---
title: "Failure Patterns"
description: "Common mistakes in voice DNA capture and how to avoid them. When something feels \"off\" about a profile, reference this document to diagnose the problem."
type: skill
canonical_url: https://claudary.paisolsolutions.com/skills/failure-patterns-1
source: "Claudary"
difficulty: intermediate
author: "Claude Code Knowledge Pack"
date: 2026-07-10T11:24:27.796Z
license: CC-BY-4.0
attribution: "Failure Patterns — Claudary (https://claudary.paisolsolutions.com/skills/failure-patterns-1)"
---

# Failure Patterns
Common mistakes in voice DNA capture and how to avoid them. When something feels "off" about a profile, reference this document to diagnose the problem.

## Overview

# Failure Patterns

Common mistakes in voice DNA capture and how to avoid them. When something feels
"off" about a profile, reference this document to diagnose the problem.

---

## The Goal

A successful Voice DNA Document:

- Captures what's **distinctively recognizable** about this writer
- Provides **actionable guidance** for the ghost writer
- Has **specific examples** demonstrating patterns
- Includes **anti-patterns** (what to avoid)
- Would make the writer say **"yes, that's me"**

When profiles fail, they usually fall into one of these patterns.

---

## 1. The Generic Profile

### Description

The profile is accurate but could apply to hundreds of writers. Nothing
distinctive is captured.

### Warning Signs

- Descriptions use vague language: "varied sentence lengths," "conversational
  tone," "clear and accessible"
- No specific examples from the writer's work
- Anti-patterns section is empty or generic
- Reader could not distinguish this writer from others based on the profile
- Ghost Writer Briefing has no actionable specifics

### Why It Fails

The ghost writer has nothing distinctive to work with. Output will be generic
because the input is generic. The 80% accuracy goal becomes impossible—you're
aiming at "average" instead of "this person."

### Example

**Generic (Bad):**

> "Uses a conversational tone with varied sentence lengths. Addresses the reader
> directly. Clear and accessible prose."

**Specific (Good):**

> "Short punchy sentences (5-10 words) followed by one longer unwinder (20-30
> words). Never uses 'however'—always 'but.' Addresses reader as 'you' at least
> once per paragraph. Ends paragraphs with concrete images, never abstractions."

### Intervention Questions

- "Can you show me a sentence that **only you** would write?"
- "What would a parody of your writing exaggerate?"
- "If I read something you wrote without your name on it, what would tip me
  off?"
- "What makes your writing **recognizably yours** versus just 'good writing'?"

---

## 2. The Contradictory Profile

### Description

The profile contains patterns that conflict with each other, making the ghost
writer's job impossible.

### Warning Signs

- Opposing patterns appear: "formal and distant" but also "warm and
  conversational"
- Different samples suggest different voices without explanation
- Ghost Writer Briefing has conflicting "do" and "don't" instructions
- User has validated contradictory elements

### Why It Fails

The ghost writer can't satisfy contradictory instructions. Output will be
inconsistent—sometimes formal, sometimes casual—in ways that feel wrong rather
than intentionally varied.

### Common Causes

- Mixing samples from different registers without realizing it
- Capturing aspirational voice alongside actual voice
- Evolution over time (old vs. new writing)
- Heavy editing by others in some samples

### Intervention Questions

- "These two patterns seem to conflict: [X] and [Y]. Help me understand when you
  do each."
- "Is this different across contexts? Blog vs. essay? Public vs. private?"
- "Some of these samples feel different to me. Were any heavily edited by
  someone else?"
- "Has your style changed over time? Which represents where you are now?"

### Resolution

Either:

1. **Separate registers:** Create different DNA documents for each mode
2. **Identify dominant pattern:** Capture the primary pattern with the variant
   as contextual
3. **Capture evolution:** Document the current preference while noting history

---

## 3. The Aspirational Mismatch

### Description

The profile captures how the writer wants to write, not how they actually write.
It's a wish list, not a fingerprint.

### Warning Signs

- Profile describes patterns not visible in samples
- User describes their voice very differently than samples demonstrate
- Heavy use of "I try to" or "I want to" language
- Named influences don't match actual samples
- Ghost Writer Briefing would produce something the user couldn't actually write

### Why It Fails

The ghost writer will produce text that sounds like someone else—the
aspirational target. The user will recognize it as "not me" even if it's "what I
wished I sounded like."

### Example

**User says:** "I write like Hemingway—short, punchy, no adjectives." **Samples
show:** Long flowing sentences with abundant adjectives.

**User says:** "I'm very formal and precise." **Samples show:** Casual,
conversational, loose with grammar.

### Intervention Questions

- "I'm noticing a gap between how you describe your voice and what I see in
  these samples. Which should I capture?"
- "Do you want the ghost writer to match how you currently write, or help you
  write more like your ideal?"
- "These samples don't quite match [description]. Is that typical, or were these
  unusual pieces?"

### Resolution Options

1. **Capture actual:** Document the real patterns; let user evolve naturally
2. **Capture aspirational:** Explicitly note this is a target, not current state
3. **Hybrid:** Actual patterns as foundation, aspirational elements as
   directional nudges

Always make clear which you're capturing.

---

## 4. The Shallow Profile

### Description

Only surface-level patterns are captured. The profile scratches the surface
without reaching what's truly distinctive.

### Warning Signs

- Only obvious patterns documented (sentence length, paragraph length)
- No signature elements or distinctive moves
- No exemplar passages
- Anti-patterns are generic ("avoids jargon")
- "What makes this writer recognizable?" has no clear answer

### Why It Fails

The ghost writer can mimic surface features but misses the deeper patterns that
make the writer recognizable. Output will feel hollow—technically similar but
spiritually wrong.

### Common Causes

- Rushed discovery session
- Not enough samples analyzed
- Skipped interview questions about distinctiveness
- Focused on what's easy to measure instead of what's distinctive

### Intervention Questions

- "We've captured some surface patterns, but I don't have a sense of what makes
  you **you**. Let's go deeper."
- "What do people who know your writing notice about it?"
- "If someone was imitating your writing, what would they get wrong?"
- "What's the thing about your writing you're most proud of? Most critical of?"

### Resolution

Go deeper on:

- Signature elements (what's uniquely theirs)
- Distinctive moves (things only they do)
- Anti-patterns (what they never do)
- The ineffable (what makes readers say "that's definitely them")

---

## 5. The Over-Specified Profile

### Description

Too many rigid rules, leaving no room for natural variation. The profile is a
straitjacket instead of a fingerprint.

### Warning Signs

- Every pattern is documented as a rule ("never," "always")
- No flexibility or context-dependence noted
- Many specific word bans and mandates
- Ghost Writer Briefing reads like a legal contract
- Output would be mechanical and stilted

### Why It Fails

Writing has natural variation. An over-specified profile forces the ghost writer
to produce stilted, mechanical output. Every sentence becomes a puzzle to
satisfy all rules rather than flowing naturally.

### Example

**Over-Specified (Bad):**

> "Sentences must be 8-15 words. Never start with 'The.' Always use
> contractions. Never use semicolons. Every paragraph must have 3-4 sentences.
> Never use words over 3 syllables."

**Appropriately Specified (Good):**

> "Tends toward shorter sentences (8-15 words typical) but occasionally uses
> longer ones for flow. Prefers contractions in casual contexts. Avoids
> semicolons. Paragraphs usually 3-4 sentences but can vary."

### Intervention Questions

- "Some of these feel like strict rules. Are they absolute, or more like
  tendencies?"
- "Is there context where you'd break this pattern?"
- "If I followed all these rules exactly, would it feel natural or forced?"

### Resolution

Convert rigid rules to:

- Tendencies with ranges
- Context-dependent patterns
- Primary patterns with noted exceptions

Add "When Uncertain" guidance that allows natural decision-making.

---

## 6. The Sample-Dependent Profile

### Description

Patterns are captured from samples but haven't been validated as generalizable.
The profile describes the samples, not the writer.

### Warning Signs

- Patterns based on single samples
- No interview confirmation
- User hasn't validated ("yes, that's me")
- Patterns might be context-specific, not general
- Mixed sample types without separation

### Why It Fails

The samples might be unrepresentative:

- Written for unusual contexts
- Heavily edited by others
- Written in a mode the user doesn't normally use
- Old writing that doesn't represent current voice

### Example

User provides an academic paper. Profile captures formal, passive-voice,
jargon-heavy patterns. But the user normally writes casual blog posts—the paper
was an anomaly.

### Intervention Questions

- "Is this sample typical of your writing, or was it unusual for some reason?"
- "The patterns I'm seeing here—are these consistent across your work?"
- "Was this piece edited heavily by someone else?"
- "Is this the kind of writing you want the ghost writer to produce?"

### Resolution

1. Validate patterns through interview before finalizing
2. Request additional samples to confirm consistency
3. Note confidence level for each pattern (confirmed vs. tentative)
4. Separate context-specific patterns from general ones

---

## 7. The Missing Anti-Patterns

### Description

Profile captures what the writer does but not what they avoid. Ghost writer has
no guardrails.

### Warning Signs

- Anti-Patterns section is empty or perfunctory
- "Never Does" list has only one or two items
- AI patterns to suppress aren't checked
- Ghost Writer Briefing has "Do" but not "Don't"

### Why It Fails

Knowing what to avoid is often more important than knowing what to do. Without
anti-patterns, the ghost writer might produce technically-matching text that
includes jarring elements the writer would never use.

### Example

Profile captures: "warm tone, conversational style, direct address." Missing:
"Never uses corporate jargon. Avoids sentimental language. Won't use 'amazing'
or 'incredible.'"

Ghost writer produces warm, conversational text but includes "let's synergize
our amazing journey together"—technically matching the positive patterns but
violating unspoken anti-patterns.

### Intervention Questions

- "What would you **never** write?"
- "What words make you cringe?"
- "If the ghost writer produced something that felt 'off,' what would it
  probably be doing wrong?"
- "What AI patterns have you noticed that bother you?"

### Resolution

1. Populate Anti-Patterns section thoroughly
2. Review `references/anti-ai-patterns.md` with user
3. Check applicable AI patterns to suppress
4. Add "Don't Do This" to Ghost Writer Briefing with equal weight to "Do This"

---

## 8. The Orphaned Profile

### Description

Profile was created but never validated or refined. It's a first draft treated
as final.

### Warning Signs

- No session history beyond initial discovery
- Readiness level says "Minimum Viable" but it's being used for full drafts
- No stress test completed
- No ghost writer feedback integrated
- Last updated date is old

### Why It Fails

Initial discovery is a foundation, not a finished product. Without validation
and refinement, the profile may contain errors, gaps, or patterns that don't
hold up in practice.

### Intervention Questions

- "When did we last update this profile?"
- "Have you used it with the ghost writer? What worked and what didn't?"
- "Are there patterns here that don't feel right anymore?"

### Resolution

1. Schedule refinement sessions after ghost writer use
2. Integrate feedback from ghost writer output
3. Update version and session history
4. Re-assess readiness level honestly

---

## Diagnostic Checklist

When a profile feels "off," run through this:

1. **Distinctiveness Test:** Could I distinguish this writer from others based
   on this profile?
2. **Contradiction Check:** Do any patterns conflict with each other?
3. **Reality Check:** Does the profile match the samples, not just aspirations?
4. **Depth Check:** Are signature elements and distinctive moves captured?
5. **Flexibility Check:** Are patterns captured as tendencies, not rigid rules?
6. **Validation Check:** Has the user confirmed these patterns are truly theirs?
7. **Anti-Pattern Check:** Are "don'ts" as thorough as "dos"?
8. **Currency Check:** Has the profile been updated based on use?

If any check fails, use the corresponding intervention questions to improve the
profile.

---

## Recovery Workflow

When you identify a failure pattern:

1. **Name the problem** clearly to the user
2. **Ask intervention questions** to gather missing information
3. **Update the profile** with what you learn
4. **Note the issue** in session history so it's not repeated
5. **Re-assess readiness** honestly

---

Source: [Claudary](https://claudary.paisolsolutions.com/skills/failure-patterns-1) · https://claudary.paisolsolutions.com
