Your tCPA & ROAS Are Going for a Ride. Are You Ready?

Welcome back to the front row.

Two Google announcements dropped last week. One has a hard deadline in 56 days. The other quietly ended a two-year industry debate. Most people running accounts and content strategies missed both.

This Week

There's a version of the next two months where everything looks fine — traffic holding, CPAs stable, GEO strategy ticking along. Then August hits and two things shift simultaneously. Your bidding system starts honoring the targets you set years ago and forgot. Your AI search visibility strategy turns out to be built on tactics Google just officially dismissed.

Neither of these is a crisis. But only if you move now.

MAIN STORY

Fasten your seatbelts, Google is taking your tCPA and ROAS on a ride. The Question Is Whether Your Account is ready for it.

Something interesting happened last week. Google made two separate announcements — one on bidding, one on AI search — that most practitioners are treating as unrelated news items. They're not. They're the same message, delivered twice.

Google is getting better at honoring exactly what you told it to do. The problem is that what you told it to do might be two years old.

The August 17 Problem

Let's start with the one that has a deadline.

On June 15, Google announced changes to how Smart Bidding handles budget-limited campaigns. Starting August 17, if your campaign has a "Limited by budget" status and runs on Target CPA or Target ROAS, the system will start delivering more consistently toward the target you actually set — not the performance it happened to achieve.

Google uses a clean example. If your Target CPA is $10 but the campaign has been delivering conversions at $5, after August 17 it will move toward $10.

Read that again.

The campaign was beating your target. By a lot. And now the system will correct toward the number you originally typed.

This sounds counterintuitive. But here's what it actually reveals: a large number of accounts have been running on targets set during a launch brief, a competitive push, or a client onboarding call that happened somewhere between six months and two years ago. The system kept over-delivering, performance looked strong, nobody reviewed the target. The number just sat there — technically wrong, operationally invisible.

Google is not doing something aggressive here. It's doing something precise. It is enforcing the contract you wrote with yourself.

The Bid Target Adjustment Tool becomes available on July 6. It surfaces your affected campaigns and gives you three clear options: keep the target as-is and accept the shift in delivery, match it to recent actual performance, or set a custom target that reflects where the business actually is now. You have from July 6 to August 17 to make this call. That's six weeks. It's enough time if you start now and not enough time if you treat this as a July problem.

Here's the practical sweep: pull every campaign currently showing "Limited by budget" status. Cross-reference your Target CPA or Target ROAS setting against actual performance in the last 90 days. Any campaign where actual performance is meaningfully better than the stated target is affected. Those are the accounts where August 17 is not a background change — it's a live performance event.

One more thing worth noting: this is not just a Search issue. The change applies across Search, Shopping, Performance Max, Demand Gen, and Travel campaigns. For Performance Max and Demand Gen specifically, you may also see shifts in how traffic is distributed across surfaces within the campaign. The system optimizing more precisely toward a target affects the mix, not just the volume.

The GEO Debate Google Just Ended

The same week, Google published its first official guidance on optimizing for generative AI features in Search. It was blunt.

If you've been buying strategies built around llms.txt files, content chunking for AI parsing, or services promising to optimize your content specifically for LLMs rather than humans — Google's guidance is direct: you can ignore all of it. Those tactics are not how Google Search decides what to surface in AI Overviews or AI Mode.

This matters because the GEO industry has been moving fast and selling confidently. New terminology, new tools, new consulting engagements. Some of it is genuinely useful. But a meaningful portion of what's been sold as AI-specific optimization is what Google just described as unnecessary or ineffective for its own products.

Google's actual position is simpler and, honestly, more useful: generative AI visibility on Google Search is still SEO. AI Overviews and AI Mode run on the same core ranking and quality systems as regular Search. If your content is well-structured, topically authoritative, technically accessible, and genuinely useful — it's already positioned for AI surfaces. If it isn't, no amount of AI-specific markup will fix it.

That's not a dismissal of GEO as a discipline. It's a clarification about what the discipline actually requires. The real work is the same: entity clarity, structured data that means something, content that directly answers questions, and the kind of authority signals — third-party coverage, consistent brand presence, cited expertise — that Google's systems have always valued.

Where GEO becomes genuinely distinct from traditional SEO is in the citation layer. A field experiment found that AI Overviews reduce outbound organic clicks by 38% on triggered queries. Zero-click rises. But brands actually cited inside those Overviews see 35% more organic clicks and 91% more paid clicks. Being in the answer is structurally different from ranking below it.

That's not an llms.txt problem. That's a content authority problem. And it's one that Google's core guidance has always addressed.

The Pattern

Here's what connects these two announcements.

Google built automation that works with your inputs — your targets, your content, your signals. For years, the system was forgiving enough that stale inputs didn't cost you much. Targets slightly off, content not quite structured right, authority signals inconsistent. The automation absorbed a lot of that slack.

What's shifting is the precision. The system is getting better at doing exactly what you told it — which means what you tell it matters more than it used to.

For advertisers: the target you set should reflect where your business actually is right now, not where it was when you launched the campaign.

For content teams: the authority you build should be real, structured, and consistently signaled — not engineered for a loophole that Google just confirmed doesn't exist.

Both announcements are Google getting sharper. The practitioners who are already operating with intention — clean targets, real authority, structured content — won't feel much. Those running on inherited settings and shortcut frameworks are the ones with work to do before August.

PROMPT OF THE WEEK

You've got Google's Bid Target Adjustment Tool coming on July 6. Before it does, get ahead of your own audit.

Use this prompt to structure the review:

"I'm auditing my Google Ads campaigns before August 17, 2026. I'll give you a campaign name, its current Target CPA or Target ROAS, and its actual performance over the last 90 days. For each one, tell me: is the gap significant enough to require action before August 17, what the risk is if I leave it unchanged, and what target I should set to maintain current performance levels. Here's the first campaign: [paste campaign name, target, actual CPA/ROAS, budget status]."

What this gives you: a structured pre-audit you can run through your whole account before July 6, so you walk into the tool knowing exactly what to adjust and why.

THIS WEEK

Google's ToS updated for the first time since 2018. Effective July 1, Google's Ads Terms of Service now explicitly state that advertiser-provided inputs — including what you type into conversational tools like Ask Advisor — may be used across AI features to improve campaign performance. Advertisers remain responsible for reviewing and approving assets generated automatically. The operational implication: treat every input into Google's AI tools as intentional campaign direction, not a casual question.

ChatGPT self-serve ads now open to SMBs. OpenAI launched a beta self-serve Ads Manager in May, and Criteo has since reduced the campaign minimum to $10,000. The format — sponsored units below responses, clearly labeled — is still early, but the CPC pricing model makes performance accountability cleaner than the original $200K impression-based entry. Worth a test budget for brands whose audience skews toward ChatGPT's Free and Go tiers.

Reddit is now a GEO strategy, not a social media strategy. Tinuiti's Q1 2026 data found Reddit accounting for roughly 24% of Perplexity's total AI citations. Brands with high Reddit presence are four times more likely to be cited by AI systems than those without community activity. This isn't about running Reddit ads — it's about whether your brand exists in the spaces where AI systems look when they need to verify what practitioners actually think about a category.

Meta AI business assistant now live globally. Meta expanded its AI business assistant in Ads Manager to all advertisers worldwide, alongside a one-click Conversions API setup that requires no technical implementation. If your Meta pixel setup has been on the to-do list for months, that excuse is now gone.

TOOL OF THE WEEK

Google's Bid Target Adjustment Tool — available July 6, 2026, inside Google Ads.

This is not a third-party tool or a workaround. It's Google's own instrument for navigating the August 17 change, built into your account dashboard.

How to use it right now: Before July 6, go to your Campaigns view, filter by "Limited by budget" status, and note which campaigns are running Target CPA or Target ROAS. Run a 90-day performance report on each. Calculate the gap between the stated target and actual delivery. When the tool drops on July 6, you'll have your list ready instead of spending the first week of July figuring out which campaigns are even affected.

Why it matters this week specifically: every day you don't start this review is a day closer to August 17 with less time to observe the impact of any target changes before the system update goes live.

MY TAKE

I've been running Google Ads accounts for years. And the honest thing I can tell you about the August 17 change is that the accounts it hits hardest are the ones that are performing well with no supervision.

That sounds wrong. But think about it. The campaigns that will be affected are the ones that felt fine — low CPAs, stable performance, no alerts. Nobody reviewed the target because nothing was broken.

The accounts that needed constant attention, that were struggling to hit targets — those are already calibrated. It's the overperformers that have stale numbers sitting in the target field.

That's the uncomfortable thing about precision. It doesn't punish bad accounts. It punishes the accounts where success made review feel unnecessary.

Check your overperformers this week. Not the ones you're worried about. The ones you stopped worrying about.

That's Issue #10 of NVM. If this gave you one thing to act on this week, share it with someone who needs it. Reply to this if you have a question — I read every one. See you next week. — Mostafa

NVM · Front row seats to AI and digital marketing

• Subscribe → nvm.marketing/newsletter
• Join the community → nvm.marketing/community
• Write for NVM → nvm.marketing/blog