Industries / Finance and business services / Advertising and marketing
Market research
Market research is the work of finding out what buyers think and do, by asking them in surveys and interviews and by counting what they buy, for the companies, public bodies and media that pay to know.
1911
The Curtis Publishing Company of Philadelphia set up a Division of Commercial Research in its advertising department and put Charles Coolidge Parlin in charge. The University of Pennsylvania, which holds the company's records, calls it "the first marketing research operation in the United States.".
33
Milestones achieved across 6 domains.
10
To watch for: announced, not yet confirmed outside.
12
Open deliverables. Nobody has shipped them.
In one minute
Where it started
Parlin's staff at Curtis interviewed shoppers, wholesalers and shopkeepers, and went through pantries and household trash to learn which brands people bought. By the mid-1930s the trade had its basic tools: the agree-or-disagree scale (1932), shop audits that counted what sold (Nielsen, 1933), and Jerzy Neyman's case for choosing samples at random (1934).
Where it stands
ESOMAR, the trade's world association, put the industry at about US$153 billion in 2024: US$62 billion for research software, US$56 billion for research services and US$35 billion for reporting. Most surveys now run online, where cheap opt-in panels had about twice the error of randomly recruited panels in a Pew Research Center test published in 2023. AI models now run interviews, code answers and stand in for respondents, and Pew's 2026 test of AI stand-ins found them off by 12 percentage points on average.
What is open
Most published claims about AI in the trade come from the firms that sell it. The open work is independent checking: AI question testers against real cognitive interviews, AI phone interviewers against human ones, synthetic "boosts" against held-back respondents, quotes in AI reports against their transcripts, and receipts and new products coded the way sales trackers code them. Each can be done with public data and models available today.
Domains
AchievedWatch forOpen
Domain
Progress
Last achieved
Watch for
Open
Surveys
Surveys ask many people the same written questions, in the same order, so that their answers can be counted.
5 · 3 · 2
Last achievedPhone surveys replace visits2020
Watch forChatbots that ask follow-ups
2
openPanels and sampling
Sampling is choosing whom to ask, and a panel is a standing pool of people who have agreed to answer surveys again and again.
7 · 2 · 2
Last achievedOpt-in samples, double the error2023
Watch forPeople simulated from interviews
2
openInterviews and focus groups
Qualitative research talks at length with a few people, alone or in groups, to learn why they think and act as they do.
4 · 2 · 2
Last achievedAI codes a focus group2026
Watch forAI interviews at scale
2
openData analysis
Data analysis turns raw answers into findings: coding written answers into categories, weighting samples, and modeling what drives choices.
7 · 1 · 2
Last achievedAI analysts meet real codebooks2026
Watch forCoding with stated confidence
2
openTracking sales
Sales tracking counts what people actually buy, from shop tills, shop audits and households that record their purchases.
5 · 1 · 2
Last achievedThe big trackers merge2023
Watch forSales through AI shopping agents
2
openReporting
Reporting is telling clients and the public what a study found and how it was done, so that others can judge how far to trust it.
5 · 1 · 2
Last achievedAI answers must be labeled2026
Watch forReports written by machine
2
open
Open in market research
3 of 12 open deliverables
Open · Surveys
Score AI pretests against interviews
A test that runs AI question checkers on survey questions that have already been through real cognitive interviews, and scores how many of the documented problems each checker finds and how many it invents.
Built with what exists
Q-Bank's question evaluation reports; the prompt designs described in the Essex paper; open-weight and commercial AI models through their APIs; Python for scoring.
Open · Panels and sampling
Test synthetic boosts on holdouts
An independent, pre-registered test of "boosting," in which a model adds synthetic respondents to a small subgroup of a survey. The test hides most real members of each subgroup, asks each method to fill the gap from the rest of the data, and compares the boosted estimates with the hidden respondents' real answers.
Built with what exists
Large public survey files such as the European Social Survey and Pew Research Center datasets; open synthetic-data tools such as the synthpop package for R and the Synthetic Data Vault for Python as baselines; commercial boosting services under trial licenses; Python for scoring.
Open · Interviews and focus groups
Check quotes in AI reports
A checker that takes an AI-written qualitative report and the transcripts it came from, finds each quoted passage, and records whether the quote is word for word, whether it is credited to the right speaker, and whether it supports the claim it is attached to. Then run it on reports from at least five AI research tools.
Built with what exists
The Anthropic Interviewer transcripts on Hugging Face; a fuzzy string-matching library such as RapidFuzz; an AI model as judge of whether a quote supports a claim, checked against people on a sample; the AI research tools under test.