Skip to main content
Methodology

Traffic Quality Score Methodology

The Traffic Quality Score is a 0–100 score that helps you quickly assess how trustworthy the visits and clicks reaching your website are, based on a combination of signals rather than a single rule.

TL;DR

The Traffic Quality Score is a 0–100 score ClickSentinel calculates for every click, used to estimate how trustworthy that traffic is. A lower score means the click carries more suspicious signals — for example coming from a VPN, datacenter, Tor, being an abnormal duplicate, or occurring at an unusual velocity. The score is a prioritization tool, not a final verdict.

Score bands

80–100
High quality
Few or no suspicious signals; behavior largely resembles real users coming through normal channels (organic, direct, social).
60–79
Medium quality
A few signals worth watching (e.g. an IP from an uncommon hosting ASN) but not enough to be treated as high risk; worth monitoring over time.
40–59
Suspicious
Several risk signals appear together (e.g. VPN combined with abnormal velocity); manual review is recommended before taking action.
0–39
High risk
A concentration of negative signals (datacenter/Tor, repeated duplicates, an IP with a bad reputation across sites); more likely to be invalid or low-quality traffic.

Signals used

  • IP intelligence: detects VPN, proxy, datacenter (AWS/GCP/Azure/Cloudflare, etc.) and Tor exit nodes based on ASN/hosting-provider data.
  • Duplicate click detection: repeated clicks from the same IP within a short time window (default under 5 minutes).
  • Velocity pattern: an abnormal click frequency from the same IP within one hour (default above 10 clicks/hour).
  • Cross-site IP reputation: an IP with a history of bad behavior on other websites also using ClickSentinel is factored in.
  • UTM/referrer context: cross-checking the declared traffic source to spot anomalies.
  • Whitelist logic: IPs an admin has already whitelisted (e.g. internal or partner IPs) are not penalized by the signals above.

What this score is NOT

  • Not legal proof that a click is ad fraud.
  • Not a replacement for Google Ads' official Invalid Click Report — for disputes or refunds, use Google Ads' own tools as the basis.
  • No guarantee of an ad refund from any platform.
  • Not an absolute "human verification" mechanism (it is not a CAPTCHA or device biometrics).
  • The score reflects the signals available at the moment of the click; the same IP can change ownership or behavior over time.

Illustrative scoring scenario

Illustrative example — not a guaranteed result

Illustrative example (not real data): a click arrives from an IP in a known datacenter range, repeats 12 times within 20 minutes on the same button, and that IP has previously been flagged on 2 other websites. With these signals combined, the system may deduct heavily across three categories at once — datacenter, duplicate and velocity — pushing the overall score into the high-risk band (0–39).

By contrast, a click from a residential ISP IP that appears only once that day and matches none of the negative signals is typically placed in the high-quality band (80–100).

How to use this score

  • Track the average score trend by day/week rather than reacting to a single click.
  • Filter and prioritize low-score clicks for manual review before taking action.
  • Consider exporting high-risk IPs to Google Ads (IP exclusion) if that fits your advertising policy.
  • Combine this data with Google Search Console, GA4 or Microsoft Clarity for a fuller picture before deciding.

Frequently asked questions

What is a good Traffic Quality Score?
Generally a score of 80 or above is considered high quality. What counts as "good" also depends on your industry and traffic sources — compare the average score over time rather than applying one fixed threshold to every case.
Why do datacenter clicks get penalized?
Real users typically connect through residential ISPs (fixed-line or mobile carriers), not from data centers. IPs in datacenter ranges (AWS, GCP, Azure, etc.) are more often associated with servers or automated tools than a real user's browser, so they are treated as a risk signal — though that does not always mean fraud.
Should I block all VPN traffic?
No, automatically blocking all VPN traffic is not recommended. Many real users have legitimate privacy or security reasons for using a VPN. ClickSentinel recommends using the VPN signal for monitoring and combining it with other signals rather than a hard block.
Does the score change over time?
Yes. The score is calculated from IP intelligence and reputation data available at the time of the click; the same IP can receive a different score at different times if its behavior or reputation changes.
How is the Traffic Quality Score different from GA4's bot filtering?
GA4 primarily filters bots based on a known bot user-agent list (IAB/ABC) at the session level. ClickSentinel scores at the click level, adding IP intelligence (VPN/datacenter/Tor), duplicate/velocity patterns and cross-site reputation — an additional signal layer, not a replacement for GA4.
Does a low score guarantee the click is fraudulent?
No. A low score is a risk signal that warrants further review, not a conclusion. Some legitimate clicks can still fall into a low-score band, for example an employee using a corporate VPN.

Check your traffic quality for free

Want to see your own website's traffic quality score? Get started for free — no credit card required.