Out of 5.2 million candidate interviews, 31.2% had a scorecard attached. That’s the industry’s normal state, and it’s the wider of the two problems in this data: most candidate interviews never get a scorecard at all.
The second problem is submission, and it sits on a different denominator. Of the scorecards that did get created, 32% were complete and 26% were blank,¹ and whether one got submitted at all depended heavily on how it got started.
Interviewers submitted 28.6% of the scorecards they wrote from scratch, against 50.3% of the ones Metaview drafted first from the conversation. That’s 1.76x as many. The drafted ones came in fuller too, at 7.85 completed fields on average to 2.61 written by hand.¹
This is an observational comparison rather than a controlled experiment. Teams that let a draft start the scorecard may run tighter processes in other ways, so we can’t isolate the effect of the draft on its own. Still, the size and direction of the gap are consistent with a draft easing the writing burden, and writing is the part of the job most likely to get skipped.
Why scorecards get left behind
Scorecards get filled in after the interview, usually once the interviewer has moved on to the next thing. Details fade. Writing from memory takes longer, and the feedback usually comes out thinner.
Reminders prompt interviewers to complete the scorecard, but they don't make it easier
A reminder can push someone to fill the scorecard in, but the interviewer still has to reconstruct the conversation from whatever notes they took and whatever they remember. Missing feedback costs you later, too: debriefs have less to work with, and any later analysis of the hiring process is weaker when the interview record is incomplete.
What changes when Metaview drafts first
A first draft turns the job from writing into reviewing. The interviewer starts from a draft of what was said instead of a blank form. They check the evidence, fix what’s wrong, add their judgment, and submit.
A first draft makes the task much easier
A draft replaces the blank scorecard, but it doesn’t finish the job. Even with a draft, roughly half are never submitted, and the team still has to review, correct, and submit each one.
- Submitted 28.6% of the time
- 2.61 fields completed on average
- Written later, from notes and memory
- Submitted 50.3% of the time
- 7.85 fields completed on average
- Reviewed and judged by the interviewer
The two groups weren’t randomly assigned. Teams that use drafts may also run tighter processes or hold interviewers to clearer expectations. The comparison shows a strong association. It doesn’t prove the draft explains the whole difference.
How the workflow works
Metaview’s AI Notetaker records the interview with consent, then drafts the scorecard from what was said. Each point links back to the moment it came from. The interviewer keeps every judgment call.
- Check the draft: confirm each point against the evidence, fix anything wrong, and cut what isn’t relevant.
- Own the rating: the interviewer chooses the rating and the recommendation. Metaview doesn’t score the candidate.
Different stages need different scorecard templates. A recruiter screen, a technical interview, and a final panel aren’t looking for the same things, so their scorecards shouldn’t read the same either.
What a completed scorecard does not tell you
A filled scorecard is easier to act on than a blank one, but a complete scorecard can still be a bad one: vague, inaccurate, or waved through without a real read of the evidence.
So treat completion as the first check, and only that. Before trusting the record, ask a few harder questions. Is the evidence specific? Did the interviewer actually edit the drafted fields? Does each rating trace back to something the candidate said? Those answers matter more than the field count.
What to measure
A single completion rate hides where the process is failing. Track each stage on its own so you can see whether the gap is missing scorecards, unfinished ones, slow feedback, or thin evidence.
Start with the last 90 days
Keep the measures separate, because each one has a different denominator:
- Coverage: of all candidate interviews, how many had any scorecard attached?
- Submission: of the scorecards created or drafted, how many were submitted?
- Completeness: of the submitted scorecards, how many were complete, partial, or blank?
- Feedback time: how long after the interview was the scorecard submitted?
- Draft review: did the interviewer review and edit the drafted fields before submitting?
- Evidence quality: does each rating cite specific, job-relevant evidence that can be checked against the conversation?
Don't collapse these into one completion rate
A team can have good coverage and poor submission, or high submission and weak evidence. Each combination needs its own fix. And a better submission rate among the scorecards that exist won’t rescue the interviews that never had one.
The two numbers hand you two different jobs. Coverage is the wider gap, and most candidate interviews still end with no scorecard at all. For submission, a first draft from the conversation is the clearest lever in this data, and it leaves the rating with the interviewer. Coverage is closer to an adoption problem, and it deserves its own plan. Measure each stage on its own denominator, and start wherever your numbers look worst.
See your real scorecard completion rate.
Metaview Reports tracks scorecard coverage, submission, and feedback time across your team.
Frequently asked questions
How much higher was submission when Metaview drafted first?
Submission was 50.3%, compared with 28.6% for scorecards written from scratch. That’s 1.76x as high, though the comparison is observational and doesn’t isolate the effect of the draft.
Does Metaview score the candidate?
No. Metaview drafts the scorecard from the conversation. The interviewer reviews the evidence, makes the rating, and submits it.
Do more completed fields mean better feedback?
Not by themselves. A full scorecard can still be vague, inaccurate, or accepted without enough review. Check whether the evidence is specific, whether the draft was reviewed, and whether each rating can be traced back to the interview.
Why are the coverage and submission figures different?
They’re measured on different denominators. Coverage is the share of all candidate interviews that have any scorecard attached. Submission is the share of created scorecards that were submitted, and it was higher when Metaview drafted the scorecard first. A team can have weak coverage and reasonable submission at the same time, so the two are tracked separately.
How should we measure scorecard completion?
Track coverage, submission, and completeness separately. Add feedback time, draft review, and a quick check of evidence quality. That’ll show whether the problem is missing scorecards, unfinished scorecards, or weak feedback.
¹ Source: Metaview’s corpus of 5.5 million captured conversations (5.2 million candidate interviews), 2026. The 31.2% figure counts candidate interviews with any scorecard attached, across the corpus as a whole rather than as a measure of any one product’s performance. Submission rates compare 26,498 scorecards written by interviewers with 93,502 scorecards where Metaview generated the first draft. Field counts and completeness figures come from platform samples. All findings are aggregate.
