Shadow Infrastructure Map

The surface we map

Shadow infrastructure is external to you. It's the books, frameworks, certifications, conferences, communities, tools, and people your audience already trusts, that exist independently of your company and that you don't own. We deliberately exclude your own programs, frameworks, and internal communities, even when your content names them: the value is the ecosystem you can enter, not the one you've already built. If your public content is largely self-referential, your map will honestly be small, and we tell you why rather than pad it with your own properties.

No analytics tool reports this, because it is not traffic; it is the implicit references and borrowed vocabulary in your own writing. An execution retainer acts on a strategy but does not surface the scaffolding the strategy assumes. This map extracts it from your content and classifies each piece by accessibility, strategic value, and how to enter.

The evidence standard

Every piece in the map is grounded in one to three verbatim quotes pulled from the content samples you supply. The analysis is instructed that a quote must be an exact substring of a provided excerpt, that a piece it cannot quote directly must be left out, and that it must not invent shadow-infrastructure pieces. Where a category yields no signal, it is flagged as a gap rather than padded.

This is the standard the extraction is generated under, specified in the prompt, and the quotes printed beside each piece are your means of checking it against the sample. The map holds the standard at generation rather than through a separate automated drop gate, so we describe it as the standard the analysis runs to, and the reported methodology_confidence reflects actual extraction quality, not aspiration.

External presence check

Every shadow infrastructure piece we name is grounded in verbatim evidence from your own content. We then run a separate external check on each piece. We do not rank pieces by popularity. A niche practitioner forum and a large public community are both real infrastructure, and the value is often in the non-obvious one.

External presence check. Each named piece is checked against live search for a real external footprint and marked verified when one is found, or, honestly, absent when none is. Absent covers any piece we cannot establish from outside your own content. We never invent a presence we cannot establish. We report coverage: how many of your measured pieces resolved to a verified external presence.

What this check is and isn't. The check draws on live search results, and we link you to the external result that establishes a piece. The check confirms whether an entity is real and findable from outside your own content. It is not a measure of how influential or how large a piece is. Graded prominence, ranking pieces by reach or authority, is a deeper measurement we are building separately, and this report does not claim it.

What signals we analyze

At intake you provide your company URL, your ICP (or a prefilled archetype), two to eight content sample URLs, and an optional focus question. The extraction then reads:

• Explicit references: cited authors, named frameworks, specific certifications, conference names, tool names
• Implicit references: vocabulary patterns and methodology language used without attribution
• Your ICP description, which narrows category-fit and prevents cross-domain noise
• Your optional focus question, which weights the extraction toward a particular angle
• Reference patterns across multiple samples: a single mention is a data point, recurring references are the signal

How it works

The pipeline reads your samples, extracts the implicit methodology (your underlying framework, stated or unstated), and identifies named pieces with verbatim evidence quotes. For each piece we record:

• Category (one of eight: book, framework, certification, award, conference, individual, tool, community)
• Evidence (one to three verbatim quotes from your samples, never fabricated)
• Accessibility tier (easy: open sponsorship or paid placement; medium: relationship-building; hard: gated, requires existing reputation)
• Strategic value (high: transforms positioning; medium: incremental, table-stakes credibility; low: noise)
• Current relationship (absent, weak, moderate, strong)
• Entry tactics (two to three from a fixed menu: sponsorship, content collaboration, reference building, community participation, direct relationship, conference engagement, gift economy, joint research)

The founder reviews every extraction before report generation. The analysis is AI-first with a human edit pass, not pure human work, and we say so because the trust breaks the moment we are vague about it.

Accuracy and limitations

• Quality depends on the density of your content samples. Sparse or narrow samples produce fewer grounded pieces and a lower-confidence map, surfaced honestly in the report rather than padded. Categories with no signal in your samples are flagged as gaps, never fabricated.
• This is a diagnostic, not an engagement plan. The report identifies entry points; executing on them is separate work.
• Accessibility and strategic-value classifications reflect the state at extraction time and can shift as conferences add gating or communities close.

Anonymized example

Example only. Fictitious company and pieces. Real reports contain customer-specific extractions and are delivered only to the engaged customer.

{
  "methodology_summary": "Distribution-first B2B SaaS studio drawing on compounded-distribution thinking, with explicit community participation as a credibility surface.",
  "methodology_confidence": "medium",
  "shadow_infrastructure": [
    {
      "name": "Compounded Distribution by R. Hadley",
      "category": "book",
      "evidence": [
        "Our distribution playbook draws heavily on Compounded Distribution by R. Hadley"
      ],
      "accessibility_tier": "medium",
      "strategic_value": "high",
      "current_relationship": "moderate",
      "entry_tactics": [
        {"tactic": "content_collaboration", "rationale": "Already citing Hadley by name; publish a methodology breakdown linking the framework to case studies."}
      ]
    }
  ],
  "flagged_gaps": [
    "no certification references",
    "no award references"
  ]
}