Daily Edition · Page 8

THE DAILY INDEX

Technology, AI, software, security, crypto and business

NEWS & TRENDS

TLDR Marketing

MIXED REPORTS ON AI ECOMMERCE TRAFFIC

AI-driven ecommerce traffic shows strong buying intent, but results are inconsistent across studies. Adobe reports AI visitors were 42% more likely to purchase and generated 37% more revenue per visit, while they spent 48% longer on site, viewed 13% more pages, and bounced 32% less. Other research shows AI traffic is still small at under 0.2% of total visits and sometimes converts worse than channels like organic search or paid search. Differences come from measurement methods, definitions of AI traffic, geography, and timing, which makes early data easy to misread.

THE GHOST CITATION PROBLEM

AI search often uses brand content without naming the brand. 62% of citations are “ghost citations” where a link appears but no brand mention is included. Across 3,981 domains, 75% were cited but only 38% were mentioned, and just 13% achieved both outcomes. Different AI engines behave very differently, as Gemini mentions brands 83.7% of the time but rarely cites them, while ChatGPT cites 87.0% of the time but mentions only 20.7%. Strong consumer brands are frequently named in answers, while sources like Wikipedia are cited often but rarely mentioned.

TIPS TO INCREASE LINKEDIN REACH

Align LinkedIn posts tightly with a clear topic that matches the profile so the algorithm can reinforce perceived expertise. Use carousels, polls, short vertical video, and newsletters, since text posts now underperform. Optimize for strong engagement signals like saves, long comments, DM shares, and 60+ second dwell time instead of likes. The first 60-90 minutes are critical, since they drive about 70% of total reach. Reply to comments within 30 minutes. Posts can circulate for 2-3 weeks, and saves 24-72 hours later often perform 4-6x better. Keep post length around 1,300-1,900 characters, since under 500 characters can reduce reach by 35%. Avoid external links, engagement pods, heavy hashtags, over tagging, and frequent posting, which can reduce distribution.

DON'T USE GENERIC LINKEDIN JOB TITLES

LinkedIn headlines that only list a role can be too broad and may miss more specific recruiter searches that combine function, channel, and specialty. Recruiters often search terms like “growth marketing + paid social” or “digital marketing + SEO”, so generic titles reduce visibility. Adding context helps match these searches, using a format like role + channel or focus + specialty/tools + industry. For example, instead of “Growth Marketing Manager”, use “Growth Marketing Manager | Paid Social & Acquisition | Meta, Performance Marketing | DTC” or “Growth Marketing Manager | Lifecycle & Retention | Email, CRM, Customer Journeys | B2B SaaS”. This makes profiles easier to find and more aligned with how recruiters actually search.

LINKS IN AI MODE NO LONGER DIRECT USERS EXCLUSIVELY TO WEBSITES

Google is changing AI Mode in Chrome so links no longer send users fully to websites. Instead, clicking a result opens a side by side view with the webpage on one side and AI Mode on the other, keeping users inside the LLM experience. Users can still navigate the site, but the AI sidebar remains active throughout. This shift lets Google layer its AI experience on top of publisher content surfaced in search.

THREADS WILL ADD BROADCAST CHANNELS-STYLE FEATURE CALLED LIVE CHATS

Threads will roll out live chats to offer a more controlled way to follow real-time discussions around events and topics. The feature lets creators start a chat, invite a few contributors, and allow followers to subscribe in a read only format. It builds on Instagram Broadcast Channels from 2023 but adds more active conversation from selected participants. It aims to reduce spam seen in hashtag-driven feeds by limiting who can post.

HOW MARKETING LEADERS CLOSE THE MMM ‘ACTIONABILITY GAP'

Marketing mix modeling is widely used but often fails to drive action. Leading organizations close this gap by turning MMM from a backward looking report into a forward system that guides spend decisions. They improve data quality and integration across fragmented channels, automate manual data work like media classification, and use incrementality testing to validate and calibrate model outputs. They also refresh MMM frequently, often quarterly, to keep pace with market change and shorten the path from insight to action.

TLDR IndexPage 8