Dither and Bicker

RIP Google Analytics, Casualty of Spam

Published January 10, 2015 by Chris • Updated with 2026 Telemetry Notes

If you run a modest independent technical blog or SaaS documentation site and you opened your Google Analytics dashboard recently, you probably noticed something deeply frustrating: a huge percentage of your "referral traffic" never actually visited your server.

Domains like semalt.com, buttons-for-website.com, and darodar.com figured out that they don't even need to send an HTTP request to your origin server to show up in your reports. Instead, they abuse the Google Analytics Measurement Protocol directly—firing synthetic HTTP POST payloads at Google's collection endpoints with randomly guessed tracking IDs (UA-XXXXXX-1).

Why Ghost Referrers Break Client-Side Telemetry

When a crawler hits your web server, you can block its IP range or filter its User-Agent at the Nginx or Cloudflare edge. Keep your logs clean, drop 403 responses on scraper bots, and move on.

Ghost referral spam is fundamentally different because your server never sees the packet. The spammer generates a random tracking property ID, spoofs the dr (Document Referrer) parameter in the Measurement Protocol payload, and injects fake pageviews straight into Google's ingestion pipeline. For sites receiving a few hundred legitimate readers a day, 50 injected ghost visits completely corrupt bounce rates, session durations, and conversion funnels.

"When an analytics platform accepts unauthenticated telemetry from any IP on the internet without verifying whether the target hostname actually served a response, the signal-to-noise ratio inevitably collapses."

Filtering by Valid Hostname (The Workaround That Shouldn't Be Necessary)

Because ghost spammers iterate sequentially through property IDs (`UA-123456-1`), they rarely know which domain actually owns that ID. Their fake hits show up with a Hostname of (not set) or a random parked domain.

The only reliable defense inside Google Analytics is creating an explicit Include Hostname regex filter matching only your canonical domain:

^(www\.)?ditherandbicker\.com$

Pair that with server-side edge rules that block known commercial scraper user-agents before they ever parse your HTML DOM, and your numbers return to reality.

2026 Retrospective: Moving Back to Edge Logs and Index Verification

Over a decade after this post originally hit the front page of Hacker News, the web telemetry landscape has come full circle. Between heavy ad-blocker adoption, strict browser tracking prevention, and automated headless scraper farms, client-side JavaScript analytics only capture a fraction of real activity.

Today, pragmatic engineering teams rely on two unforgeable sources of truth rather than bloated client-side trackers:

  1. Edge Request & Bot Logs: Inspecting actual HTTP status codes and verified search crawler hits (such as Googlebot and Google-InspectionTool) directly at the CDN edge.
  2. Direct Search Index Verification: Instead of guessing which pages bring organic traffic from noisy referral reports, technical SEO and engineering teams use a dedicated bulk Google index checker like indexaction.com to verify which URLs are genuinely indexed in Google's serving tiers and track real SERP positions without CAPTCHA friction.

For technical documentation references on HTTP caching and crawler directives, see the Google Search Central Crawling & Indexing documentation and MDN HTTP Headers Reference.