Product Designer - Maternity Leave

AI shopping attribution before retailers shift spend. By DataTip · Published September 9, 2026 TL;DR: Retailers should not treat AI shopping traffic as incremental acquisition until they define how discovery, recommendation, referral, assisted revenue, direct conversion, cannibalized demand, and completed purchase are credited. Reported plans by NIQ and Similarweb to measure ChatGPT and Gemini activity, plus Salesforce-reported Canadian usage, show why governance is needed but do not prove incremental performance. * An AI referral records a measurable touchpoint, not automatic evidence of new demand. * Keep discovery, recommendation, referral, assisted revenue, direct conversion, cannibalization, and incrementality as separate attribution categories. * Treat NIQ and Similarweb measurement plans as emerging inputs, not accepted industry standards. * Do not shift acquisition spend until comparison rules define how AI activity is evaluated against existing channels. AI shopping attribution is becoming a budget question, not merely a reporting question. When an AI assistant helps someone discover a product, recommends it, sends a referral, or contributes to a completed order, what should receive acquisition credit? Retailers should not treat AI shopping traffic as incremental acquisition until they define how discovery, recommendation, referral, assisted revenue, direct conversion, cannibalized demand, and completed purchase are credited. Reported plans by NIQ and Similarweb to measure ChatGPT and Gemini activity, plus Salesforce-reported Canadian usage, show why governance is needed but do not prove incremental performance. The answer matters because an AI-originated visit does not automatically represent new demand. It may assist a purchase that would have arrived through another channel, or influence discovery without producing a transaction. Retailers need consistent definitions before shifting acquisition spend. What are NIQ and Similarweb planning to measure on ChatGPT and Gemini? NIQ and Similarweb are reportedly developing measurement systems for shopping activity on ChatGPT and Gemini, with availability targeted for Q4 2026. That suggests AI-assisted shopping is moving from an emerging concept toward a planned measurement category. It does not establish a current industry standard or prove incremental performance. Reported coverage describes the planned measurement work. The supplied source material does not provide the methodologies, definitions, sample sizes, or detailed findings behind those systems. That boundary matters. A future reporting product may help retailers observe AI shopping activity, but measurement availability and budget attribution remain separate decisions. Why might AI influence discovery before retailers can measure it consistently? Reported consumer usage indicates that AI may already affect product discovery before comparable measurement is widely available. Retail Insider, citing Salesforce research, reports that nearly 4 in 10 Canadian shoppers are using AI for shopping. This is a reported Canadian finding, not a universal adoption rate, and it does not establish that AI use leads to completed purchases. The commercial question is not simply whether AI shopping exists. It is whether retailers can distinguish an AI-assisted touchpoint from genuinely incremental acquisition. Without that distinction, a new channel label may only relocate credit from search, direct traffic, marketplaces, or another existing source. The activity may be real. Its budget meaning is not yet settled. What should AI shopping attribution distinguish? Retailers should keep the stages of influence separate rather than treating every AI touchpoint as a conversion. Discovery, recommendation, referral, and completed purchase describe different events, and each can have a different relationship to demand. A practical attribution vocabulary should distinguish: * AI-assisted discovery: AI helps a shopper identify a product or category, without necessarily producing a referral or purchase. * AI recommendation: An AI environment suggests a product. That may influence consideration without creating a measurable visit. * AI referral: A shopper reaches a retailer through an AI-generated link or identifiable handoff. This records traffic source, not proof of new demand. * Completed purchase: The shopper finishes an order after an AI-related interaction. The retailer still needs to assess whether another channel or existing intent would have produced it. * Assisted revenue: Revenue associated with an AI touchpoint that helped the journey but did not necessarily cause the transaction. * Direct conversion: A completed purchase credited to an AI interaction under defined rules. Direct credit is not the same as incrementality. * Cannibalized demand: A purchase credited to AI activity that would likely have occurred through an existing channel or without the AI touchpoint. * Incremental acquisition: Demand that would not have occurred without the relevant activity, established through a comparable framework rather than assumed from referral data. This avoids a basic category error: confusing visibility with causation. An AI referral can be observable and still be assisted, cannibalized, or inconclusive. Why does measurement availability not prove incremental acquisition? A new measurement system may show that shoppers interacted with ChatGPT or Gemini. It cannot, by itself, show that those interactions created additional demand. Reporting describes what was observed; attribution governs what receives budget credit. The supplied coverage does not establish how NIQ or Similarweb will define an AI shopping event, connect it to a purchase, separate new demand from existing intent, or compare it with other acquisition sources. Retailers should apply the same discipline they would use for any other acquisition decision. What should retailers establish before reallocating spend? Before treating AI shopping traffic as an incremental acquisition channel, retailers should document what each touchpoint means and how it compares with existing channels. The goal is not to reject AI commerce measurement. It is to prevent an emerging data source from becoming an unexamined budget claim. The governance standard should specify: * which events qualify as discovery, recommendation, referral, assistance, and conversion; * when AI-influenced revenue receives partial or full credit; * how potential cannibalization is identified and reported; * what evidence is required before activity is classified as incremental acquisition; and * how future measurement systems will feed the framework without replacing it. NIQ and Similarweb's reported plans, together with the Salesforce figure reported for Canadian shoppers, show why AI commerce measurement deserves attention. They do not settle the attribution question. Retailers should use emerging measurement systems as inputs to documented governance, not as proof that AI-generated traffic adds demand. Key takeaways. * An AI referral records a measurable touchpoint, not automatic evidence of new demand. * Keep discovery, recommendation, referral, assisted revenue, direct conversion, cannibalization, and incrementality as separate attribution categories. * Treat NIQ and Similarweb measurement plans as emerging inputs, not accepted industry standards. * Do not shift acquisition spend until comparison rules define how AI activity is evaluated against existing channels. Practical tips. * Create a shared attribution glossary that marketing, finance, analytics, and commerce teams use consistently. * Report AI-assisted activity separately from completed purchases so influence is not confused with conversion. * Flag AI-attributed demand whose counterfactual outcome is unknown rather than assigning it full incremental credit. * Review future measurement-system outputs against documented rules before using them in budget decisions. Before AI shopping activity affects acquisition budgets, document the definitions and comparison rules your teams will use to distinguish assistance from incremental demand.

NIQ and Similarweb advance Agentic Commerce Measurement for the AI shopping era. New solution will connect AI-driven discovery to consumer intent, product content, AI-driven traffic and sales conversion CHICAGO - NIQ (NYSE: NIQ), a leading consumer intelligence company, today announced further advancements in its Agentic Commerce Measurement strategy through a collaboration with Similarweb (NYSE: SMWB). The companies are collaborating on a new solution to help brands, retailers...

Similarweb launches AI Ads Intelligence for ChatGPT and Google AI. Rather than leaving advertisers to guess where their ads are appearing, Similarweb's AI Ads provides a direct look at what's winning placement. Similarweb has announced its newest Ad Intelligence data set, AI Ads, that fills an important gap for advertisers trying to understand the changing competitive landscape for ads in ChatGPT and Google's AI Mode and AI Overviews. With AI Ads, Similarweb customers can see a competitive analysis of ads on the most used conversational AI platforms, ChatGPT, Google AI Mode & Google AI Overview, worldwide and for specific markets. "This is the rare moment when a major ad channel is still wide open. The advertisers who can see what's happening with AI ads now will have a real head start. Until today, nobody could see it at all," said Harel Amir, General Manager & Head of Product for Similarweb Ad Intelligence. To-date, 26% of ChatGPT responses already carry a sponsored ad (shown to users in the ChatGPT Free and Go tiers), and nearly 30% of ad-eligible Google AI Mode queries already show ads as well. Finally, over 40% of Google searches now trigger an AI Overview, and those AI Overviews often include ads. Why It Matters Most major ad channels have a public transparency layer: Meta has the Ad Library. Google has its Ads Transparency Center. TikTok has the Creative Center. ChatGPT, AI Mode, and AI Overview, however, have none of that - not for competitors, and often not even for the advertiser's own team. Adding AI Ads to Similarweb's Ad Intelligence fills that gap with an Ads Gallery that shows what ads are running and in what context. For AI Mode & AI Overview, advertisers participating in Google's broader digital advertising programs are blind to ad placements in Google AI answers. Most advertisers don't know whether their own ads are running in that context, let alone how they stack up against competitors. Google's visibility tools don't answer that question. For ChatGPT, ads only show up when intent exists and is clarified from the conversation, and the decision comes down to the combination of ad creative, copy, and landing page. Rather than leaving advertisers to guess where their ads are appearing, Similarweb's AI Ads provides a direct look at what's winning placement. What makes Similarweb's approach unique is that these insights are drawn from real user panel conversations, not synthetic prompts. The upcoming releases of the AI Ads dataset will add advertisers' true share of voice, ad categories, and conversational intent. "When we started advertising on ChatGPT, we were flying blind, no visibility into who else was in the auction or what was working," said Jonathan Bar Vardi, Head of Strategy at Natural Intelligence, a data-driven marketing specialist that has become a top buyer of ChatGPT ads. "Similarweb changed that. It shows what other advertisers don't know: who's spending, where they're appearing, and what's driving performance." Sophia Bennett is a news curator at Martechvibe, covering global developments across marketing technology, digital transformation, AI, data, and customer experience trends. View More