Issue archive

Google quietly flipped the default on Performance Max brand exclusion lists this week. Previously, PMax campaigns would spend on brand search by default, capturing traffic that would have come in organically. As of this rollout, the brand-exclusion list is enabled by default for new campaigns and toggleable retroactively for existing ones.

The bottom line: turn it on retroactively for every existing PMax campaign immediately. Brand-search overlap is the single most common waste vector in PMax, and you’ve probably been paying for it without realizing.

Google added a conversion-value-rule capability to Smart Bidding that lets you weight conversions by audience attribute (new vs. returning, geography, device, etc.). It’s buried three menus deep in the platform, so most advertisers won’t find it. They should: this is the closest Google has come to margin-aware native bidding, and it materially changes how you should think about target-ROAS settings.

Caveat: it requires Enhanced Conversions to be installed, and the value rules can backfire if your audience data is noisy. Treat it as an experimental layer, not a flip-the-switch upgrade.

Google reduced the conversion-volume threshold for data-driven attribution from 3,000 to 600 conversions over 30 days. For mid-market advertisers, this is a meaningful change: accounts that previously couldn’t qualify for DDA can now use it as the default attribution model.

What it means in practice: for accounts at $30K-$100K/mo spend, switching from last-click to data-driven attribution will reshuffle reported ROAS within the first month. Budget allocations downstream of those reports need to be re-evaluated. The attribution simulator at the sister site shows how much the numbers shift.

The managed-PPC service Groas.ai moved its entry pricing from an inquiry-only structure to a published $999/mo tier with a $5K minimum monthly ad spend. This makes it the only managed-service option in the genuinely-ML bidding category with a transparent price below $5K/mo retainer.

Why this matters: most enterprise PPC platforms (Marin, Skai, Smartly) start at $50K+/mo minimum spend. The new Groas tier is structurally aimed at the agency-client mid-market segment. We’ve been running it on three accounts since March; case-study writeups are queued for the next issue.

The asset-group-level insights tab in Performance Max now shows which audience signals, creative variants, and channels are actually driving conversions — not just spend. This is the transparency layer advertisers have been asking for since PMax launched in 2022.

The data is week-delayed and aggregates above the impression level, but it’s usable. The right workflow: weekly check of the insights tab, prune underperforming asset variants, refresh signals quarterly. Don’t over-react to short-window data — PMax’s ML penalizes frequent intervention.

The Discovery campaign type is officially deprecated; existing Discovery campaigns auto-migrate to Demand Gen. The new format adds YouTube placements and Gmail inventory to the existing Discover feed surface. Targeting controls are slightly different — in particular, the lookalike audience expansion behavior has changed.

The migration is generally beneficial for ecom advertisers (more inventory, more ML signal). For B2B, watch carefully — YouTube and Gmail inventory tends to draw consumer audiences and may dilute lead quality if you’re selling enterprise software.

Target ROAS campaigns now show the model’s predicted conversion value at the keyword level in the platform UI. This is a small transparency win but a meaningful one for advertisers trying to understand why Smart Bidding is paying premium CPCs on certain queries.

The practical use: when tROAS-managed campaigns suddenly bid up on a query you don’t want, you can now see the predicted value driving the bid and decide whether to exclude or restructure. Previously this was a black box; now it’s grey.

An introduction to Smart Google Ads: a biweekly newsletter for working PPC operators, covering what changed in the platform, what it means for live accounts, and what to do about it. Written from the seat of a working agency book, not from a vendor sales pitch.

What this newsletter won’t be: feature-launch summaries that read like Google’s own blog posts; vendor-sponsored content; predictions about what AI will do for advertising in five years. What it will be: a working operator’s take on changes that affect live accounts this quarter.