Google's 80% Sales Lift Claim and the AI Ads Reality Check

Brands lifting online sales by 80% with AI-powered ads sounds like the kind of stat that gets screenshot-pasted into board decks. Before yours becomes one of them, it’s worth understanding exactly what that number means — and what it’s quietly standing in front of.

What Google Actually Said (and Didn’t)

At Shoptalk Spring, Google Ads VP of Retail Courtney Rose cited early results showing some brands achieving up to 80% lifts in online sales using AI-powered campaigns, as reported by Digiday. The operative word is “some.” Google’s Performance Max and AI-driven bidding tools have genuinely matured — automated asset generation, audience signal layering, and cross-channel budget allocation are meaningfully better than they were 18 months ago.

But an 80% lift is a best-case outcome, not a median. It almost certainly reflects brands with clean first-party data feeds, well-structured product catalogs, and sufficient conversion volume for the algorithm to learn from. For a mid-size retailer in Manila or Jakarta still running fragmented CRM data into a Google Merchant Center account that hasn’t been audited since 2024, the algorithm has less to work with — and the results will reflect that.

The tactical implication: before scaling AI campaign spend, run a first-party data health check. Feed quality, audience list size, and conversion signal richness are the actual inputs that determine whether you land near that 80% or nowhere close to it.

The Supply Chain Problem That AI Can’t Fix

Here’s the uncomfortable context sitting behind any AI ads conversation: the programmatic infrastructure those ads run through is still deeply compromised. AdExchanger’s recent interview with Danny Spears, COO of publisher alliance Ozone, describes the current moment as a “perfect storm” — signal loss from third-party cookie deprecation, macro economic pressure on CPMs, and traffic quality degradation compounding simultaneously.

Spears uses the phrase “corrosive intermediaries” to describe the layers of SSPs, exchanges, and resellers that extract margin from every impression while contributing questionable value to either advertiser or publisher. Google’s AI may optimise your bidding elegantly across this ecosystem, but it cannot clean up what it’s bidding into. An AI campaign that efficiently buys cheap impressions across a supply chain riddled with made-for-advertising inventory is just a more sophisticated way to waste budget.

For Southeast Asian advertisers, this matters acutely. The region’s open web programmatic supply is thinner and less audited than US or European markets, which means MFA exposure and domain spoofing rates can run higher. Layering AI optimisation on top of an unaudited supply path is putting a precision engine into a leaky tank.

---

What a Disciplined AI Ads Stack Actually Looks Like

The brands realistically achieving outsized results from Google’s AI tools tend to share three structural characteristics that have nothing to do with campaign settings.

First, they’ve consolidated their signal architecture. First-party data from loyalty programs, CRM platforms, and on-site behaviour is flowing cleanly into Google’s ecosystem via enhanced conversions or a Customer Data Platform with a proper Google integration. Shopee and Lazada seller data, where applicable, is being used to model audiences rather than ignored.

Second, they’ve done supply path optimisation before scaling spend. This means using Google’s own Brand Safety controls aggressively, activating publisher allowlists for Display and Discovery placements, and treating Performance Max’s “URL expansion” feature with appropriate suspicion until you’ve validated which placements it’s actually landing on.

Third, they’re treating AI campaign outputs as hypotheses, not conclusions. The algorithm will tell you what’s converting — your job is to interrogate whether those conversions are genuinely incremental or whether you’re paying a premium to reach customers who would have found you anyway. Incrementality testing, even simple geo-based holdout experiments, should be running in parallel with any AI campaign at meaningful scale.

The Broader Strategic Moment

Spears’ “perfect storm” framing from the AdExchanger piece is worth sitting with longer than the AI ads headline. Signal loss isn’t a temporary inconvenience — it’s a structural reset of how digital advertising infrastructure works. Third-party cookie deprecation in Chrome, iOS ATT enforcement, and the slow fragmentation of cross-platform identity resolution are collectively making the old targeting playbook obsolete.

Google’s AI tools are, in part, a response to this — they’re designed to perform without the granular individual-level signals that programmatic used to rely on. That’s genuinely useful. But it also means the brands that will compound gains from AI advertising are the ones investing now in the first-party data foundations and clean-room infrastructure that give the algorithm something real to work with. In Southeast Asia, where mobile-first audiences are distributed across LINE, TikTok, Shopee, and Grab ecosystems, that data foundation is harder to build — but the competitive advantage of having it is proportionally larger.

The 80% headline is real for somebody. The more interesting question is whether your data infrastructure, supply path hygiene, and measurement discipline are positioned to make you that somebody — or whether you’re just about to spend more money, faster, on the same underlying gaps.

---

Key Takeaways

• Audit first-party data quality and conversion signal richness before scaling Google AI campaigns — feed quality determines whether you approach the 80% ceiling or miss it entirely.
• Run supply path optimisation in parallel with AI campaign activation; algorithmic efficiency cannot compensate for compromised programmatic inventory.
• Build incrementality testing into your AI campaign measurement from day one — the algorithm optimises for reported conversions, not genuine business growth.

The AI advertising era isn’t coming — it’s already the default operating environment. The brands pulling real returns from it aren’t the ones with the biggest budgets; they’re the ones that treated data infrastructure and supply chain integrity as prerequisites, not afterthoughts. The question worth asking your team this quarter: if Google’s algorithm is only as good as what you feed it, what exactly are you feeding it?

---

At grzzly, we work with growth teams across Southeast Asia on exactly this intersection — first-party data strategy, programmatic supply path audits, and building the measurement infrastructure that makes AI advertising actually accountable. If you’re scaling Google AI campaigns and want a clear-eyed read on whether your stack is set up to deliver, Let’s talk.