Developing
ExpressVPN says its sample and method are insufficient to generalise; Google has not responded in the material reviewed.
In 60 seconds
- 102 queries (100 from Google Year in Search lists 2020-2025, plus "VPN" and "Antivirus") were run with Trends set to worldwide, varying only the VPN server location (US, Japan, India).
- Exact weekly match across all three countries over one year (52 weeks, April 2025 to April 2026): 39.9% overall, 47.0% for terms and 32.8% for topics; over five years, 66.9% for terms and 55.4% for topics.
- Within one point: 78.7% for terms and 64.8% for topics; Japan and India agreed on nearly all queries, while US pairings matched on under half of terms.
- Three devices on the same server returned identical terms scores in 81.73% of weekly cases.
Impact Layer
- What changed
- A study suggests Google Trends output may vary with the user's IP address, not only the query.
- Why it matters
- SEOs use Trends for demand and seasonality, so unexplained differences can mislead planning.
- Who is affected
- SEOs, content strategists and market researchers who use Google Trends.
- Martech stack area
- Analytics, SEO & Search
- Stack impact
- Check important Trends comparisons from more than one location or session, and use ranges and direction, not exact scores.
- Business impact
- Low A VPN vendor's research with undisclosed authors. A one-point difference counted as a mismatch, the report does not investigate causes or say which country is accurate, and ExpressVPN says it cannot generalise.
- Related vendors
- What to watch
- Google's response and independent replication.
ExpressVPN’s research team published “Consistency in Google Trends: How results can vary across IP addresses” on October 21, 2026, according to PPC Land. No authors are named.
Method
- 102 queries: 100 randomly drawn from Google Year in Search lists (2020-2025), plus “VPN” and “Antivirus”.
- Google Trends set to worldwide; only the VPN server location changed (US, Japan or India).
- Logged-out Chrome Incognito CSV exports; a match required identical weekly scores, so a one-point difference counted as a mismatch.
Key figures
- Over one year (52 weeks): exact match across all three countries 39.9% overall, 47.0% for terms and 32.8% for topics.
- Over five years (260 weeks): 66.9% for terms and 55.4% for topics.
- Within one point (one year): 78.7% for terms and 64.8% for topics.
- Japan and India agreed on nearly all queries; US pairings matched on under half of terms.
- Three devices on the same server returned identical terms scores in 81.73% of weekly cases.
Not stated
Why US results differ (the report does not investigate), which country’s score is accurate, exact export dates, Google’s response and whether the findings generalise.
Our read
ExpressVPN sells VPNs and recommends querying from several addresses. Use Trends for direction, and repeat a critical query before you act on it.
Sources
- PPC Land: Google Trends scores differ by IP in 60% of weekly checks, ExpressVPN finds (opens in a new tab)
Source for the method, findings, caveats and gaps.