rampstackco/claude-skills / skills/discovery-research-synthesis
discovery-research-synthesis
VerifiedTurning research artifacts into actionable PM insight. Customer interviews, user research notes, support ticket reviews, sales call transcripts, survey data, in-app feedback, all synthesized into the decisions they are meant to inform. The discipline of moving from raw discovery data to clear product direction without losing signal in the synthesis or fabricating insight that was not actually there. Triggers on research synthesis, customer interview synthesis, user research analysis, discovery readout, research insights, sales call analysis, support ticket analysis, qualitative data analysis. Also triggers when a team has done research but cannot turn it into decisions, when synthesis is producing pretty decks but no roadmap movement, or when an upcoming PM decision needs to be grounded in research already conducted.
| File | Kind | Size |
|---|---|---|
| README.md | markdown | 369 B |
| SKILL.md | skill md | 24.6 KB |
| references/common-discovery-synthesis-failures.md | markdown | 11.3 KB |
| references/from-pattern-to-product-implication.md | markdown | 9.8 KB |
| references/pattern-naming-patterns.md | markdown | 8.8 KB |
| references/research-types-and-when-each-fits.md | markdown | 10.4 KB |
| references/synthesis-review-and-validation.md | markdown | 10.7 KB |
| references/synthesis-sequence-walkthrough.md | markdown | 11.2 KB |
| references/tagging-and-clustering-discipline.md | markdown | 9.3 KB |
| references/when-to-gather-more-data.md | markdown | 11.0 KB |
| references/writing-for-decisions-not-decks.md | markdown | 10.4 KB |
references/common-discovery-synthesis-failures.md
View on GitHub# Common discovery synthesis failures
Eleven-plus failure patterns with diagnoses and cures. The cross-cutting pattern: most synthesis failures share a single root, which is treating synthesis as documentation work rather than as decision work.
---
## Failure 1: We did the research but never made decisions from it
**Symptom.** Discovery cycle completed; transcripts and notes filed; the team's roadmap discussions in the following months never reference the research.
**Diagnosis.** Data-dump or insight-theater pattern. Synthesis sequence either skipped entirely (data-dump) or stopped at findings without reaching decisions (insight-theater).
**Cure.** Run the synthesis sequence end-to-end on the existing artifacts. Even months after the original research, retroactive synthesis can produce decisions the original research was meant to inform.
**Prevention.** Before commissioning the next research cycle, name the decisions the research is intended to inform. Build the synthesis sequence into the discovery plan, not as an afterthought.
---
## Failure 2: Our synthesis deck got a polite review and was never referenced again
**Symptom.** A polished readout deck. A 30-minute presentation. Positive feedback in the meeting. Six months later nobody references the synthesis in product decisions.
**Diagnosis.** Insight-theater. Output was performative rather than load-bearing. Optimized for the presentation moment, not for reference use.
**Cure.** Convert the deck content into a document with patterns leading, evidence in appendix, decision-input section. Recirculate as a reference document.
**Prevention.** Future synthesis ships as documents, not decks. The presentation is a deck if needed; the synthesis is the document.
---
## Failure 3: We have 30 named patterns and no clear priorities
**Symptom.** The synthesis lists many patterns. The team agrees they are interesting but cannot prioritize. The synthesis fails to drive a roadmap shift.
**Diagnosis.** Category bloat. Under-clustering produced inflation. Many of the 30 patterns are variations on a smaller number of underlying patterns.
**Cure.** Re-cluster. Take the pattern list back to the clustering stage and look for the 4-8 underlying patterns the data actually supports. Synthesize from the consolidated list.
**Prevention.** Cap pattern count at 4-8 for typical synthesis. When more than 8 patterns emerge, that is a signal to recluster, not to ship with bloat.
---
## Failure 4: Every pattern reads as something we already knew
**Symptom.** The synthesis names patterns that read as truisms. Team members who attended the readout could have produced the patterns without the research.
**Diagnosis.** Pattern names captured truisms rather than the specific observations the research surfaced. The analytical work of pattern naming was skipped.
**Cure.** Re-examine the data for what was actually surprising. Often the strongest patterns surface in the contradictions, the unexpected sub-segments, the workarounds users built. Pattern names should commit positions the team would not have committed before the research.
**Prevention.** During pattern naming, apply the "what the team did not know" test. Patterns that read as common-knowledge are probably under-named.
---
## Failure 5: The patterns contradict each other and we cannot reconcile them
**Symptom.** Two patterns point in opposite directions. The synthesis is silent on the contradiction or hand-waves it.
**Diagnosis.** Often segment misfit: different segments have different needs, and the synthesis did not surface the segment distinction. Sometimes context misfit: different use contexts produce different patterns.
**Cure.** Re-examine the data for the segment or context dimension that explains the contradiction. The contradiction often becomes the most valuable finding once explained.
**Prevention.** Tag dissent and contradictions explicitly during the tagging stage. Surface them through clustering. Address them in synthesis rather than burying them.
---
## Failure 6: Research participants did not recognize themselves in our synthesis
**Symptom.** When shown the synthesis, participants say "I did not say that" or "this does not match my experience."
**Diagnosis.** Synthesis fabricated patterns or projected interpretations not in the data. The pattern naming reached for what the team wanted to find rather than what the participants actually said.
**Cure.** Restart with closer evidence-tracking. Each pattern should be traceable to specific moments in specific artifacts. Patterns that cannot be traced are projections, not observations.
**Prevention.** Build participant review into the synthesis process. Synthesis that participants do not recognize fails the validation; revise before publishing.
---
## Failure 7: The synthesis is detailed but the team cannot tell what to do
**Symptom.** Patterns are clear and well-supported. The team agrees on the patterns. The team cannot derive the product implications.
**Diagnosis.** Implications missing or vague. The synthesis stopped at observations without bridging to product responses.
**Cure.** Add implications to each pattern. Each implication should be specific, falsifiable, cost-acknowledged, and tied to a specific decision.
**Prevention.** The synthesis sequence's implication stage is non-negotiable. Synthesis that ships without implications is incomplete.
---
## Failure 8: We hedged every pattern and the synthesis reads as wishy-washy
**Symptom.** "Users may sometimes occasionally seem to indicate." Every claim caveated. The synthesis avoids commitment.
**Diagnosis.** Hedge stacking. The researcher did not commit to what the data showed.
**Cure.** Rewrite with conviction. Name patterns as if the data supports them. If the data does not support the position, name a smaller position the data does support; do not hedge an oversized claim.
**Prevention.** During pattern naming, apply the conviction test. Hedge stacking signals a pattern that should either be sharpened or dropped.
---
## Failure 9: Our quote walls were impressive but the deck was 60 slides
**Symptom.** Beautiful quote pulls; designed personas; 60 slides. Readers admired the deck. Few remember the patterns.
**Diagnosis.** Insight-theater. Evidence dominated; analysis receded. Quote walls without analytical bridge.
**Cure.** Restructure as a document. Patterns lead; evidence in appendix. Quote walls converted into evidence-summary paragraphs supporting each pattern.
**Prevention.** The format choice is upstream. Default to documents; reserve decks for short executive summaries paired with substantive documents.
---
## Failure 10: The synthesis recommended 20 product changes; we shipped one
**Symptom.** Synthesis presents many implications. Team picks the easiest or most-vocal one to ship. The rest sit in the synthesis document unaddressed.
**Diagnosis.** Wishlist synthesis. Implications did not acknowledge cost or prioritize within themselves. The synthesis offloaded prioritization to the team without giving them a starting point.
**Cure.** Re-prioritize within the synthesis. Each implication gets a cost note. The synthesis recommends a sequence: which implications to address now, which to defer, which to drop. The team has a starting point for prioritization.
**Prevention.** During implication writing, surface cost. The synthesis owner uses available context to estimate cost; pairs with engineering for harder estimates. Cost-blind synthesis is incomplete.
---
## Failure 11: Adjacent teams disagreed with our patterns and we did not engage
**Symptom.** Synthesis ships. Support team says "this does not match what we see in tickets." Sales team says "our prospects say something different." The synthesis fails to align with the teams closest to the customer.
**Diagnosis.** Adjacent-team review skipped or compressed. The synthesis reached publication without surfacing perspectives the research team did not have.
**Cure.** Pause publication. Run the adjacent-team review even after the fact. Revise based on inputs.
**Prevention.** Build adjacent-team review into the synthesis workflow. Make it non-negotiable. The teams closest to the customer have ground-truth observations the research team needs.
---
## Failure 12: Data gaps buried in the synthesis
**Symptom.** Synthesis presents conclusions where the data did not support them. The recommendations sound confident; the data is thin in places the synthesis did not name.
**Diagnosis.** The we-have-enough trap. Sunk-cost reasoning suppressed honest gap-naming.
**Cure.** Audit each pattern for data sufficiency. Where gaps exist, name them explicitly. Where recommendations rest on under-supported patterns, downgrade or hedge them honestly.
**Prevention.** The synthesis owner's job is to be honest about what the data can and cannot support. The team relies on the synthesis owner for this honesty.
---
## Failure 13: Synthesis sequence stages skipped
**Symptom.** Synthesis arrives at "patterns" without going through the tagging and clustering stages. Patterns feel imposed on the data rather than emerging from it.
**Diagnosis.** Stages skipped. Researcher jumped to pattern-naming based on impressions from reading transcripts.
**Cure.** Restart the sequence from tagging. Patterns from skipped sequences are usually projections; tagging-grounded patterns are observations.
**Prevention.** The sequence is non-negotiable. Every synthesis runs through transcribe, tag, cluster, name, imply, so-what.
---
## Failure 14: Synthesis owner became defensive in the review
**Symptom.** Adjacent teams raise concerns; synthesis owner argues against each one; the review session produces no revisions; synthesis ships unchanged.
**Diagnosis.** Defensive review posture. The owner treated review as defense of the synthesis rather than as input to it.
**Cure.** Owner takes notes; does not defend in the moment; revises based on inputs over the following days. Reviewer concerns surface in the revision; the synthesis improves.
**Prevention.** Establish the review session's purpose explicitly: input, not approval. The synthesis owner's job in the session is to listen.
---
## The cross-cutting pattern
Most synthesis failures share a single root: treating synthesis as documentation work rather than as decision work.
Documentation work prioritizes capturing the research; producing a deliverable; performing for an audience. Output: artifacts that look like synthesis but do not drive decisions.
Decision work prioritizes producing direction; committing positions; informing prioritization. Output: synthesis that team members reference for months, that drives roadmap shifts, that earns the research investment.
The fix for almost any synthesis failure starts with the same question: what decision is this synthesis supposed to inform? When the answer is clear, the synthesis becomes load-bearing. When the answer is "we should document the research," the synthesis fails as decision input.
---
## Methodology-level choices that stay in the public skill
The fourteen failure patterns with diagnoses and cures. The cross-cutting documentation-vs-decision pattern.
## Implementation choices that stay internal
Specific reviewer training on synthesis-failure detection. Specific automation that flags synthesis structures (e.g., document length checks, hedge-frequency analysis). Specific reviewer dashboards. The team's own conventions for catching synthesis failures pre-publish. These vary by team.
Related skills
| Skill | Stars | Updated | Works with | Status | Install |
|---|---|---|---|---|---|
| add-product-e2e-evalpaperclipai/paperclipAdd or extend a Paperclip full-stack runner E2E workflow, fixture, matcher, or report evidence path for local or Daytona execution. | Stars: 115k | Updated | Claude CodeCodex | Verified | Install |
| add-runner-evalpaperclipai/paperclipAdd or extend a Paperclip Runner protocol evaluation definition, roster, assertion, or report fixture with provenance and narrow validation. | Stars: 115k | Updated | Claude CodeCodex | Verified | Install |
| diagnose-why-work-stoppedpaperclipai/paperclipCompatibility entry point for failed, stalled, looping, or over-recovered Paperclip work. Use fix-failed-paperclip-run for the full diagnosis, regression, repair, real verification, and green-PR workflow; honor an explicitly requested diagnosis-only scope. | Stars: 115k | Updated | Claude CodeCodex | Verified | Install |
| fix-failed-paperclip-runpaperclipai/paperclipDiagnose and fix failed, stuck, slow, or incorrectly blocked Paperclip tasks and runs. Use with a task/run URL or failure report to trace the root cause, prove a regression red then green, check second-order effects, verify the real workflow, and deliver a PR with passing checks. | Stars: 115k | Updated | Claude CodeCodex | Verified | Install |