopping is usually their highest-volume campaign type. Do not assume Shopping has the feature until you see it live in the relevant account. The same broad-targeting discipline pairs well here as it does with AI Max for Search and the DSA sunset.
Smart Bidding Exploration — Google-reported lifts (PMax / Shopping)
Source: Google internal data from a 31-day test window (Mar 11 – Apr 11, 2025), cited via PPC Land and Google’s own announcement · vendor-stated, not independently verified · results vary by advertiser06 — The 5–30% ROAS BandThe operational spec nobody publishes plainly.
Smart Bidding Exploration runs inside a ROAS tolerance band of 5% to 30%. The tolerance is how far below your Target ROAS the algorithm is permitted to bid when it chases unproven queries. The effective floor is simple arithmetic: floor = Target ROAS × (1 − tolerance). Google’s own example uses a 200% target with a 10% tolerance, which produces a 180% effective floor (200 × 0.90 = 180). Set the tolerance too tight and exploration barely moves; set it too loose and you fund queries that may never pay back.
| Target ROAS | Floor @ 5% tolerance | Floor @ 15% tolerance | Floor @ 30% tolerance |
|---|---|---|---|
| 200% | 190% | 170% | 140% |
| 300% | 285% | 255% | 210% |
| 400% | 380% | 340% | 280% |
| 500% | 475% | 425% | 350% |
Two prerequisites are easy to miss and both block the feature from working as intended. First, Smart Bidding Exploration needs unconstrained budgets — the system alerts you if a campaign is budget-limited, because a budget ceiling stops the algorithm from ever reaching the exploratory queries. Second, you should remove CPC bid limits before turning it on, because bid caps block access to the less competitive queries exploration is meant to find.
How far below target it may bid
You set the tolerance. Floor = Target ROAS × (1 − tolerance). Tighter means less exploration; looser means more reach but a lower efficiency floor on unproven queries.
Or 2–3 conversion cycles
Allow at least 2 to 3 conversion cycles, or a minimum of 7 days, before judging performance. Reading it sooner is reading noise from a system still re-learning.
Recommended A/B duration
Run a structured experiment for at least six weeks to get a stable read. Pair with broad targeting — broad match, Dynamic Search Ads, or AI Max — to give exploration query surface.
07 — The Setup SequenceTurn it on in the right order.
Both features reward a little discipline before you flip the switch. The sequence below is the order we run across managed accounts, and it sits inside our broader paid media services, where peak governance is a standing routine rather than a seasonal scramble. The same broad-targeting hygiene that powers exploration is what we apply when implementing AI Max text guidelines.
Confirm budgets are unconstrained
Smart Bidding Exploration needs headroom — the system alerts you if a campaign is budget-limited. Clear the constraint first, or the algorithm never reaches the exploratory queries the feature exists to find.
Remove CPC bid limits
Bid caps block access to the less competitive queries exploration targets. Remove them before activation so the algorithm can bid into the unproven space your tolerance band permits.
Widen the query surface
Give exploration room to work with broad match keywords, Dynamic Search Ads, or AI Max. For a launch, layer a Promotion Mode window on top so the push and the reach run together.
Then set the tolerance and wait. Choose a ROAS tolerance from the 5–30% band, schedule any Promotion Mode window with its start and end dates, and hold for at least 2 to 3 conversion cycles — a minimum of seven days — before reading the result. For a proper read on exploration, run a structured experiment for six weeks or more. Google Ads coach Jyll Saskin Gales, cited by PPC Land, welcomed the exploration expansion and said she was looking forward to testing Promotion Mode, while urging caution around the separate August 17 change — a fair posture for any beta you are about to put in front of client spend.
08 — ConclusionPlan the peak before the peak plans you.
Promotion Mode rewards the teams that schedule the peak in advance.
Strip away the announcement noise and the peak-and-reach story is tidy. Promotion Mode gives you a self-closing way to spend harder through a defined event without leaving a budget bump live three weeks later. Smart Bidding Exploration gives you a governed way to reach unproven queries inside a tolerance band you control. Both are deliberate levers — the opposite reflex from the August 17 change, which asks you to confirm and hold your targets.
The practical move is to keep them straight. Build a peak calendar, route each event through the decision matrix, and reach for Promotion Mode only when an event fits its 3-to-14-day shape and warrants both looser efficiency and more budget. For sub-day spikes, signal with a seasonality adjustment; for permanent demand, change the standing target. And remember the lift figures are vendor-stated from a single test window — plan from your own account data, not from a headline percentage.
These are betas. Availability and behaviour will shift, the in-product steps are not fully public, and the August 17 change is still an upcoming, gradual rollout. Validate each feature on one campaign before standardising it across a roster, and treat the documented specs here as the floor to test against — not the final word.