📡Digital Marketing in 2026
What changed this year: AI search, signal loss, agentic commerce and the measurement that survived · 28 terms
AI Search
8How AI tools answer searches, how brands are seen and linked in those answers, and the bots that read sites for them.
AI Answers
3Searches that end in a written answer, and how that answer is put together from many searches.
A search that ends on the results page. The person gets the answer and clicks nothing.
A chat tab in Google Search. It answers long questions in words, with source links.
One question split into many small searches. They run at once and feed one answer.
AI Search Metrics
3The numbers for AI search: being named, being linked and the visits that follow.
How often AI answers name your brand. You count it across many buyer questions.
A source link inside an AI answer. It points to a page the answer used.
Visits to your site that come from links inside AI answers. Your reports often name the AI tool as the source.
AI Crawl Control
2Files and rules that decide which AI bots may read a site, and what they are pointed to.
A bot that fetches web pages for AI tools. Some train models, and some answer questions.
A Markdown file at a site's root. It points AI tools to the pages that matter most.
Signal Loss
7Ad tools seeing less of the buyer's path, the privacy changes behind it, and how the gap is filled.
Ad tools seeing less about people. Blocked cookies and privacy prompts hide part of the path.
Browser Privacy
2Browser cookies that follow people across sites, and the plans to show ads without them.
A cookie set by another site's tracker. It can follow you from one site to the next.
Google's set of private ad tools for Chrome and Android. Google said in October 2025 it would retire most of them.
Mobile App Privacy
2Phone rules that make apps ask before tracking, and the private way installs from ads are counted.
Apple's rule that apps must ask before tracking you. Saying no hides your ad ID.
Apple's way to count app installs from ads. It never names the person.
Conversion Modeling
2Estimating the sales that tracking could not see, and passing each visitor's consent to the tags.
Sales an ad tool estimates because it could not see them. It adds them to the report.
Passes each visitor's cookie choice to Google's tags. Tags then act on that yes or no.
Agentic Commerce
7Shopping done by AI agents for people, selling through chat, and the product data agents read.
An AI agent shops for a person. It finds, picks and pays, within set limits.
Selling through chat. The buyer asks, gets advice and buys in the same conversation.
An AI agent that shops for a person. It compares, picks and can pay once approved.
A file that lists every product, with its price, stock and link. Shopping tools read it.
Commerce Protocols
3Open standards that let agents find products, check out and pay at shops they were never wired to.
An open standard from OpenAI and Stripe. It lets AI agents check out at a shop.
An open standard for payments made by AI agents. Signed records prove what the person asked for.
An open standard from Google and big shops. Agents can buy at any shop that uses it.
Marketing Measurement
3Ways to judge marketing that still work with less tracking: tests, mix models and using them together.
A mix model, lift tests and ad tracking, used together. Each one checks the others.
An ad platform's test of extra sales. Random people see the ads, and others are held back.
Google's free tool for building a marketing mix model. Teams run it on their own data.
AI Advertising
3Ads made by AI, campaigns run by the platform's AI, and the ad itself choosing who sees it.
Ads written or drawn by AI. A tool makes many versions from one brief.
The ad itself picks its audience. The platform shows each ad to the people who react.
A campaign where the platform's AI picks the people, the places and the bids. You only set the goal and the budget.