{"site":{"name":"Koji","description":"AI-native customer research platform that helps teams conduct, analyze, and synthesize customer interviews at scale.","url":"https://www.koji.so","contentTypes":["blog","documentation"],"lastUpdated":"2026-06-07T15:12:55.329Z"},"content":[{"type":"documentation","id":"a05f6103-f98e-4214-af59-979d4fb911a5","slug":"secondary-research-guide","title":"Secondary Research: The Complete Guide to Desk Research for Product and UX Teams","url":"https://www.koji.so/docs/secondary-research-guide","summary":"Secondary research (desk research) is the practice of extracting insights from existing data sources before conducting primary research. A 5-step process: define the research question, identify high-signal sources, extract systematically, categorize findings as well-established/contested/unknown, and write a synthesis summary. Internal sources include analytics, support tickets, and previous research. External sources include industry reports, competitor reviews, academic papers, and online communities. Secondary research cannot replace primary research — it lags behind reality, covers market averages not specific users, and cannot surface unarticulated needs. AI-moderated interview platforms like Koji have eliminated the speed advantage that secondary research historically held over primary research.","content":"# Secondary Research: The Complete Guide to Desk Research for Product and UX Teams\n\n**Bottom line:** Secondary research (also called desk research) is the practice of gathering insights from existing data sources — industry reports, academic studies, competitor publications, analytics, and internal documents — before conducting primary research. Done well, it saves weeks of redundant effort, focuses your interviews on genuine unknowns, and makes your primary research dramatically more productive.\n\nEvery research project should start with secondary research. Not because it replaces talking to users — it does not — but because it tells you what is already known so you can focus your user interviews on what is not.\n\n## What Is Secondary Research?\n\nSecondary research is any research that analyzes or synthesizes data originally collected for a different purpose. You are not generating new data — you are extracting insights from existing sources.\n\nIt is called \"desk research\" because historically you could conduct most of it without leaving your desk (as opposed to fieldwork, which requires going out to observe and interview real people in their context).\n\n**The fundamental distinction:**\n\n| | **Secondary Research** | **Primary Research** |\n|---|---|---|\n| Data source | Already exists | You collect it |\n| Time to insights | Hours to days | Days to weeks |\n| Cost | Low | Medium to high |\n| Answers | What is known about the market/domain | What your specific users experience |\n| Limitations | Lags behind reality; covers averages | Requires participant recruitment |\n| Best for | Context-setting, hypothesis formation | Validating hypotheses, uncovering your users' specific needs |\n\n## Types of Secondary Research Sources\n\n**Internal sources** — data your organization already has:\n- Product analytics and behavioral usage data\n- Previous research studies and synthesis reports\n- Customer support ticket logs and recurring themes\n- Sales call recordings, win/loss notes, and CRM data\n- NPS and CSAT survey verbatims\n- Onboarding and activation funnel data\n- Customer success team knowledge\n\n**External sources** — data available outside your organization:\n- Industry analyst reports (Gartner, Forrester, IDC, Nielsen)\n- Academic research papers and peer-reviewed journals\n- Government statistics and public datasets\n- Competitor blogs, case studies, help documentation, and job postings\n- App store reviews, G2, Capterra, and Trustpilot reviews\n- Online communities (Reddit, LinkedIn groups, industry Slack communities, forums)\n- News coverage, press releases, and earnings call transcripts\n\n## Why Secondary Research Comes Before Primary Research\n\nThe argument for starting with secondary research is straightforward: do not ask questions you already know the answers to.\n\nWhen product teams skip secondary research and go straight to interviews, they routinely spend time asking questions that have already been thoroughly answered in existing reports, confirming obvious hypotheses rather than uncovering new ones, and missing important context that would have changed the direction of their research entirely.\n\nAccording to Nielsen Norman Group: \"Secondary research in UX tells the team what is already known, surfaces existing answers, and points the primary work toward the gaps that still matter.\" This is the compounding effect of good desk research — it does not just save time; it improves the quality of everything that follows.\n\nConsider two researchers preparing to study enterprise software adoption:\n\n**Researcher A** (skipped secondary research): Spends the first 5 interviews asking about the general buying process, rediscovering well-documented patterns about committee decisions and procurement involvement.\n\n**Researcher B** (completed secondary research): Knows committee buying is already well-documented. Asks focused questions about the specific friction points in the technical evaluation phase — a gap she identified in the literature — and surfaces insights no existing report contains.\n\nResearcher B produces more original, more actionable findings with the same number of interviews.\n\n## The 5-Step Secondary Research Process\n\n### Step 1: Define Your Research Question\n\nSecondary research without a clear question produces interesting but unfocused notes. Be specific: \"What do we already know about why enterprise software pilots fail to convert to paid?\" produces better desk research than \"learn about enterprise buying.\"\n\nWrite your research question at the top of your notes document before you begin searching. Return to it repeatedly to keep your review focused.\n\n### Step 2: Identify Your Highest-Signal Sources\n\nNot all sources are equal. Prioritize by credibility and relevance:\n\n- **For market behavior and quantitative benchmarks**: Gartner, Forrester, IDC, Statista, government datasets\n- **For UX and user behavior patterns**: Nielsen Norman Group, Baymard Institute, academic UX journals\n- **For competitive landscape**: Company blogs, G2/Capterra reviews, LinkedIn job postings (which reveal what competitors are building), changelog pages\n- **For your own users**: Internal analytics, support tickets, previous research reports, sales call notes\n- **For industry discourse and unarticulated needs**: Reddit communities, Slack groups, LinkedIn posts, industry newsletters\n\nPrioritize recent sources. Reports more than two years old may describe a market that has changed substantially.\n\n### Step 3: Extract Systematically, Not Passively\n\nDo not just read and highlight. For each source, capture in a structured notes document:\n\n- The key claim or finding\n- Source details (author, publication, year, URL)\n- Your confidence in the source's rigor (peer-reviewed journal vs. vendor-published report vs. community post)\n- Whether this finding confirms, contradicts, or is silent on your specific research question\n\nStructure matters. You will reference these notes when designing your interview guide, and the structure will determine how usable those notes are.\n\n### Step 4: Identify the Three Categories of Information\n\nAfter reviewing multiple sources, you will see findings fall into three categories:\n\n**Well-established**: Multiple credible sources agree on this. No need to re-investigate in primary research. Treat as a known assumption at the start of your study.\n\n**Contested**: Sources disagree, or the evidence is thin and inconclusive. Worth exploring in primary research to understand the disagreement through your specific users' lens.\n\n**Unknown / unexplored**: Nobody seems to have studied this in the context that matters to you. This is your primary research opportunity — the questions where user interviews can surface genuinely original insight.\n\nYour interview guide should focus primarily on contested and unknown topics. This is how secondary research directly improves the quality and originality of your primary research.\n\n### Step 5: Write a Secondary Research Summary\n\nBefore beginning primary research, write a 1–2 page synthesis document that includes:\n- Key established facts relevant to your question\n- Hypotheses that emerged from the secondary research\n- Specific gaps that primary research will address\n- Revised scope for primary research based on what you learned\n\nShare this with stakeholders before primary research begins. It demonstrates rigor, surfaces misaligned assumptions early, and sets realistic expectations for what primary research will and will not answer.\n\n## When Secondary Research Is Not Enough\n\nSecondary research has significant limitations that make primary research essential, not optional:\n\n**It covers averages, not your users**: Industry reports describe markets. They do not describe your specific users in your specific product context. A report stating that \"60% of enterprise buyers involve IT in the purchase decision\" does not tell you whether your buyers do — or what specifically triggers IT involvement in your category.\n\n**It lags behind reality**: Published reports typically reflect data collected 6–18 months before publication. Fast-moving markets — AI tools, consumer platforms, emerging categories — can change dramatically in that window.\n\n**It cannot answer behavioral questions at depth**: Secondary data tells you what happened. Only primary research can tell you why, and in enough depth to act on it. \"Users abandon checkout at the payment step\" is visible in your analytics. Why they abandon requires a conversation.\n\n**It reflects what people publish, not what they experience**: Organizations publish case studies of their successes, not their failures. Online reviews over-represent extreme experiences (delighted or furious). Job postings reflect aspiration, not current reality. Secondary data has systematic selection biases that primary research must correct.\n\n**It cannot surface unarticulated needs**: By definition, secondary research only reveals what is already known. If you need genuinely novel insight — understanding a new user segment, discovering jobs-to-be-done that nobody has named yet, validating a hypothesis nobody has tested — only primary research delivers.\n\n## Secondary and Primary Research in Practice: A Product Discovery Example\n\nHere is how a product team might combine secondary and primary research for a new feature decision:\n\n**Research question**: Should we build a native analytics dashboard, or integrate with existing tools?\n\n**Secondary research phase (2–3 days)**:\n- Review internal analytics: how many customers have connected third-party integrations?\n- Read Gartner and Forrester reports on the analytics tool market in this category\n- Review G2 and Capterra competitor reviews: what do users say about analytics in similar products?\n- Search relevant Reddit communities: what do practitioners say about analytics workflows?\n- Review competitor job postings: are they hiring data engineers? (signals investment direction)\n\n**Synthesis**: Secondary research reveals that third-party integration adoption is low industry-wide, but G2 reviews consistently mention that built-in analytics lack the export/sharing capabilities practitioners need. The gap: nobody has studied specifically how teams share insights with non-technical stakeholders.\n\n**Primary research focus**: Rather than asking broad questions about analytics needs, the interview guide focuses specifically on the insight-sharing and reporting workflow — a gap identified through secondary research. This produces original findings in a fraction of the interviews it would have taken without the prior context.\n\n## How Modern AI Research Tools Eliminate the Speed Gap\n\nOne historical argument for prioritizing secondary research over primary was speed: you could review five reports in a day, but scheduling and conducting five user interviews took a week.\n\nThat asymmetry has largely been eliminated by AI-moderated research platforms.\n\nWith Koji:\n- Launch an interview study in under 30 minutes\n- Interviews run asynchronously — participants complete them when it is convenient for them, 24/7\n- AI-generated thematic analysis is available within hours of the last interview completing\n- Structured questions across Koji's 6 question types (open-ended, scale, single choice, multiple choice, ranking, and yes/no) capture both qualitative depth and quantitative benchmarks in a single session\n\nThe result: primary research now takes roughly the same calendar time as thorough secondary research. Teams no longer face a meaningful trade-off between the two methods — they can run both in tight sequence.\n\n**A modern integrated research workflow:**\n\n1. **Day 1**: Define research question. Identify top 5–8 secondary sources.\n2. **Days 2–3**: Review and synthesize secondary sources. Identify gaps.\n3. **Day 3**: Configure a Koji study targeting the specific gaps. Write structured questions using Koji's 6 question types to capture both qualitative themes and quantitative benchmarks.\n4. **Days 4–7**: Interviews run asynchronously. Participants complete sessions at their convenience.\n5. **Day 7**: Review AI-generated thematic analysis, read standout transcripts for quotes.\n6. **Days 8–9**: Synthesize secondary + primary findings into recommendations.\n\nTotal cycle time: approximately 9 days for a complete, defensible research project with both secondary context and primary user evidence.\n\n## Key Secondary Research Sources to Bookmark\n\nFor product and UX researchers, these sources consistently deliver high-signal secondary data:\n\n| Source | Best for |\n|--------|---------|\n| Nielsen Norman Group (nngroup.com) | UX methods, user behavior patterns, research rigor |\n| Baymard Institute (baymard.com) | E-commerce UX, checkout and form design benchmarks |\n| State of User Research Report (User Interviews, annual) | Research practice benchmarks, team structures |\n| Google Scholar (scholar.google.com) | Academic research on human behavior and cognitive patterns |\n| Statista (statista.com) | Statistical data across industries and geographies |\n| G2 / Capterra / Trustpilot | Aggregate user reviews as qualitative secondary data |\n| Gartner / Forrester (paid) | Enterprise market sizing, buyer behavior, technology adoption |\n| Reddit (relevant subreddits) | Unfiltered practitioner discourse, unarticulated frustrations |\n\n## Related Resources\n\n- [Structured Questions in AI Interviews](/docs/structured-questions-guide) — How to design primary research using Koji's 6 question types to fill the gaps secondary research reveals\n- [User Interview Guide](/docs/user-interview-guide) — The complete guide to conducting primary user interviews\n- [Product Discovery Research Guide](/docs/product-discovery-research-guide) — How to integrate secondary and primary research across the full product discovery process\n- [Research Brief Template](/docs/research-brief-template) — How to document secondary research findings and define primary research scope\n- [How to Analyze Qualitative Data](/docs/how-to-analyze-qualitative-data) — Systematic analysis frameworks for making sense of primary research data\n- [Generative Research Guide](/docs/generative-research-guide) — When and how to use primary generative research to surface unarticulated user needs\n\n\n## Further reading on the blog\n\n- [B2B Customer Research: The Complete Guide for Product Teams (2026)](/blog/b2b-customer-research-guide-2026) — B2B customer research is harder than B2C — you are navigating buying groups of 10+ stakeholders, gatekeepers, and enterprise procurement cyc\n- [Customer Research Done Right: A Complete Guide for Product Teams](/blog/customer-research-done-right-a-complete-guide-for-product-teams) — Customer research is the foundation of every successful product decision. Learn the types, methods, and best practices that help product tea\n- [Concept Testing: The Complete Guide for Product Teams (2026)](/blog/concept-testing-guide-2026) — Concept testing validates whether your idea is worth building before you build it. This guide covers methods, question templates, analysis a\n\n<!-- further-reading:blog -->\n","category":"Research Methods","lastModified":"2026-05-13T00:26:36.807295+00:00","metaTitle":"Secondary Research: The Complete Guide to Desk Research for Product Teams","metaDescription":"Learn how to conduct secondary research (desk research) before user interviews — including internal and external sources, a 5-step process, when primary research is essential, and how to combine both methods for faster, more original insights.","keywords":["secondary research","desk research","secondary research methods","primary vs secondary research","what is secondary research","desk research UX","secondary research before user interviews"],"aiSummary":"Secondary research (desk research) is the practice of extracting insights from existing data sources before conducting primary research. A 5-step process: define the research question, identify high-signal sources, extract systematically, categorize findings as well-established/contested/unknown, and write a synthesis summary. Internal sources include analytics, support tickets, and previous research. External sources include industry reports, competitor reviews, academic papers, and online communities. Secondary research cannot replace primary research — it lags behind reality, covers market averages not specific users, and cannot surface unarticulated needs. AI-moderated interview platforms like Koji have eliminated the speed advantage that secondary research historically held over primary research.","aiPrerequisites":["No prior research experience required — this is a starting-point guide"],"aiLearningOutcomes":["Understand the difference between secondary and primary research","Identify the best secondary research sources for product and UX questions","Apply a 5-step desk research process to any new research project","Recognize when secondary research is not sufficient and primary research is required","Integrate secondary and primary research into a fast, repeatable discovery workflow"],"aiDifficulty":"beginner","aiEstimatedTime":"12 minutes"}],"pagination":{"total":1,"returned":1,"offset":0}}