Rerouted: Who Moves, Who Benefits, and Who Checks the Numbers

Written By Moe Pouyan

The Research Week · October 15 to 23, 2026

People are rerouting. They are moving routine purchases online, moving their news habits into the feed, and moving their expectations about the economy from "coping" toward "worse ahead". Organizations are rerouting too, toward apps, AI and automation. The research from the past two weeks suggests the people being rerouted are not always the ones who benefit.

This first issue covers new work from Ipsos, Gallup, Pew Research Center, NIQ and GfK, J.D. Power, Bain, Leger and Forrester, plus Research Live's reporting on the insights profession itself.

65%expect inflation to stay above normal until after next year, or never return to normalIpsos, 33 countries
15%of US adults order groceries online every week, up from 4% in 2019Gallup
6.3% vs 0.2%employment growth per point of AI exposure, men vs womenPew Research Center
51%of insights practitioners do not trust AI to run any research task without human reviewVoxpopme and MRII, via Research Live

Shift one: coping now, and the present is catching down

The ninth Ipsos Cost of Living Monitor, released October 22 across 33 countries, opens with what looks like resilience: 38% say they are comfortable or doing alright, against 27% finding things difficult. The outlook undercuts it. Seventy percent expect to pay more for energy and utilities in the next six months, 65% think inflation will not return to normal until after next year or ever, and 76% blame wars and their consequences. Among lower-income households the balance flips, with 38% struggling and 24% comfortable.

Newer trackers suggest the present is now catching down to those expectations. The US LSEG/Ipsos Primary Consumer Sentiment Index fell 5.3 points to 45.2 on October 19, its lowest since October 2011, with the Current sub-index down 7.6 points. In the UK, the GfK Consumer Confidence Index from NIQ dropped 11 points to -24 on October 23, its largest monthly fall since July 2016, with declines across income, age and gender groups. Bain's US Consumer Health Update found composite spending intent below neutral for the first time in nearly two years.

Spending is getting more selective rather than stopping. Gallup's Consumption Habits survey shows weekly restaurant dining down from 53% in 2019 to 39%, with the steepest drop among upper-income adults (65% to 48%). And Deloitte's holiday survey still expects a 4% rise in household holiday spending, with goods up and experiences down.

Why it matters. Satisfaction and wellbeing trackers that ask only about today will miss the turn. Expectation questions, and the gap between present and future ratings, are where the movement shows up first.

Shift two: digital routes up, human routes down

Gallup's survey also shows routines moving online. Weekly in-person grocery shopping has fallen to 74%, the lowest in its trend, while weekly online ordering has risen from 4% in 2019 to 15%. The shift is concentrated among women, whose weekly in-store shopping fell from 84% in 2022 to 69%.

Information is rerouting the same way. Pew finds 58% of US adults now call social media an important way to keep up with news, up from 43% in 2022, and 20% regularly get news on TikTok, up from 3% in 2020.

The cost of the shift shows up in service. The J.D. Power 2026 Canada Small Business Banking Satisfaction Study, released October 22, found mobile banking satisfaction up 43 points since 2022 to 719 on a 1,000-point scale. Over the same period, the share put on hold when calling a representative rose from 34% to 49%, and proactive advice slipped to 56% from 60% a year earlier. Customers with an assigned relationship manager saw satisfaction fall 17 points since 2022.

Why it matters. Better apps do not offset weaker human service; they cover different moments. Score digital and human channels separately, and follow customers into the channels they are moving toward.

Shift three: AI's benefits and AI's checks

A Pew Research Center analysis of government labour data compared 2017 to 2019 with January 2023 to June 2026. For each extra point of AI exposure, men's employment grew 6.3% and women's 0.2%. Among workers with a high school diploma or less, the split was +8.8% for men and -9.5% for women; among bachelor's degree holders, neither changed significantly. Pew notes this is not proof that AI caused the shift.

Consumers are drawing their own lines. Leger's AI Index finds 18% of Canadians now turn to AI first when searching online, triple the share when tracking began in 2024. Yet 70% prefer following a real person over an AI-generated creator, and 34% say they are less likely to buy from a brand using AI-generated content. Forrester's 2027 predictions expect half of brands to make human-generated content a top priority.

Researchers are applying the same caution to their own work. A State of Insights report from Voxpopme and MRII, reported by Research Live, found 51% of practitioners do not trust AI to run any research task without human review, and 28% discarded or re-fielded a study over data quality in the past year. On the say-do gap, 16% use no formal check and 33% did not know. The sample is small (126) and non-random.

Why it matters. AI's gains are landing unevenly, and audiences notice. Programs that report only aggregate AI adoption or satisfaction will miss who is falling behind and how much trust each AI touchpoint costs.

A note on method: name the measure

Gallup's grocery trend compares separate random samples, so it shows the population changing, not individuals switching. Pew's social media figure measures how important people say the feed is for news, not how often they use it. And a foundation of US survey weighting is under review: the Census Bureau has proposed rules that the Census Project says would remove race and ethnicity questions. Public comment closes November 2.

Also worth knowing

  • Two more synthetic-respondent products launched: Artificial Societies' Rothko, reported with no validation figures, and Highlight's Pulse, built on modeled consumer "Twins" (October 22).
  • NIQ found the average price paid for European tech and durables in Black Friday week 2025 was €147, against €131 across the year (October 22).
  • A Bentley University and Gallup survey found trust in businesses to act in society's best interest slipped 2 points to 41% (October 21).
  • An Ipsos KnowledgePanel poll found 49% of Americans have heard a great deal or quite a bit about AI (October 20).

Reading these numbers together

These studies differ in country, sample, mode and timing, from a 33-country online tracker to US telephone polls, sales data, analyst forecasts and a small practitioner poll. Their figures should not be compared directly. What they share is a direction, and each points to one action.

For researchers running trackers, add expectation questions and report the gap between present and future ratings. For CX teams, score human and digital channels separately and add the journeys customers are moving into, from grocery delivery to social feeds. For anyone reporting on AI, break results down by group and check AI outputs before they reach a decision. The averages look calm; the routes underneath them are changing fast.

Sources

Moe Pouyan
Next

From Question to Decision: How Market Research Works, With a Full Worked Example