# Insights Engineer

> Insights Engineer is about a new job in consumer insights: the person who builds the system that answers the question, instead of answering it again next month. A manifesto, a definition, and a two-week assessment for teams.

An insights engineer is a consumer insights professional who builds the system that answers a research question, instead of answering it again next month. They start from a decision the business has to make, build on work the team already does by hand, and keep score in hours given back and decisions changed. The title follows the pattern of analytics engineer, forward-deployed engineer and GTM engineer.

## Core pages

- [The Insights Engineer: a manifesto](https://insightsengineer.ai/manifesto): the founding argument for the role.
- [What is an insights engineer?](https://insightsengineer.ai/what-is-an-insights-engineer): definition, comparison with adjacent roles, skills and FAQ.
- [The two-week assessment](https://insightsengineer.ai/assessment): for insights teams. Recurring asks mapped, hours baselined, first use case and blueprint handed back.
- [About](https://insightsengineer.ai/about): why the site exists and how it is made.

## Writing

- Nothing published yet.

## Glossary

- [Agentic research](https://insightsengineer.ai/glossary/agentic-research): Research where AI agents plan, field, analyse and report, with a human setting the question and reviewing the output.
- [AI-moderated interview](https://insightsengineer.ai/glossary/ai-moderated-interview): A qualitative interview run by a conversational AI that asks follow-up questions in real time.
- [Always-on research](https://insightsengineer.ai/glossary/always-on-research): Research that runs continuously and reports on a schedule instead of arriving as a one-off study.
- [Analytics engineer](https://insightsengineer.ai/glossary/analytics-engineer): A data role that turns raw data into clean, tested, documented datasets, usually in SQL.
- [Automate the second time](https://insightsengineer.ai/glossary/automate-the-second-time): The rule that anything asked twice becomes a system, and anything asked monthly becomes a read that runs before anyone asks.
- [Automated quantitative research](https://insightsengineer.ai/glossary/automated-quantitative-research): Platforms that run advanced survey methods end to end, from questionnaire to sample to analysis, without a specialist vendor.
- [Brand tracker](https://insightsengineer.ai/glossary/brand-tracker): A recurring survey that measures awareness, consideration, usage and perception of a brand and its competitors.
- [Codeframe](https://insightsengineer.ai/glossary/codeframe): The set of themes or codes used to classify open-ended responses or transcript passages so they can be counted and compared.
- [Concept test](https://insightsengineer.ai/glossary/concept-test): A study that puts a product, service or message idea in front of consumers to gauge appeal, relevance and purchase intent before money is spent.
- [Conjoint analysis](https://insightsengineer.ai/glossary/conjoint-analysis): A quantitative method that infers how much people value each feature of a product by asking them to choose between bundles.
- [Consumer insights](https://insightsengineer.ai/glossary/consumer-insights): The function inside a company that understands customers' attitudes, motivations and behaviour.
- [Consumer insights manager](https://insightsengineer.ai/glossary/insights-manager): The mid-to-senior role that runs a brand's research programme and its vendors. The insights engineer takes over this role.
- [Decision memo](https://insightsengineer.ai/glossary/decision-memo): A one-page document that pairs a finding with a specific recommended action and the person who owns it.
- [Decisions moved](https://insightsengineer.ai/glossary/decisions-moved): A measure of insights value, counted as the business decisions that changed because of a finding and its recommendation.
- [Discussion guide](https://insightsengineer.ai/glossary/discussion-guide): The structured list of topics, questions and probes a moderator follows in a qualitative interview or group.
- [Ethnography](https://insightsengineer.ai/glossary/ethnography): Observing people in their own context, at home, in store or at work, instead of asking them questions in a research setting.
- [Forward-deployed](https://insightsengineer.ai/glossary/forward-deployed): Working inside the business team that owns the decision, instead of in a central research function down the hall.
- [Hours returned](https://insightsengineer.ai/glossary/hours-returned): A measure of insights value, counted as the operator hours given back by automating a recurring task.
- [Hypothesis-first brief](https://insightsengineer.ai/glossary/hypothesis-first-brief): A research starting point that names the decision, who makes it and the baseline before anyone touches data.
- [Implicit testing](https://insightsengineer.ai/glossary/implicit-testing): Methods that measure reactions people cannot or will not report directly, such as response-time association tests.
- [Incidence rate](https://insightsengineer.ai/glossary/incidence-rate): The percentage of people screened who qualify for a study. Low incidence makes fieldwork slow and expensive.
- [Insight agent](https://insightsengineer.ai/glossary/insight-agent): An AI agent that answers business questions from a company's past research and can run new research when the answer is not there.
- [Insight product](https://insightsengineer.ai/glossary/insight-product): A deliverable that keeps working after it ships, such as a dashboard, an agent, a decision tool or a scheduled read.
- [Insights analyst](https://insightsengineer.ai/glossary/insights-analyst): The traditional consumer insights role. Scopes studies, manages vendors, analyses results and presents the findings.
- [Insights engineer](https://insightsengineer.ai/glossary/insights-engineer): A consumer insights professional who builds the system that answers a research question, instead of running another study.
- [Insights repository](https://insightsengineer.ai/glossary/insights-repository): A searchable, tagged store of past research, findings and raw data that the whole organisation can query.
- [Insights stack](https://insightsengineer.ai/glossary/insights-stack): The data sources, tools and systems an insights function runs on, seen as infrastructure and not as a list of vendors.
- [MaxDiff](https://insightsengineer.ai/glossary/maxdiff): Maximum difference scaling. Respondents pick the best and worst item from small sets, which produces a ranked list with real spacing.
- [Net Promoter Score](https://insightsengineer.ai/glossary/net-promoter-score): A loyalty metric based on one question. How likely are you to recommend us, scored 0 to 10.
- [Open-end coding](https://insightsengineer.ai/glossary/open-end-coding): Assigning themes to free-text survey answers or transcripts so they can be counted and compared.
- [Qual at scale](https://insightsengineer.ai/glossary/qual-at-scale): Qualitative research run at sample sizes that used to be possible only in quantitative work.
- [Research design as code](https://insightsengineer.ai/glossary/research-design-as-code): Treating survey instruments, discussion guides, codeframes and analysis as versioned, testable, reusable files.
- [Research ops](https://insightsengineer.ai/glossary/research-ops): The operational side of research, covering recruiting, tooling, consent, repositories and process.
- [Research panel](https://insightsengineer.ai/glossary/research-panel): A pre-recruited pool of people who have agreed to take part in surveys and studies, usually for an incentive.
- [Segmentation](https://insightsengineer.ai/glossary/segmentation): Dividing a market or customer base into groups that differ in needs, attitudes or behaviour so each can be targeted differently.
- [Social listening](https://insightsengineer.ai/glossary/social-listening): Monitoring public conversation on social platforms, forums and review sites for mentions, themes and sentiment about a brand or category.
- [Synthetic panel](https://insightsengineer.ai/glossary/synthetic-panel): A group of synthetic respondents built to mirror a target audience.
- [Synthetic respondent](https://insightsengineer.ai/glossary/synthetic-respondent): An AI-generated persona that answers research questions in place of, or ahead of, a human respondent.
- [Text analytics](https://insightsengineer.ai/glossary/text-analytics): Automated analysis of unstructured text, such as reviews, support tickets and open-ends, for themes and sentiment.
- [Triangulation](https://insightsengineer.ai/glossary/triangulation): Checking a finding against more than one source or method before acting on it.

## Optional

- [RSS feed](https://insightsengineer.ai/feed.xml)
- [Sitemap](https://insightsengineer.ai/sitemap.xml)