TE
Travis Eric
HomeNow
Work with meBook a 30-minute call
Travis Eric LogoTravis Eric

AI consulting and custom software, based in Fort Collins. Explore the work. Bring your next idea.

Travis Eric AI Consulting(720) 770-2979

Navigate

  • Home
  • Now
  • Projects
  • How I work

Deeper

  • Work with me
  • About
  • Share your experience
  • Case Studies
  • Resources
  • Impossible Vision
  • Technical Stack
  • Manifesto
  • Writings

Follow the work

New systems and field notes, sent when they ship.

Communities

AI Builder's LabView membershipTeneo Publishing CollectiveView details
@TravisEric_Travis Eric@TravisericLive Builds@travis_eric
PrivacyTerms© 2026 Travis Eric
/
/
Back to writings
Skills and playbooks

Copy Arena — draft, read, refine, launch

Travis EricOctober 10, 20263 min read
Share this post:

When I set out to improve the copy on my website, I turned to a practical system called Copy Arena. This system eliminates the guesswork from selecting effective headlines by testing them with real readers. The process begins with drafting several headline candidates and ends with launching the most successful one. The idea was inspired by a founder's directive logged on September 12, 2026, in ledger F82: "we run a bunch of copy ideas and refine them until we find the best one."

The Problem It Solves

Before implementing Copy Arena, I faced a recurring issue: multiple headline changes within a few days, none of which were validated by readers. This was evidenced by a failed headline test where none of three blind company readers recognized the live headline (Traviseric.com/design-review/home-v2-2026-09-09/headline-ledger.html). A headline must resonate with readers within five seconds and prompt them to take action.

Output of Copy Arena

The system produces a ranked list of headline candidates, each evaluated by three personas over a five-second read. Each candidate receives scores on specific criteria such as for_me, promise_understood, and action_named, ranging from 0 to 5. It also captures the exact words that either disengaged or captured the reader's attention. The results, including a recommended winner, are saved to design-review/<topic>/copy-arena-<date>.html, ensuring I have a documented history of headline performance.

Implementation Steps

1. Define the Reader

Before drafting any headlines, it's crucial to establish the reader personas. Each site has one primary reader, and I create three personas based on this reader using files like JUDGE-PROMPTS.md and READER-PERSONAS.md. These personas are detailed with specifics such as their business type and what they value most. For my site, traviseric.com, the reader is defined by ledger F80 as a company with its own systems.

2. Draft the Candidates

I draft between six to twelve headline candidates, each comprising a full first screen: eyebrow, headline, lede, primary action, and a secondary action. The drafts vary in approach, focusing on the reader's situation, the mechanism offered, fears alleviated, and proof points. Importantly, each candidate adheres to the claim lock and honesty doctrine, avoiding exaggerated claims. These drafts are stored in candidates.json.

3. Render for Reader Experience

Using the command:

node client-factory/scripts/copy-arena/shoot-candidates.mjs --url <preview-or-local-url> --candidates candidates.json --out <dir>

I render each candidate as it would appear to a reader. This involves swapping the hero text on the real page and saving images of the page's appearance in both phone and desktop formats. This step ensures the five-second test is based on visual impact rather than mere text.

4. Conduct Blind Reads

For each candidate and persona combination, I employ an agent with the Pass A prompt from JUDGE-PROMPTS.md to evaluate the image. Judges work independently, ensuring unbiased feedback. The output is collected as a JSON block for analysis.

By following this methodical approach, I've transformed my headline selection process from guesswork to data-driven decision-making. Copy Arena provides a structured way to test and refine marketing copy, ensuring that the final choice resonates with real readers.

Join Pattern Insights

Get weekly insights on AI architecture, pattern recognition, and building platforms without permission.

On this page

  • The Problem It Solves
  • Output of Copy Arena
  • Implementation Steps
  • 1. Define the Reader
  • 2. Draft the Candidates
  • 3. Render for Reader Experience
  • 4. Conduct Blind Reads
Tags:
build-in-public

Read next

October 9, 2026

signup-doors-walk: every door, including the moment after submit

The "signup-doors-walk" skill is an essential runbook for FortListOS, designed to walk through every signup door as a customer would experience it. This...

Read it

October 8, 2026

goal-finder — derive the goals a repo needs to finish

The goal-finder tool is designed to reconcile a repository's documented intentions with its current state to derive prioritized, verifiable goals. It's...

Read it

October 7, 2026

Motion Studio: the video harness

The overnight build system for Motion Studio reveals a clear, efficient process for producing video content as code. At the heart of this process is the...

Read it

Join the Discussion

Have thoughts on this post? I'd love to hear them! Join the conversation on X where we can discuss AI architecture, pattern recognition, and building platforms.

Discuss on X

Or reach out directly at @TravisEric_