Signal & Noise blog

How to generate user or customer personas from real interview and survey data with AI, the persona types that exist, and how to avoid baking in bias from skewed input.

AI bias in qualitative analysis means a model's errors correlate with who said what. What the research on GPT, Llama and Claude found, and how to catch it.

The Critical Incident Technique (CIT) asks for specific remembered events, not general opinions. What it is, how to run it, and how to analyse incidents at scale.

How to analyse patient and member free-text feedback at scale with AI, while handling health data anonymisation and care-touchpoint mapping properly.

Skimle Ask is a simple AI interviewer for quick polls, like Doodle for opinions or Google Forms for data. Qual at scale: how to launch one and analyse results.

How engineering leaders use AI to find recurring root-cause themes across dozens of incident postmortems, with traceability back to source.

How IR teams and analysts use AI for earnings call transcript analysis: management language, analyst Q&A themes, quarter-over-quarter shifts.

How Monitoring, Evaluation, and Learning (MEL) teams can analyse beneficiary interviews, FGDs, and MSC stories at scale, with donor traceability and native multi-language support.

Qualitative evidence synthesis combines findings across published studies. This guide covers meta-ethnography, thematic synthesis, and CASP appraisal.

How to analyse large FOIA releases and leaked documents with AI: find recurring themes and contradictions, and trace every claim to its source page.

How to synthesise customer advisory board feedback into recurring themes with full traceability, instead of relying on a manual spreadsheet nobody keeps updated.

How to analyse skip-level meeting notes at scale and tell a systemic problem from one person's complaint, with full traceability back to source.

AI for public comment analysis helps agencies in the US, EU, and elsewhere triage and code thousands of regulatory submissions with full traceability.

Descriptive coding assigns summary labels to qualitative data passages using a fixed codebook. Here is how to build one, apply it consistently, and use AI to scale the process.

An interview guide is a structured set of questions and topic areas used to run qualitative interviews consistently. Here is how to write one, what to include, and a template you can use.

Customer discovery interviews reveal whether the problem you're solving is real and who it matters to most. Learn how to find participants, run interviews, and turn findings into confident product dec...

Analysing hundreds of written consultation responses requires systematic coding, metadata by respondent type, and clear audit trails. This guide covers the full workflow for public and regulatory cons...

A practical guide to consulting primary research: scoping the right questions, designing the interview guide, managing fieldwork, and synthesising 20+ interviews into a clear recommendation.

Step-by-step guide to doing thematic analysis in Excel — spreadsheet setup, colour coding, and pivot tables for qualitative data. Plus when Excel stops working and what to use instead.

A thematic analysis results section presents each theme as a heading, supported by 2-4 verbatim quotes with analytical commentary. This guide shows the structure, worked examples, and what reviewers l...

JTBD research reveals why customers hire or fire products by exploring the progress they're trying to make. Learn the Switch interview structure, how to code JTBD data, and how to act on the findings.

Voice of customer research captures customers' expectations and experiences to drive product, marketing, and retention decisions. Learn how to design, collect, and analyse VoC data at scale.

Content analysis counts and categorises text systematically; thematic analysis interprets patterns of meaning. Learn the 5 key differences and when to choose each method.

IPA explores lived experience through small, purposive samples and deep individual case analysis. Learn when to use IPA, how the 6-step process works, and how it differs from thematic analysis.