Quote

Origin & evolution

The case for verbatim language is strongest in qualitative method, where exact wording is treated as data rather than decoration. Work on reporting qualitative social research sets out the functions a direct quotation performs: it gives the reader the evidence to judge the researcher's interpretation, it lends the account credibility by showing the analysis rests on real speech, and it preserves nuance that paraphrase flattens. In ethnography and discourse analysis the syntax itself is evidence, so altering it erases part of the record.

There is a live methodological debate about how far to clean a quote. Methods writing on transcription distinguishes true verbatim, which keeps every "um" and false start, from clean verbatim, which removes fillers while leaving word choice intact. The first protects analytical fidelity, the second improves readability and can spare a participant from looking inarticulate in print. Most teams clean lightly for reporting while keeping the true verbatim in the repository, so the unedited record is always recoverable.

In product practice the quote earns a second role beyond evidence: persuasion. Tomer Sharon's atomic-research model, set out in his foundations piece, treats a clip or transcript line as the evidence attached to an observation👁️ObservationUser ResearchA specific behaviour or statement observedView reference →, the thing that makes a finding replayable for a doubting stakeholder🧑‍💼StakeholderTeam & OrganisationA person with influence over the productView reference →. A roadmap🛣️RoadmapProduct SpecificationA strategic plan of features and milestonesView reference → argument backed by a customer's own sentence travels further inside an organisation than the same point in a researcher's summary.

How it works in practice

A research team reports that onboarding is too slow. The claim lands softly until they attach the recording: a new admin, four minutes into setup, saying "honestly, I was about to close the tab and email your sales guy." Engineering, which had treated onboarding speed as a nice-to-have, reprioritises it within the week. The exact words did the work. "About to close the tab" names the stakes more sharply than any churn statistic, and "email your sales guy" reveals the fallback path the company never wanted users to take. A paraphrase, "the user found onboarding frustrating," would have changed nobody's mind.

Quote vs. its neighbours

• Observation is the recorded event, which may be a behaviour with no speech at all. A quote is specifically the participant's language, and it often serves as the evidence pinned to an observation.
• Evidence is the broader category of proof behind a finding, including screen recordings, analytics, and screenshots. A quote is one form of evidence, the verbal one, and the most quotable in a readout.
• Insight🔦InsightUser ResearchA synthesised finding from researchView reference → is the synthesised conclusion a quote can illustrate but never equals. A single vivid line is one person's experience; an insight is the pattern that holds across many. Treating a striking quote as an insight is cherry-picking dressed as research.

In the graph

In the Unified Product Graph, a quote is the verbatim evidence node of the User Research region, and the edges keep it tethered to what it proves rather than floating as a pull-quote. It backs a recorded event through observation_evidenced_by_quoteObservationevidenced byQuotehierarchy, and it can substantiate the structures the research feeds: quote_evidences_needQuoteevidencesNeedcross-domain links a sentence to the requirement it reveals, and quote_evidences_jobQuoteevidencesJobcross-domain ties it to the progress the user is trying to make. A sales or success objection🚧ObjectionGo-To-MarketA common sales objectionView reference → traces to its source through objection_sourced_from_quoteObjectionsourced fromQuotecross-domain, so a stated barrier carries the customer's own phrasing into go-to-market work. The link quote_relates_to_jobQuoterelates toJobcross-domain keeps a user's language attached to the job🔨JobUserJob To Be Done: what the user is trying to accomplishView reference → it concerns, which means any claim downstream can be checked against the words that started it.