How we measure our audience
Every figure we share with sponsors, partners, and government outreach buyers is measured first-party on our own infrastructure and defined on this page. This is the same data dictionary our own reporting is built on. No number leaves this site without a definition and a limitation label.
Collection principles
First-party and privacy-safe
All measurement runs on our own infrastructure. No third-party analytics scripts, no ad pixels, no data brokers, and no cross-site tracking of any kind.
No personal data, ever
We never store IP addresses, raw browser fingerprints, names, or emails in the analytics store, and we never collect any VA claim, health, or disability detail from visitors.
Human by construction
Bots do not execute the measurement beacon, so reportable traffic is human by design. Known crawlers, AI scrapers, uptime monitors, and vulnerability scanners are additionally filtered at the edge and counted separately.
Honest labels on every figure
Anything that is an estimate is labeled an estimate. We never present self-reported or modeled figures as verified fact, and we never inflate a number to make a sale.
A visitor who accepts analytics cookies receives a random first-party identifier that contains no personal information. A visitor who declines is counted with a daily-rotating salted pseudonym that cannot identify a person and resets every day. Returning-visitor status is only computed for consented visitors, and reports say so.
Metric definitions
The definitions below are the source of truth for our media kit and every campaign report.
| Metric | How it is measured | Limitation, always disclosed |
|---|---|---|
| Page views | Count of human pageview events measured by our first-party beacon. | Verified human only; bots excluded before counting. |
| Unique visitors | Distinct anonymous visitor tokens in the reporting window. | Visitors who decline cookies are counted with a daily-rotating pseudonym, which can slightly over-count uniques. Labeled accordingly. |
| Sessions | Visit sessions separated by 30 minutes of inactivity. | Derived client-side from the beacon. |
| Returning visitor share | Distinct visitors previously seen on an earlier day, divided by unique visitors. | Computed for cookie-consented visitors only. |
| U.S. traffic share | Distinct visitors whose request carried a U.S. country code, divided by unique visitors. | Country comes from a coarse edge header at request time. IP addresses are never stored. |
| Geography | Visitors grouped by country and U.S. state or region. | Country is reliable; region granularity depends on the network edge. |
| Devices | Mobile, tablet, or desktop class derived from the browser at request time. | Coarse class only. The raw user-agent string is not stored. |
| Traffic sources | Organic, direct, referral, social, email, or paid, classified from the referrer and UTM parameters. | A classification heuristic, consistent with industry practice. |
| Average engagement time | Mean active, visible time per page from a visibility-based timer. | A proxy for attention. Not identical to other vendors' definitions. |
| Top content | Page views grouped by page path. | No limitation. |
| Newsletter subscribers | Confirmed double opt-in subscribers with recorded consent. | Pending and unsubscribed addresses are reported separately, never blended in. |
| Audience composition | Self-reported role (veteran, family member, advocate, and so on) from an optional anonymous survey at newsletter signup. | A self-reported ESTIMATE with the response rate disclosed. Never presented as a verified figure. |
| Campaign clicks | Verified human clicks on a dedicated tracking link. | We report clicks we measured ourselves. We do not claim ad impressions or viewability. |
| CPC and CPA | Campaign cost divided by tracked clicks, and by attributed conversions. | Shown only when the actual campaign cost is entered. No modeled costs. |
Traffic quality
We report three plain totals: reportable human (verified human page views), excluded invalid (known crawlers, AI scrapers, uptime monitors, and vulnerability scanners, counted coarsely with no personal data), and raw (the sum of both). Sponsors only ever buy against the human number.
Legitimate search engine crawlers such as Googlebot and Bingbot are classified and excluded from human reporting but are never denied access to public pages.
What we never do
- Sell or share visitor data, or run cross-context behavioral ad tracking.
- Store IP addresses, names, emails, or any VA claim or health detail in analytics.
- Claim ad impressions, viewability, or any figure we did not measure ourselves.
- Let sponsors influence editorial content, rankings, or reviews.
- Present a self-reported or modeled estimate as a verified fact.
We are happy to walk through the methodology and share a current audience report.
The Veteran Benefit Desk is operated by Aurelis LLC. We are not affiliated with the U.S. Department of Veterans Affairs or any government agency.