Google: Target And Budget Changes Do Not Reset Smart Bidding Learning

By Elena Vasquez

Changing a tCPA target, a tROAS target or a budget on a steady-state Google Ads campaign does not trigger "Learning" status and does not erase bidding history, Product Manager Tal Akabas said in an interview on Google's Ads Decoded podcast. Ads Product Liaison Ginny Marvin hosted the roughly 40-minute conversation. It covered a wave of bidding and budgeting updates ahead of Black Friday Cyber Monday.

Here is what PPC teams need to know from the episode.

What does Google say about target changes?

Akabas addressed the learning myth directly at 25:14. On an existing, steady-state campaign, a target or budget change does not wipe out historical data or trigger "Learning" status. He explained that the backend already holds bid landscape data (the model of how bids perform across auction prices) and can scale bids in proportion to target changes.

Caveats remain. New campaigns still need ramp-up time for their mix of assets, targeting and goals, Akabas said at 26:03. New or reactivated bid strategies, composition changes and significant conversion-goal changes can trigger "Learning" status. For new conversion-oriented campaigns, Google advises starting with Maximize Conversions or Maximize Conversion Value rather than Maximize Clicks.

Performance can still fluctuate as bids and traffic mix change. The guidance: allow for conversion delay and avoid repeated target changes within one conversion cycle without a clear business need.

Should advertisers use Seasonality Adjustments for BFCM?

No, for standard holiday spikes. Akabas said at 29:12 that advertisers "don't necessarily need them just because it's BFCM." Google's year-over-year models already anticipate normal holiday surges.

A Search Engine Journal study of BFCM accounts from 2022 to 2024 supports restraint. Accounts using seasonality adjustments saw CPCs rise roughly twice as much as accounts that did not. The adjustment cohort's ROAS fell in each year studied, though its revenue grew faster. The comparison is observational and does not isolate causation.

Save adjustments for anomalies the AI cannot anticipate, such as an off-cycle flash sale or a discount far deeper than last year's.

What is Promotion Mode and Smart Bidding Exploration?

Promotion Mode is rolling out to Search and Performance Max, Akabas said at 19:20. It schedules a promotional window with dedicated budget and bid changes, combining faster ramp-up with Smart Bidding Exploration during the window.

Smart Bidding Exploration, now rolling out to Retail PMax and coming to Shopping (17:13), gives the algorithm a small ROAS tolerance—typically around 10% effective ROAS flexibility per Google's FAQs—to bid into new queries and incremental auctions. The requirement: an eligible tROAS campaign with an unconstrained budget. If the budget becomes constrained, the mechanism is disabled.

How do Channel Prioritization and Budget Allocation work?

The new controls cover Search, Shopping, Performance Max and Demand Gen (01:08). Channel Prioritization sorts channels into High and Low Priority tiers. Budget Allocation adds an optional, best-effort daily spend ceiling for Low Priority channels. All channels start as High Priority.

The mechanics depend on whether budget limits spend, as Marvin clarified on LinkedIn. With a budget limit, spend flows to High Priority channels first. Budget shifts to Low Priority channels only when they are predicted to convert at or better than the target.

Without a budget limit—Google's recommended setup—Channel Prioritization changes nothing, because Smart Bidding already pursues every eligible conversion at your target. Akabas noted at 03:37 that it still works on any day the budget limits spend. This creates what the source calls a catch-22: Google recommends uncapped, demand-led budgets with tCPA and tROAS (32:26), but an uncapped budget leaves Channel Prioritization nothing to do.

Advertisers on uncapped campaigns must use Budget Allocation to restrict Low Priority channels. A Low Priority channel set to 0% receives no spend, per Google's help documentation. In Performance Max, where individual channel opt-out is unavailable, that functions as the opt-out.

Several fine-print details warrant attention:

• The "ghost budget" trap. Allocation percentages apply to the average daily budget setting, not actual spend. A 10% Display cap on a $5,000 budget that normally spends $500 a day still permits $500 in daily Display spend—100% of typical volume.
• Best-effort enforcement. Caps reset daily without rollover and may be exceeded during traffic volatility.
• Shared budgets. Channel Prioritization works with shared budgets only if every campaign is the same type.
• Verification. Check cost by channel in reporting, including the "Where your conversions come from" Sankey chart in PMax Channel performance reports.
• Experiments. Both controls work in Custom Experiments with restrictions; shared budgets are unsupported and all channels sit on one tier.

One discrepancy: Google's help doc says Budget Allocation is not supported for Demand Gen, yet Akabas suggested at 04:52 that Demand Gen advertisers who opted out of Display could test it back in at lower priority or with an allocation.

What changes in February 2027?

Google replaces the existing Display Expansion allocation behavior for Search campaigns starting February 2027, with the episode describing a February-March transition (09:09). Today, Display Expansion treats Display as a low-priority outlet for unused Search budget (07:17). Advertisers may see spending shift between Search, Search Partners and Display.

Two fixes exist: a one-click migration button Google plans to add, or manually setting Display to Low Priority with a 50% allocation. Marvin confirmed on LinkedIn that the manual setting preserves current behavior.

Action item: audit manager accounts now, flag legacy campaigns with Display Expansion, and check for Target Impression Share campaigns—Google says the Display Network will no longer serve on those after the change.

What about Customer Lifecycle Goals?

Three updates make lifecycle data easier to use:

• NCA reporting-only mode (13:15): measure new versus returning customers without bidding on them.
• Re-engagement only mode (16:04): live in PMax, coming to Search and Shopping; focuses budget on churned or disengaged customer lists.
• Loyalty mode (16:37): generally available for PMax and Shopping; pairs loyalty lists with differential bidding and member pricing callouts.

How can teams bridge the CFO-CMO gap?

Demand-led budgeting pairs uncapped budgets with strict targets (30:10). Akabas said at 31:45 that missing efficient conversions early can mean paying more for them later. Three steps help reconcile this with fixed finance budgets:

1. Anchor targets in real financials. Ground conversion values in margins and customer lifetime value, or added headroom scales unprofitable spend.
2. Use filtered planning tools. Performance Planner (35:25), the Missed Opportunities Report (37:23) and Recommended Investment Strategy V2 (36:25) now allow filtering by conversion goal or campaign. Google Analytics cross-channel budgeting tools connect to Meta, Pinterest, Reddit and Snap (38:42).
3. Simulate with pacing guardrails. Give campaigns daily headroom on peak days, then layer scripts or rules that alert teams, dial back caps or raise tROAS when spend runs ahead of monthly trajectory.

The practical conclusion for the season: leave steady-state targets alone, skip Seasonality Adjustments for standard spikes, verify that budget settings let Channel Prioritization and Exploration function at all, and put the February 2027 Display migration on the calendar.