Coffee & Commerce #59: Holiday Shopping Trends, Target Layoffs & ChatGPT’s Conversion Rates

Good afternoon,

Here’s the headlines for this week: 31 October 2025

ECOMMERCE

Over 56 % of online holiday-week sales are projected to come via mobile devices, and BNPL is expected to account for about $20.2 billion in purchases—up ~$2 billion from 2024.
Some major retailers have already scaled up AI for personalization and automation, and analysts estimate AI-driven e-commerce could reach ~$263 billion globally this season.
Online holiday sales are projected to grow ~9 % to reach ~$305-310 billion, while overall holiday-season retail growth is estimated at just ~2.9 %-3.4 %, signalling that online is out-pacing the broader market.

Atlas introduces “agent mode,” allowing users to automate online tasks like shopping or booking, shifting digital visibility from search engines like Google to AI-driven browsers.
Marketers are urged to adapt websites and product data for AI readability—using protocols like ChatGPT’s “agentic commerce” and ensuring sites are navigable by agents that can interpret JavaScript and complete tasks.
Retailers such as Instacart, Walmart, and Target are integrating with ChatGPT for direct AI-powered shopping, while brands must make social content authoritative and up-to-date since LLMs now index it to influence purchase decisions.

Traffic referred from ChatGPT to e‑commerce sites is high, but the conversion rate to sales is very low compared to traditional channels (like search, email, affiliates).
The study suggests that while ChatGPT drives curiosity and clicks, many users hesitate to complete purchases, pointing to a lack of trust or readiness to buy via that route.
For marketers: generating AI‑driven referral traffic may not translate into revenue unless you combine it with strong trust‑building, clear conversion paths, and tailored attribution models to track its real contribution.

TECH

Meta Platforms is rolling out a paid subscription in the UK allowing users of Facebook and Instagram to opt-out of personalised advertising.
The cost is £2.99 per month if subscribed via web; a higher fee applies if done via app stores to account for store-commission fees.
Users who don’t subscribe will continue under the ad-supported model, meaning their data will continue to fuel targeted advertising unless they opt for the paid option.

RETAILERS

Retailers enter a generative AI feedback loop

Retailers are responding to the rapid rise of generative AI tools (chatbots, recommendation engines) by altering product pages, content and imagery so they show up better when consumers ask AI chatbots for shopping advice.
Instead of keyword-searching (“black coffee mug”), shoppers are increasingly using more descriptive, occasion- or feeling-based queries via AI bots. This means e-commerce sites need richer descriptive content and smarter relevance to match those new query types.
The use of generative AI also gives retailers opportunities to create digital personas or synthetic market research via AI, but also brings new complexity: models and platforms may carry bias, and the decision-funnel for shoppers is being disrupted by stimulus-based rather than memory-based pathways.

Retailers are rolling back generous “free return” policies. What was once a competitive advantage is becoming harder to sustain given rising shipping, logistics, and tariff costs.
According to the National Retail Federation (NRF), the U.S. returns rate is expected to be about 15.8% of sales in 2025 (≈ $849.9 billion), slightly down from 16.9% in 2024.
Key factors driving policy changes: high reverse-logistics costs (especially in e-commerce), misuse or fraud of lenient return practices (notably among younger consumers), tariff burdens, and attempts to shrink excess return volumes.

Amazon plans to replace over 500,000 U.S. jobs with robots, aiming to automate 75% of its operations by 2027, while continuing to expand its product sales.
Advanced robotic fulfillment centers, like the Shreveport facility, serve as templates; they drastically reduce human labor needs while increasing efficiency, with new retrofits expected to cut thousands of jobs in existing warehouses.
The shift could reduce blue-collar jobs, particularly affecting Black workers, while Amazon promotes technical roles and community engagement to manage public perception and create “future” jobs in robotics and engineering.

Many Amazon delivery contractors are quitting or struggling as expenses for insurance, vehicle maintenance and accident-related costs soar, eroding earnings even as delivery volumes remain high.
The company raised payments per package by 20% (to 12 cents), but contractors say the increase is too little, too late, and some view it mainly as a holiday retention tactic.
Contract terms, performance metrics and limited resale options make it hard for delivery firms to exit profitably, leaving some to diversify into other businesses or operate at shrinking margins.

Amazon is rolling out a feature called “Help Me Decide” that presents users with a single recommended product once they’ve browsed multiple similar items — along with why that product is the best choice.
Alongside the main recommendation, Amazon shows alternative options (e.g., a cheaper variant or a higher-end version) to give users decision-flexibility rather than one rigid suggestion.
The tool is intended to streamline the purchase decision process by reducing browsing fatigue and simplifying choice — aiming to push shoppers toward conversion by leveraging generative AI/behavioural data.

Target is cutting about 1,800 corporate roles — including eliminating 800 open positions — representing roughly 8% of its global corporate staff, as part of a broader reorganization.
The moves are tied to weak and declining sales, with comparable sales down in 9 of the last 11 quarters, and a leadership push (by incoming CEO Michael Fiddelke) to simplify operations, reduce overlap, and speed decision‑making.
The layoffs do not affect store or supply chain workers; affected corporate employees will receive pay and benefits through January 3 and severance packages.