Dwelling as Conversion

The Problem: Conversion as Interruption

The data told a clear story. Of 174 visitors to the agency homepage, only 21 clicked through to the contact page (5.6%). Of those, only 3 submitted the form. Meanwhile, experiments on .space and .io held attention—users explored, scrolled, engaged.

The pattern was revealing: users who arrived at .agency from experiments didn't convert—they navigated to .ltd seeking philosophical grounding. They wanted to understand before they bought. The sales page interrupted their dwelling.

Heideggerian Diagnosis: Vorhandenheit

In Heidegger's terms, the agency homepage was present-at-hand (Vorhandenheit)—it announced itself as a sales mechanism rather than receding into transparent use. The visitor became aware of being sold to. The tool became conspicuous.

""

The experiments on .space embody the opposite: ready-to-hand (Zuhandenheit). Users explore the Threshold Dwelling visualization, test the Motion Ontology tools, run the Triad Audit—and in doing so, they're already experiencing what CREATE SOMETHING builds. The tool recedes; insight emerges.

The agency homepage broke this pattern. Instead of continuing the dwelling experience, it demanded a mode switch: stop exploring, start deciding. The result? Users bounced to .ltd for more exploration rather than converting.

Phase 1: Assessment as Active Dwelling

Our first insight: the fix wasn't better copy—it was recognizing that conversion is dwelling continued. Instead of asking users to switch from exploration to transaction, we built an interactive assessment that applies the Subtractive Triad to the visitor's own business. The user reflects; insight emerges from their answers.

The three questions mirror the three levels of the Subtractive Triad. Q1 identifies what is accumulating (implementation, artifact, or system-level). Q2 applies Rams' test: "What would you remove if there were no consequences?" Q3 connects back to the whole: "What's stopping you?"

Implementation: The Hermeneutic Circle in Code

The assessment is itself an application of the Subtractive Triad. We adapted patterns from the experiment runtime components on .space, applying DRY at the implementation level.

// Q1 options map to triad levels
const accumulatingOptions = [
  { id: 'tech_debt', label: 'Technical debt', triad: 'implementation' },
  { id: 'process_sprawl', label: 'Process sprawl', triad: 'artifact' },
  { id: 'feature_creep', label: 'Feature creep', triad: 'artifact' },
  { id: 'decision_paralysis', label: 'Decision paralysis', triad: 'system' },
  { id: 'disconnected_systems', label: 'Disconnected systems', triad: 'system' },
  { id: 'manual_tasks', label: 'Manual repetitive tasks', triad: 'implementation' }
];

// Q3 options map to services
const blockerOptions = [
  { id: 'uncertainty', label: "Uncertain about impact", service: 'advisory' },
  { id: 'resistance', label: 'Organizational resistance', service: 'transformation' },
  { id: 'where_to_start', label: "Don't know where to start", service: 'automation' },
  { id: 'bandwidth', label: 'No internal bandwidth', service: 'partnership' },
  { id: 'complexity', label: 'Technical complexity', service: 'agentic-systems' },
  { id: 'web_first', label: 'Need web presence first', service: 'web-development' }
];

Each question's options encode domain knowledge. The recommendation isn't random—it emerges from the intersection of what's accumulating (the problem domain) and what's blocking removal (the intervention type).

// When user completes assessment, they're directed to contact
// with assessment context preserved
const contactUrl = `/contact?service=${result.recommendation.service}&assessment=${sessionId}`;

// Contact form shows:
// "Based on your assessment, we recommend {service}"

// On contact submission, assessment is marked as converted
await db.prepare(
  `UPDATE assessment_responses SET converted_to_contact = 1 WHERE session_id = ?`
).bind(assessmentId).run();

Phase 1 Insight: Dwelling Through Questions

The crucial insight behind the assessment was that asking questions is dwelling. When a user answers "What are you accumulating?" they're not being interrogated—they're being invited to reflect. The assessment becomes a mirror, and in that mirror, the user sees their own situation more clearly.

""

This is why the assessment works where static copy failed. Static copy tells the user what they should want. The assessment asks what they're experiencing. The difference is the difference between Vorhandenheit (object to be convinced) and Zuhandenheit (tool for self-understanding).

The assessment components (AssessmentRuntime, AssessmentStep, InsightReveal) remain in the codebase for potential future use—particularly for visitors who enter via specific service pages and want a deeper engagement. But for the homepage, we discovered something even more aligned with our philosophy.

Phase 1 Analytics

The assessment implementation tracked timing and conversion without interrupting the experience:

CREATE TABLE assessment_responses (
  session_id TEXT NOT NULL,
  accumulating JSON,        -- Q1 selections
  removal_insight TEXT,     -- Q2 free text
  blockers JSON,            -- Q3 selections
  recommended_service TEXT,
  triad_level TEXT,
  time_on_q1_ms INTEGER,
  time_on_q2_ms INTEGER,
  time_on_q3_ms INTEGER,
  total_time_ms INTEGER,
  converted_to_contact INTEGER DEFAULT 0
);

This allowed us to measure not just if users converted, but how they dwelt. Which questions took longest? Which answers correlated with conversion? The assessment generated data for refinement.

Implementation: Words as Data

The key insight was treating words as data with a keep property:

const words = [
  { text: 'We', keep: true },
  { text: 'help', keep: false },
  { text: 'businesses', keep: false },
  { text: 'identify', keep: false },
  // ...corporate fluff marked keep: false...
  { text: 'remove', keep: true },
  { text: 'what', keep: true },
  { text: 'obscures.', keep: true }
];

// Scroll progress drives phase transitions
let phase = $derived(
  scrollProgress < 0.1 ? 'reading' :
  scrollProgress < 0.35 ? 'striking' :
  scrollProgress < 0.5 ? 'fading' :
  scrollProgress < 0.85 ? 'coalescing' : 'complete'
);

Each word calculates its own strike and fade progress based on scroll position. Words marked keep: false get a strikethrough that grows from left to right, then fade out entirely. The kept words remain, growing bolder as the message distills.

{#each words as word, i}
  {@const wordStrike = (strikeProgress - i * 0.03) / 0.3}
  {@const wordFade = word.keep ? 1 : 1 - fadeProgress}
  <span
    class="word"
    class:keep={word.keep}
    class:hidden={!word.keep && phase === 'complete'}
    style="--strike: {wordStrike}; --fade: {wordFade};"
  >
    {word.text}
    {#if !word.keep}
      <span class="strike"></span>  // CSS width: calc(var(--strike) * 100%)
    {/if}
  </span>
{/each}

From Asking to Showing

The evolution from assessment to TextRevelation represents a deeper understanding of the principle:

""

The assessment asked about subtraction. The TextRevelation demonstrates it. The medium becomes the message—not as decoration but as meaning. Understanding comes through observation, not interrogation.

The Meta-Application

There's a recursive elegance here. The Subtractive Triad was developed to guide our work. Now it's being used as the content of a conversion tool. The framework that helps us build things is the same framework that helps prospective clients understand what they need.

This is what it means for the hermeneutic circle to complete: the philosophy (.ltd) informs the research (.io), which validates the practice (.space), which applies to service delivery (.agency), which tests and evolves the philosophy.