Traffic classification
Every visit is sorted into one of five categories: Real, Crawler, Bot, Competitor or Click Farm. Crawlers are legitimate bots like Googlebot and are always whitelisted and never blocked. The dashboard also shows finer states such as Bounced, Over-threshold and Paranoia.
Classification answers a different question from the threat score. The score says *how suspicious* a visit is; the category says *what kind of traffic* it is. Both appear on every session in your analytics.
The five categories
| Category | What it means | Blocked? |
|---|---|---|
| Real | A legitimate human visitor | No |
| Crawler | A legitimate bot — Googlebot, AI and LLM crawlers, link-preview bots, uptime monitors, SEO tools, archive crawlers | Never |
| Bot | Confirmed malicious automation | Yes |
| Competitor | A competing business clicking your ads to drain budget | Yes |
| Click Farm | Organised, large-scale coordinated fraud | Yes |
Crawlers are never blocked
Blocking Googlebot would deindex your site. Crawler detection is deliberately broad — Google service bots, GPTBot, ClaudeBot, PerplexityBot, Meta-ExternalAgent, social link-preview bots, uptime monitors, SEO tools and archive crawlers are all recognised and permanently whitelisted, whatever their threat score looks like.
Finer dashboard states
- Bounced
- The visitor left almost immediately with no meaningful engagement — a common fraud signature, but also something real visitors sometimes do.
- Over-threshold
- The visit scored above your configured threat threshold and was blocked on that basis.
- Paranoia
- Blocked by Paranoia Mode rather than by score — a repeat paid-ad click inside your configured window. See Paranoia Mode.
Seeing the breakdown
The dashboard's threat breakdown shows counts per category for the selected period, and the crawler count is reported separately so filtered legitimate bots never inflate your fraud numbers. You can drill into any category from Analytics.
Frequently asked questions
Why is my crawler count so high?
That is normal and healthy. Search engines, AI assistants, monitoring services and SEO tools all crawl continuously. They are separated out precisely so they do not distort your fraud rate.
How is a competitor told apart from a real visitor?
By pattern rather than identity — repeat ad clicks with no conversion intent, engagement that does not match the campaign, and returns across sessions from a clustered person. AdProtektor cannot know which company someone works for.
Related articles
How detection works
Person-based fraud detection: fingerprinting, identity clustering across IPs and devices, and the 150+ behavioural signals behind every threat score.
Threat score and threshold
What the 0–100 threat score means, how the configurable threshold controls blocking, and when to move it.
Analytics
The analytics tabs — sessions, persons, blocked IPs and persons, countries, campaigns, keywords and devices.