River
Y CombinatorBacked by Y Combinator
FREE TEMPLATE

Interview Process and Scorecard Template

Five documents and three sheets, including an assignment sheet where every competency carries a weight and exactly one interviewer's name.

Free download  ·  No account needed

Competency to Interviewer Assignment, one row

One competency, one weight, one name, and the minutes it is allowed

Weights fixed and dated before the first candidate is booked. Nothing on this row is edited after the panel meets.

From Register Row

The numbered requirement it came from. A competency with no row behind it does not enter the loop, and a competency that belongs in the loop but is not on the register sends you upstream to fix the register first.

Weight

A number, and all the weights sum to 100. This is the column that turns a list of competencies into a decision rule, and it is the one every blank scorecard leaves out.

Owner

One interviewer, by name. They prepare the test, run it, and write the rating with its evidence. A second name is a second read: allowed once, on the heaviest competency, written on the row with its reason and taken blind.

How It Is Tested

The specific thing that settles it. A work sample with a seeded fault. A ten-minute live walkthrough with four planted questions. A graded 200-word status note. Whatever it is, somebody has to build it before the loop opens, and that somebody is the owner.

Minutes

Divided by weight, this is the check that catches a five-point competency booked for a full hour.

Reads And Minutes In The Audited Loop

The two columns that make the redesign land. What the loop you run today already spends on this competency, from the interviewers' own guides, sitting next to what it is worth.

A hiring loop is a budget nobody writes down. Four interviewers, an hour each by default, and whatever they happen to cover is what the decision rests on. Unassigned, that budget converges on what a generalist can probe with no preparation, so the competencies needing a test built in advance never get run. This pack weights every competency and puts exactly one interviewer's name on each, which is the relative importance a job analysis is expected to establish rather than a house preference.

In the worked example, Marloe Logistics had a Senior Implementation Consultant requisition open 21 weeks, and its four panel interviews came to 210 minutes per candidate. Eighty-five of those minutes went to written communication, read by all four interviewers, against a competency worth 20 of 100 weight points. The reconciliation work sample, worth 25 points, was never run: it needed two prepared extracts and nobody owned preparing them. Forty points had no evidence behind them, and every competency traced to a row on that role's requirement register.

That gap decides hires. On the five competencies the old loop could see, the panel's favourite led the runner-up 93.3 to 81.7. Reading the sixth inverted the order, 86.25 to 82.5, and the sixth was the work sample carrying a quarter of the decision. Written for recruiters and hiring managers whose requisition has been open five months. If the loop has already run, debrief synthesis works from the scorecards you have, and recruiting funnel analysis says whether the loop is the constraint at all.

Forty weight points nobody tested, and the hire it would have cost

The Competency to Interviewer Assignment, the Scorecard it feeds, and the Loop Feedback Log.

Competency to Interviewer Assignment

Illustrative for a fictional warehouse software company, Marloe Logistics. Senior Implementation Consultant, 21 weeks open. Weights fixed 4 February.

CompetencyRowWtReads
audited
Mins
audited
Min per
weight pt
Owner nowMins
now
C1 Owns two go-lives end to endMH 325362.52.50Ferreira35
C2 Reconciles two extracts unaidedMH 425000.00Tan50
C3 Writes a VP-readable status noteMH 820485.04.25Graded sample0
C4 Trains twenty warehouse staffMH 1115000.00Haddad30
C5 Debugs an EDI 940 or 945 failureNH 510242.54.25Berg20
C6 Names an integration's real costNH 125120.04.00Tan10
Totals10010210.0all owned160
Settled by a document, no loop minutes: MH 7 travel figure, MH 9 named prior sector.

Forty of the 100 weight points had no read at all, and both belong to competencies whose test has to be prepared in advance. C3 took 40.5 percent of the loop for 20 points. Across the loop, 126.7 of 210 minutes went to ground somebody else had already covered, which over nine panel candidates is 19.0 of 31.5 interviewer-hours.

Scorecard

Anchored four-point scale. 4 means they did it, named the specifics, and named what went wrong. Every rating carries the evidence that produced it.

CompetencyWtAdelekeQuinteroHalloran
C1 Owns two go-lives end to end25433
C2 Reconciles two extracts unaided25243
C3 Writes a VP-readable status note20433
C4 Trains twenty warehouse staff15342
C5 Debugs an EDI 940 or 945 failure10433
C6 Names an integration's real cost5343
Weighted, all six100330 of 400
82.5%
345 of 400
86.25%
285 of 400
71.25%
The five the old loop could see7593.3%81.7%70.0%

C2 is the difference and it is worth 25 points. Quintero found the seeded unit-of-measure fault in 18 minutes and traced it to the conversion applied on receipt. Adeleke found it at minute 41 and could not explain the variance without prompting. On everything the old loop was capable of reading, Adeleke wins by 11.6 points.

Loop Feedback Log

One row per stage, so the loop can say which component produced which outcome rather than reconstructing it later.

StageCompetency ownedEnteredAdvancedRateRatings 4/3/2/1
Recruiter screenMH 7 and 9, by document642234.4%not rated
Writing sample, gradedC3221463.6%3/9/8/2
Hiring managerC114964.3%2/6/5/1
Technical work sampleC2 and C69555.6%1/4/3/1
Enablement walkthroughC49666.7%2/4/2/1
Customer successC5, plus the C1 second read9777.8%2/5/2/0
Offerall six weighted3133.3%not rated

Submission

Thirteen of fourteen scorecards arrived before the debrief. The fourteenth landed a day late and its rating moved from 3 to 2 in the room, which is the whole reason the pre-debrief chase exists. The narrowest component is the technical work sample at 55.6 percent, and it is the component the audited loop did not have.

What is in the pack

01

Competency to Interviewer Assignment

Every competency with its register row, its weight, its one owner, its method and its minutes. The audited loop's reads and minutes sit in the columns beside them, which is what gets a redesign agreed to rather than debated.

02

Scorecard

One row per candidate per competency, four written anchors, and an evidence cell on every rating. A rating with no evidence is recorded as missing rather than as a low score, because a missing rating is a coverage hole.

03

Loop Feedback Log

Scorecard submission against the debrief date, plus a per-stage record of who entered, who advanced and how the ratings fell. Where a selection process carries adverse impact, that record is what shows which components carry it.

04

Interview Guides by Stage

One guide per stage, naming the competencies it owns, both rating anchors rather than only the top one, and an explicit list of what this stage does not cover.

05

Panel Brief

The assignment table on one page, plus the rule that a question about somebody else's competency goes into the debrief note instead of into the interview, and the reason it costs you twice.

06

Debrief Format

Forty-five minutes in a fixed order: coverage first, then the disagreements, then the weighted picture. The total is read once and set aside rather than voted on or averaged again.

07

Offer Recommendation

The deciding competency with both ratings, the ways the runner-up was stronger stated with their weights, the gap being accepted with a named owner, and the date the weights were fixed. Pricing the number itself belongs to the offer and pay equity pack, which runs its check before anything is spoken.

08

How a Competency Earns a Slot

The four conditions, including the one that takes work off the panel: a requirement a document already settles gets confirmed once in writing and never spends an interview minute.

How it works

  1. 1

    Send what you have

    The role's requirements, whatever shows what each interviewer already asks, and how many candidates reached the panel. Their own saved guides or last few scorecards are enough, and two interviewers' worth is plenty to start.

  2. 2

    The audit runs first

    Each interviewer's booked minutes get split across whatever their material actually covers. Out comes weight points read, weight points missed by name, minutes per weight point, and the minutes spent on ground already covered.

  3. 3

    One name per competency

    Weights are set and dated, one owner per competency, at most one second read on the heaviest and taken blind. Anything a document already settles leaves the panel entirely.

  4. 4

    The loop runs and records itself

    Guides, brief and scorecard are all written from the assignment sheet, and every stage logs what it tested. The debrief then argues about specific evidence instead of averaging four impressions.

Frequently asked questions

Why not just download an interview scorecard template?

Because a blank scorecard is the thing that already exists everywhere. Every free version hands the same form to four interviewers, which is precisely how four people end up rating the same competency. The assignment sheet is the missing half: which competency each interviewer owns, what it is worth, decided before anybody is booked.

How do you set the weights without guessing?

From what the role does most and what failure costs most, one line of reasoning per competency, then dated. The check is minutes divided by weight. If a competency worth five points is booked for a full hour, the weights and the schedule disagree and one of them is wrong.

Four interviewers all want to ask about the same thing. Now what?

Show them the audit. In the worked example the panel had read written communication four times and the reconciliation work sample zero times, and the work sample carried 25 of 100 weight points. That comparison moves people; an argument about rigour does not. Once each competency has an owner, the behavioral question bank scores the questions filling those slots.

Does one owner per competency mean no second opinions?

There is one second read, budgeted on the heaviest competency, written on the row with its reason, and taken blind so the second reader never sees the first rating. The debrief then has one specific disagreement to work through rather than four opinions formed about the same answer.

What if the role has no requirement register behind it?

Competencies get derived from the job description instead, and the pack says what that costs: a requirement with no test behind it produces a competency with no anchor. Settling it on the job description first is cheaper than rating six candidates against something nobody can score.

What format are the downloaded files?

Five Word documents and three CSV sheets in one zip, with no conversion needed. The documents open in Word, Pages and Google Docs. The sheets open in Excel, Numbers and Google Sheets. A PDF version is available for reading and circulating rather than filling in.

Is it free, and what does Edit with AI add?

The download is free and needs no account. Edit with AI installs the same pack as a private Space and runs the audit against your own material: your interviewers' guides, your requirements, your panel counts. Where the candidates arrive through external recruiters, the agency register measures what those submissions are worth once the loop has run.

Find out what your loop is actually testing

Send the role's requirements and whatever shows what each interviewer already asks. The first thing back is the weight points nobody is reading, named, and what they cost in interviewer-hours.

Edit with AI