Competitive Analysis
AI vs Search Traffic in 2025: What the New Data Means for CPG Brands
Fresh tracking shows Google still dominates referral traffic by a huge margin, but it’s edging down, while ChatGPT’s share (still tiny) keeps compounding. For CPG, that means you cannot ignore AI assistants: treat Answer Engine Optimisation like product work, not blog work.
9 November 2025
8 min read
The headline is stability with a drift. Google continues to send the bulk of discovery traffic, yet month-over-month movement suggests a slow redistribution toward assistant-driven answers. ChatGPT is the clearest grower among AI sources; volumes are small today, but intent quality is often high when the model recommends a product with clear use-case fit. This isn’t an “either/or” shift, think portfolio. You need to win both the classic SERP shelf and the emerging “answer” shelf.
Where are assistants moving the needle already? Content-rich categories with unambiguous attributes (think ingredients, certifications, sizing, or use-conditions) show earlier gains because models can parse and cite them confidently. In replenishment-heavy CPG, once you’re included in a recommended short list, inertia works in your favour; if you’re absent, you’re effectively invisible.
The practical change is from ranking to recommendation. Traditional SEO tuned pages to win positions on a list; AEO tunes product truth so models can resolve a shopper’s task in one answer. That means cleaner PDPs, explicit “who it’s for” and “how to use it,” conservative and verifiable claims, and consistent attributes everywhere your product lives. When the assistant can quote you with confidence, you appear more; when it hesitates, you don’t.
What should teams do this quarter? Ship robust structured data across your site and retail feeds, align Amazon attributes with your own schema, and write answer-ready copy that states benefits, constraints, and safety guidance plainly. Measure assistant-influenced traffic as its own channel and pair every change with a control group of SKUs so you can prove uplift in citations, retailer rank movement, and sell-through. Reproduce tests over several weeks to smooth out seasonality and model drift.
None of this replaces Google hygiene. Classic SEO still pays the bills; the same structured content that helps ChatGPT helps you appear in Google’s AI-style surfaces as well. One investment powers two shelves. The early wins we see come from teams that keep the basics tight while layering in disciplined AEO experiments.
For finance leads, the frame is simple: the risk of inaction is exclusion from compounding recommendation loops; the cost is mostly content and structure you likely need anyway; the payback shows first in higher-intent assistant clicks and second in downstream marketplace performance. Treat it like product: small releases, clear measurement, and repeat.
Our view at LMO7 is that AI search isn’t replacing Google in 2025 but it is quietly reshaping the path to purchase. The brands getting ahead are the ones standardising product truth, writing for tasks not topics, and proving impact with control-vs-test. That’s how you earn AI shelf-space before your competitors realise it exists.
Check out all the data here: https://chatgpt-vs-google.com/
Where are assistants moving the needle already? Content-rich categories with unambiguous attributes (think ingredients, certifications, sizing, or use-conditions) show earlier gains because models can parse and cite them confidently. In replenishment-heavy CPG, once you’re included in a recommended short list, inertia works in your favour; if you’re absent, you’re effectively invisible.
The practical change is from ranking to recommendation. Traditional SEO tuned pages to win positions on a list; AEO tunes product truth so models can resolve a shopper’s task in one answer. That means cleaner PDPs, explicit “who it’s for” and “how to use it,” conservative and verifiable claims, and consistent attributes everywhere your product lives. When the assistant can quote you with confidence, you appear more; when it hesitates, you don’t.
What should teams do this quarter? Ship robust structured data across your site and retail feeds, align Amazon attributes with your own schema, and write answer-ready copy that states benefits, constraints, and safety guidance plainly. Measure assistant-influenced traffic as its own channel and pair every change with a control group of SKUs so you can prove uplift in citations, retailer rank movement, and sell-through. Reproduce tests over several weeks to smooth out seasonality and model drift.
None of this replaces Google hygiene. Classic SEO still pays the bills; the same structured content that helps ChatGPT helps you appear in Google’s AI-style surfaces as well. One investment powers two shelves. The early wins we see come from teams that keep the basics tight while layering in disciplined AEO experiments.
For finance leads, the frame is simple: the risk of inaction is exclusion from compounding recommendation loops; the cost is mostly content and structure you likely need anyway; the payback shows first in higher-intent assistant clicks and second in downstream marketplace performance. Treat it like product: small releases, clear measurement, and repeat.
Our view at LMO7 is that AI search isn’t replacing Google in 2025 but it is quietly reshaping the path to purchase. The brands getting ahead are the ones standardising product truth, writing for tasks not topics, and proving impact with control-vs-test. That’s how you earn AI shelf-space before your competitors realise it exists.
Check out all the data here: https://chatgpt-vs-google.com/