{"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":"2026-05-18T12:55:35.185Z"},"content":[{"type":"documentation","id":"d59e13c6-bd74-486d-834c-e819031dd6a0","slug":"interview-not-counting","title":"Interview Not Counting","url":"https://www.koji.so/docs/interview-not-counting","summary":"Explains why some Koji interviews do not count toward usage. The quality gate requires a score of 3+ on a 1-5 scale. Covers scoring dimensions (relevance, depth, coverage, completion, structured quality), what happens to low-quality interviews, and tips for improving interview quality.","content":"# Interview Not Counting\n\nIf you see an interview in your project but it is not counting toward your usage or appearing in your analysis, it most likely did not meet the quality threshold. This is intentional -- Koji's quality gate ensures that only meaningful interviews contribute to your research insights.\n\n---\n\n## Why Interviews Do Not Count\n\nKoji assigns a quality score to every completed interview. Interviews that score below a threshold of **3 on a scale of 1 to 5** are considered low-quality and are excluded from your usage count and aggregate analysis.\n\nThis is a **feature, not a bug.** The quality gate protects your research data from being diluted by interviews where the respondent did not engage meaningfully.\n\n---\n\n## What Causes a Low Quality Score\n\nSeveral factors can result in a quality score below 3:\n\n### Minimal Responses\n\nThe respondent gave very short, single-word, or one-sentence answers throughout the interview. Meaningful qualitative research requires detailed, thoughtful responses. An interview where every answer is \"yes\", \"no\", or \"fine\" does not provide useful insight.\n\n### Off-Topic Responses\n\nThe respondent's answers did not relate to the interview topics. This can happen when someone is not genuinely interested in providing feedback and clicks through the interview to complete it quickly.\n\n### Insufficient Length\n\nThe interview ended very early -- either the respondent left after just a few exchanges or was disconnected. A very short interview does not have enough content to generate meaningful analysis.\n\n### Gibberish or Spam\n\nThe respondent entered random text, keyboard mashing, or repetitive nonsensical content. Koji's quality assessment detects these patterns and scores the interview accordingly.\n\n### Copy-Pasted Content\n\nResponses that appear to be copy-pasted from external sources rather than genuine answers to the interview questions.\n\n---\n\n## How to Check an Interview's Quality Score\n\nYou can view the quality score for any completed interview:\n\n1. Open the interview from your project's interview list.\n2. Look for the **Quality Score** indicator, typically displayed as a numeric score or a visual indicator near the interview header.\n3. Interviews below the quality threshold are marked accordingly.\n\nThe quality score is also available via the API when you [retrieve interview data](/docs/completing-interviews-via-api).\n\n---\n\n## What Happens to Low-Quality Interviews\n\nInterviews that do not meet the quality threshold:\n\n- **Are not counted** toward your plan's interview usage or billed credits.\n- **Are excluded from aggregate analysis** -- they do not influence theme extraction, insight generation, or project-level reports.\n- **Are still accessible.** You can view the transcript and the quality score. Nothing is deleted.\n- **Can be reviewed.** You might find the transcript useful for other purposes even if it did not meet the quality bar.\n\nThis means you are never \"charged\" for a low-quality interview. Your interview credits are preserved for interviews that actually contribute to your research.\n\n---\n\n## Improving Interview Quality\n\nIf you are seeing a high rate of low-quality interviews, consider these adjustments:\n\n### Improve Your Interview Guide\n\nA well-written interview guide elicits better responses. See [How the Quality Gate Works](/docs/how-the-quality-gate-works) for detailed guidance on writing guides that encourage thoughtful answers.\n\n- **Ask open-ended questions.** Questions that can be answered with \"yes\" or \"no\" lead to short responses.\n- **Use structured question types.** Scale, ranking, and choice questions from the [structured questions guide](/docs/structured-questions-guide) give respondents clear interaction points.\n- **Provide context.** Help respondents understand why their detailed feedback matters.\n- **Set expectations.** Let respondents know upfront that the interview takes a certain amount of time.\n\n### Screen Your Respondents\n\nIf you are recruiting respondents externally, quality issues may indicate a recruitment problem:\n\n- **Use screening criteria.** Filter for respondents who have relevant experience with your product or topic.\n- **Verify engagement.** Look for respondent panels or recruitment methods that prioritize quality over volume.\n- **Set expectations.** Tell respondents what the interview involves and how long it takes before they start.\n\n### Use the Right Interview Mode\n\nSome respondents express themselves better verbally. If you are seeing short text responses, consider enabling voice mode. Conversely, some respondents prefer typing and may give more detailed written answers.\n\n### Check Your Distribution Method\n\nHow respondents arrive at the interview matters:\n\n- **Personal invitations** tend to yield higher quality than anonymous open links.\n- **Context-specific triggers** (like a post-purchase survey) capture respondents when they have relevant recent experience.\n- **Incentive alignment** -- make sure any incentives you offer do not inadvertently attract respondents who rush through without engaging.\n\n---\n\n## [Understanding Quality Scores](/docs/understanding-quality-scores)\n\nThe quality score reflects multiple dimensions of the interview:\n\n| Factor | What It Measures |\n|---|---|\n| **Relevance** | How well the responses relate to the interview questions and research topic |\n| **Depth** | The level of detail and thoughtfulness in the responses |\n| **Coverage** | How many of the key interview topics were addressed |\n| **Completion** | Whether the interview progressed through the full conversation flow |\n| **Structured Quality** | The quality and consistency of responses to structured questions (scales, choices, rankings) |\n\nEach factor is scored individually, and the composite score determines whether the interview meets the threshold. An interview can score well on some factors and poorly on others -- the overall score reflects the combined assessment.\n\nFor a deeper dive into how scoring works, see [How the Quality Gate Works](/docs/how-the-quality-gate-works).\n\n---\n\n## Reports and Quality Filtering\n\nReports also filter by the quality gate. Only interviews scoring 3 or above are included in report generation. This ensures your research reports reflect high-quality data and are not skewed by low-effort responses.\n\nSee [Understanding Usage Limits](/docs/understanding-usage-limits) for more about how quality gating interacts with your credit usage.\n\n---\n\n## Frequently Asked Questions\n\n### Can I override the quality gate?\n\nNo. The quality gate is applied automatically and cannot be disabled. This is a deliberate design choice to protect research integrity. Low-quality interviews are still visible in your project -- they just do not count toward usage or appear in aggregate analysis.\n\n### Does the quality gate affect billing?\n\nYes. Interviews that score below 3 are **not billed**. You are only charged credits for interviews that meet the quality threshold. This protects you from paying for interviews that do not contribute meaningful research data.\n\n### Can I see why a specific interview scored low?\n\nYes. Open the interview and check the quality breakdown. Each scoring dimension (relevance, depth, coverage, completion, structured quality) has its own score, so you can see which aspect pulled the overall score down.\n\n### What if a legitimate interview scored below 3?\n\nThis is rare but can happen if the respondent gave unusually brief responses despite being engaged. The threshold is set to balance between filtering noise and retaining useful data. If you believe an interview was unfairly scored, the transcript is still available for manual review.\n\n## Further reading on the blog\n\n- [Koji vs Listen Labs: AI Interview Platforms Compared (2026)](/blog/koji-vs-listen-labs-2026) — Listen Labs raised $69M and powers enterprise research at brands across the Fortune 500. Koji is the accessible AI-native alternative starti\n- [Koji vs Looppanel: End-to-End AI Research vs Analysis-Only (2026)](/blog/koji-vs-looppanel-2026) — Looppanel is a great AI analysis tool — but it only handles half the workflow. You still need a separate platform to actually run interviews\n\n<!-- further-reading:blog -->\n","category":"Troubleshooting","lastModified":"2026-05-13T03:20:03.689176+00:00","metaTitle":"Interview Not Counting — Koji Docs","metaDescription":"Learn why interviews score below the quality threshold and do not count toward your Koji usage.","keywords":["interview not counting","quality score","quality gate","low quality interview","interview excluded"],"aiSummary":"Explains why some Koji interviews do not count toward usage. The quality gate requires a score of 3+ on a 1-5 scale. Covers scoring dimensions (relevance, depth, coverage, completion, structured quality), what happens to low-quality interviews, and tips for improving interview quality.","aiPrerequisites":["how-the-quality-gate-works"],"aiLearningOutcomes":["Understand why an interview does not count","Check quality scores","Improve interview quality","Adjust guides and recruitment to reduce low-quality rates"],"aiDifficulty":"beginner","aiEstimatedTime":"6 min read"}],"pagination":{"total":1,"returned":1,"offset":0}}