GWI's headline is "Trust every decision", and its own September research found 81% of marketers say AI makes it easier for outdated information to influence decisions. Yet gwi.com gives five different sizes for its research: 2M+ interviews a year on the homepage and /data, 1.4M+ annual surveys on the Agent Spark page, over 5 million interviews on the Agentic Portfolio page, "almost a million internet users" further down /data, a 940K+ Core sample, and 35BN or 45BN data points representing 2.8, 3 or 3.5 billion people. AI assistants, journalists and procurement teams copy whichever figure they find first.
First action
List every number on the homepage, /data, /data-coverage, Agent Spark, Agentic Portfolio and About pages in one spreadsheet.
How to measure it
Key pages using the single fact source
- Now
- –
- Target
- 100%
- By
- 30 days
Steps
Collect
Every size, market and data-point claim across the site, blog boilerplate and press kit.
Content Marketing Manager 1 day
Define
One definition per number: interviews, respondents, data points, people represented, markets.
Chief Insight Officer 3 days
Publish
A dated /data/facts page with the current wave and method, plus Dataset markup.
Content Marketing Manager 3 days
Centralize
One HubSpot module for stat blocks so every page reads the same values.
Director of Marketing Operations 1 week
Fix copy
Agent Spark, Agentic Portfolio, /data, About, education hub and the MRI Simmons post.
Content Marketing Manager 3 days
Update boilerplate
Press kit, sales decks and job posts use the same figures.
Head of Comms 2 days
Check AI answers
Ask ChatGPT, Claude and Copilot how big GWI's survey is, monthly.
Director of Marketing Operations monthly
Refresh
Update the sheet with each quarterly wave.
Chief Insight Officer quarterly
Why it works
Write one dated fact sheet that defines each number (interviews vs. respondents vs. data points, Core vs. all data sets, people represented) with the current wave, publish it at /data/facts with Dataset markup, and point every page's stat block to a single source that updates once a quarter. The new Director of Marketing Operations owns the single source and the quarterly refresh; the Senior Data Journalist uses only those numbers in pitches.
A company that sells verified, sourced insight is judged on its own consistency first. One set of numbers removes a question from every enterprise security and procurement review, makes AI answers about GWI correct, and gives PR a clean boilerplate.
Ready to use
ChecklistGWI fact sheet: the numbers to agree
[ ] Interviews per year (all studies) and Core annual sample [ ] Markets: Core vs. all data sets (50+ or 53) [ ] Data points per year (35BN or 45BN) [ ] People represented (2.8B, 3B+ or 3.5B) [ ] Waves per year and date of the latest wave [ ] Years of trend data [ ] One-line method with a link [ ] Owner and next refresh date
Risks and how to handle them
Different numbers are each correct in context
Keep them, but label each one (Core sample vs. all interviews) on the facts page.
Pages drift again
Stat blocks read from one module, so a change is made once.
Evidence
- Your websiteThe size of GWI's research differs from page to page: the Agent Spark page says "direct from 1.4M+ annual surveys" and "53 markets"; the Agentic Portfolio page says "over 5 million interviews a year", "45BN total data points annually" and data "representing 3.5 billion people"; the /data page says "2M+ interviews annually" and, a few lines down, "almost a million internet users" a year; the Core data set shows "940K+ Annual sample" and "3B+ Internet users represented"; the About page says "over three billion people"; the GWI vs MRI Simmons post says "over 2.8 billion consumers".Source
- Your websiteHomepage (checked October 21, 2026): H1 "Know your audience. Trust every decision."; stats "50+ global markets", "2M+ interviews annually", "4 waves of data per year", "35BN total data points annually"; Agent Spark answers "pulled from 35 billion data points"; a "Sign up for free" button that leaves gwi.com for signin.globalwebindex.com.Source
- Your press releaseGWI research published September 30, 2026 (450+ marketers in the UK, US and Singapore): 72% say their organization made a significant decision on AI insight that later proved wrong; 37% report financial losses; only 37% verify AI insights "most of the time"; 81% agree AI makes it easier for outdated information to influence decisions and 76% have used data they later realized was out of date. Jason Mander: "Verification means knowing where an insight comes from, how the underlying data was collected and whether it's still up to date."Source
- HiringDirector of Marketing Operations (London, posted October 15, 2026), reporting to the VP Growth Marketing: "Marketing Ops at GWI is a growth engine, not a service function." Owns the marketing tech stack, "lead and account scoring and ensuring actionable intent signals are surfaced to Sales", data quality between Salesforce and HubSpot, and "campaign infrastructure and attribution reporting", "ensuring we have a clear source of truth for performance"; implements "automations that measurably reduce manual work or improve conversion". Stack named: HubSpot, Salesforce, Clay, Claude, Intercom, Qualified, GA, paid media and "SEO/GEO tooling".Source
- HiringSenior Data Journalist (PR & Comms, London, posted October 15, 2026): "Support proactive pitches and media enquiries with data points"; "support ABM and demand gen efforts for high-value prospects"; "Build bespoke data stories for high-value accounts, enabling Marketing and Revenue teams to deliver personalized campaigns"; outputs include white papers, reports, podcasts and op-eds; supports GWI's "positioning at the forefront of AI-powered insights in an increasingly agent-driven world".Source