What this site is

r2clickthrough is an independent AI marketing intelligence publication. No vendor sponsorships dictate coverage, no inflated benchmark claims pass without scrutiny, and no "10 AI tools you must try" copy gets published without first-hand testing.

The remit is narrow on purpose: AI tools and workflows for working marketers and developers. That means honest reviews of AI agents (Manus, Claude, ChatGPT Operator), code assistants (Cursor, Copilot, Codex variants), CRO tooling (Optimeleon, VWO comparisons), and the open-source tutorials that bridge the gap between "this tool exists" and "here is how to actually run it tomorrow morning."

How coverage gets chosen

Three filters:

  • Useful to a working marketer or developer this quarter. Speculative AGI think-pieces live elsewhere.
  • Verifiable. If the post claims a 65% token reduction or an 86.5% GAIA score, the source is named and linked.
  • Genuinely tested where possible. Reviews that include a smoke test disclose the methodology, the credit burn, and what broke.

Editorial methodology

Every benchmark stat carries a named source. Every tool review either includes a first-hand test or is labeled clearly as a curated overview. Posts updated in the last 90 days carry a dateModified field; older posts get a freshness pass on a rolling basis. The full source list and editorial standards are auditable on request.

Disclosure policy

Some outbound product links may be affiliate links. They never change the coverage decision and they are flagged inline on any post where they appear. r2clickthrough has no paid placements, no sponsored reviews, and no off-the-record arrangements with the tool vendors covered here.

Where a tool is reviewed and the reviewer received free access or extended trial credits, that disclosure appears in the post body, not buried in a footer.

Who writes here

Posts are bylined Harry Richter, Marketing Technology Analyst. Coverage focuses on tools tested in real marketing operations workflows: paid media auditing, content production pipelines, CRO experimentation, and the developer-tool overlap that increasingly lands on marketing teams.

Corrections and contact

Errors are fixed in-line with a visible correction note and an updated dateModified timestamp. Vendor responses to negative reviews are published as updates rather than rewrites. To flag a factual error, request a correction, or pitch a tool for review, email the address on the Tags page.

What is explicitly not on the menu

  • Listicles assembled from press releases without tool access
  • "AI is going to kill X profession" trend-bait
  • Generic prompt-engineering posts that recycle the same 12 examples
  • Sponsored content disguised as editorial

Read a review, run the tool, tell us if the conclusion held up. That is the loop.