Iterative Depth

This framing will be tested and refined by each lens. Expect it to evolve.

Step 2 — Select lenses

Choose 3-5 lenses from the eight below. Select based on the nature of the problem — not every lens applies to every problem. The table includes guidance on when each lens is most valuable.

The eight lenses

Lens selection rules

• Always include at least one "human" lens (User or Business) and one "system" lens (Technical or Edge case)
• If the problem involves money, include Business
• If the problem involves security or reliability, include Adversarial
• If the problem has long-term consequences, include Temporal
• If the problem feels complex, include Simplicity as a counterweight
• If you're unsure which lenses to pick, start with User + Technical + Adversarial — this combination catches the most issues

State which lenses you selected and why.

Step 3 — Analyse through each lens

For each selected lens, follow this exact structure:

### Lens [N]: [Name]

**Focus question:** [The specific question this lens asks about this problem]

**Analysis:**
[Deep analysis from this perspective. Not a surface-level paragraph — spend real effort here.
Ask and answer at least 3 sub-questions specific to this lens and this problem.]

**Findings unique to this lens:**
1. [Something this lens revealed that no previous lens surfaced]
2. [Another unique finding]
3. [Another unique finding]

**Contradictions with previous lenses:**
- [Where this lens disagrees with a finding from a previous lens, and why]
- [If no contradictions: "No contradictions — this lens reinforces [specific finding] from [lens name]"]

**Risks identified:**
| Risk | Likelihood | Impact | Mitigation |
|---|---|---|---|
| [risk] | High/Med/Low | High/Med/Low | [what to do] |

**Open questions:**
- [Questions this lens raised that need answers from stakeholders, data, or further investigation]

Rules for per-lens analysis

• Go deep, not wide. Each lens should produce genuine insight, not a paragraph that restates the obvious. Ask "so what?" after every finding — if the answer is "obviously," dig deeper.
• Findings must be specific to this problem. "Users want good UX" is not a finding. "Users in this workflow switch between three tools, so a solution that requires a fourth tool will face adoption resistance" is a finding.
• Contradictions are the most valuable output. When two lenses disagree, you've found a tension that requires a deliberate trade-off. Don't smooth over contradictions — highlight them.
• Each lens must produce at least one finding not surfaced by any previous lens. If a lens adds nothing new, either the analysis was too shallow or the lens was a poor choice — go deeper or swap it.

Step 4 — Synthesise

After completing all lenses, synthesise the findings. This is where the real value emerges.

Convergent findings

What appeared across multiple lenses? These are your highest-confidence conclusions.

| Finding | Supported by lenses | Confidence | Implication |
|---|---|---|---|
| [finding] | [lens names] | High / Medium | [what this means for the decision] |

A finding supported by 3+ lenses is almost certainly important. Act on it.

Tensions and trade-offs

Where do lenses contradict? These require deliberate choices.

| Tension | Lens A says | Lens B says | Recommended resolution | Rationale |
|---|---|---|---|---|
| [topic] | [position] | [position] | [which side to favour] | [why] |

For each tension, make a recommendation. Don't leave trade-offs unresolved — the whole point of this analysis is to reach a better decision.

Blind spots filled

What did later lenses reveal that earlier ones missed? This demonstrates the value of multi-lens analysis and highlights areas where single-perspective thinking would have failed.

| Blind spot | Revealed by | Would have caused | Prevention |
|---|---|---|---|
| [what was missed] | [lens name] | [consequence of missing it] | [how to catch this in the future] |

Revised problem framing

Based on the analysis, rewrite the problem statement from Step 1. It should now be sharper, more nuanced, and account for what you learned.

Original framing: [from Step 1]
Revised framing: [incorporating insights from all lenses]
Key difference: [what changed and why it matters]

Refined success criteria

Based on all lenses, define what "success" looks like:

| Criterion | Source lens | Metric | Threshold |
|---|---|---|---|
| [criterion] | [which lens identified this] | [how to measure] | [what "good" looks like] |

Recommendation

State a clear recommendation with:

**Recommendation:** [what to do]

**Confidence:** [High / Medium / Low] — based on [what evidence]

**Key risks:** [top 2-3 risks and their mitigations]

**Next steps:**
1. [Immediate action]
2. [Follow-up action]
3. [Validation step]

**What would change this recommendation:**
- [If X turns out to be true, reconsider Y]
- [If Z happens, switch to approach W]

Rules

• The lenses are a tool for thinking, not a template to fill in. If a lens isn't generating insight, go deeper before moving on.
• Contradictions between lenses are not problems — they're the most valuable output. They reveal trade-offs that must be made consciously.
• Every finding must be actionable. "This is complex" is not a finding. "This has three competing constraints that require a trade-off between X and Y" is actionable.
• Don't anchor on the first lens. The purpose of multi-lens analysis is to challenge your initial perspective, not confirm it. If every lens agrees with the first one, you're not trying hard enough.
• State your confidence level honestly. An analysis with "Medium confidence due to missing user data" is more useful than one with false certainty.
• The synthesis matters more than any individual lens. Spend at least as much effort on Step 4 as on any single lens in Step 3.

Related Skills

• /council — when iterative-depth reveals irreconcilable tensions between lenses, convene a council to debate the trade-offs.
• /first-principles — when a lens reveals that the problem framing itself is wrong, decompose to fundamentals.