The One-Way Video Interview, From Invite Link to Submitted Recordings

A one-way video interview is an asynchronous, pre-recorded screen: you read a fixed set of prompts, record each answer under a timer, and submit the set before a deadline, with no interviewer on the other end. This guide is for a candidate who has just received an invite link and wants to run the format end to end without improvising into a countdown. It gives you the procedure keyed to the four constraints that actually decide the outcome: a preparation window, a hard recording cut, a fixed retake budget, and a review that is usually a sped-up skim or a transcript score.

Most ranking pages for this format are question listicles. They tell you the 25 questions you might be asked and ignore that you answer them against a 30-second prep clock, a cap that cuts you off mid-sentence, and often zero retakes. The questions are the least constrained part of the job. What follows is the executable version, built so you can rehearse to the clock instead of discovering it live.

What is a one-way video interview and what are its constraints?

A one-way video interview is a recorded assessment where you answer preset prompts under per-question timers and submit the recordings, with no live interviewer. The format is defined by four numbers you must read before you start: the prep window, the recording cap, the question count, and the retake budget.

These numbers vary by platform and by employer settings, so treat any single source as a range rather than a rule. The documented norms converge but do not agree exactly, which is why reading your own spec screen is step one.

The spread is real. A HireVue walkthrough states 30 seconds to prepare and up to 3 minutes to answer. The largest documented norm set, drawn from 2,550,105 recorded responses by 627,999 candidates, found most interviews had four or five questions, 30 seconds of prep, two minutes to answer, and rarely permitted re-recording. One reported BlackRock configuration ran roughly three minutes of prep, a 90-second answer, about three questions, a five-day deadline, and no retakes. Your invite will tell you which of these worlds you are in. Do not assume.

How are one-way recordings actually reviewed?

They are reviewed in two layers: a human skim, often at 1.5x or 2x playback speed, and, where enabled, a transcript-based automated score. Neither layer rewards a slow build to a point. Both reward the answer stated first.

Recruiters and hiring managers watch or skim recordings at increased playback speed, frequently guided by transcripts. Where HireVue's video assessment scoring is enabled, the company says it analyses the response transcript rather than facial, video, or audio features. The practical consequence is a single discipline: answer the question in the first sentence, then support it. A context-first answer dies under a 1.5x skim and scores weakly on transcript relevance, because the reviewer or model hits the relevant words late or not at all.

This also kills a widespread piece of folklore. HireVue discontinued facial analysis, announced in January 2021, and nonverbal data had contributed only about 0.25% to predictive power in most cases, rising to 4% for customer-facing roles, before being retired. Candidates who over-manage their smile and eye contact are optimizing a signal the system no longer scores.

Because scoring moved to words, the highest-leverage preparation is a clean, direct transcript, not polished facial affect. That is a different rehearsal target than the one most video-interview advice assumes.

How long should each answer be?

Target 60 to 90 seconds and do not fill the cap. Under 45 seconds reads as thin; using every second of a 3-minute cap almost always means rambling, and a reviewer skimming at 2x disengages long before you finish.

The cap is a trap, not a target. On a 2-minute cap, a 90-second answer uses about 75% of the time. On a 3-minute cap, it uses about 50%, and at 2x playback a reviewer hears it in roughly 45 seconds. The limit measures restraint, not stamina. Inside that 60 to 90 seconds, spend your time on the Action, because that is what is being scored, not the setup.

The sources disagree on the exact split, but they agree on the shape: Action is the single largest block in every one, at 50 to 60% of the answer. The Situation and Task together are setup, and setup is where skimmers lose patience. So you can use any of these allocations, but keep Action dominant and keep the whole thing under 90 seconds.

The cap measures restraint, not stamina; a full-cap answer is exactly where a 2x skimmer disengages.

Here is the structure to internalize. Note that the first sentence is the answer itself, before any Situation.

[Direct answer, one sentence]: The clearest example of that is when I [outcome in one line].
Situation (10 to 20%): Briefly, the context was [one sentence].
Task (10 to 15%): My job was to [one sentence].
Action (50 to 60%): So I [the specific moves you made, in order; this is the bulk of the answer].
Result (15 to 20%): The result was [quantified outcome].

Run the interview: the end-to-end procedure

The procedure runs in two phases: preparation across the deadline window, and execution in a single recording session. Build your story bank and rehearse days early, then set up, practice, and record in one focused sitting.

A note on order. Some sources put research and prep before setup, and that is right for the days-ahead work. HireVue's flow puts the practice question only at the session start, so on the day, steps four and five sit adjacent: set up the room, then immediately run the practice question so the first real prompt is not your first contact with the interface.

The story bank is what substitutes for a retake

Prepare 5 to 8 core STAR stories that can answer multiple question types, then record each 8 to 10 times. By repetition seven to ten you internalize the structure while keeping delivery natural. This matters because, in the 2.5-million-response norm set, re-recording was rarely permitted. The retake budget most guides obsess over frequently does not exist, so rehearsal is your retake.

How do the prep countdown and retake budget actually work?

The prep countdown auto-starts the recording when it expires, whether you are ready or not, and the retake budget is set by the employer and shown to you before each question. You cannot negotiate either; you plan around them.

Once the allotted prep time is up, the platform automatically begins recording your answer. That is why the prep window is for selecting a rehearsed story and jotting 2 to 3 bullets, not for composing. Candidates who try to write a fresh answer in 30 seconds freeze and get caught mid-thought when recording starts.

Retakes vary by employer configuration: some allow one or two re-recordings per question, many allow none, and the interface tells you the rules per question before you record. Read that line first. If retakes exist, impose a discipline on yourself: cap your own retakes at the stated limit, and spend them only on genuinely ruined or disrupted takes, never on tone or a slightly awkward phrase. There is research support for the format offering retakes at all, so when they are available, use them deliberately rather than compulsively.

How this goes wrong: failure modes and false positives

Most failed one-way interviews fail on format, not content. The candidate had good stories and lost on the clock, the cap, or the setup. Here are the documented ways it breaks and the check for each.

• Reading the spec wrong. The candidate assumes retakes exist and burns the first prompt casually; many templates allow none. Check: confirm the per-question retake count on the intro screen before question one.
• Filling the cap. A 3-minute answer reads as rambling to a 2x skimmer. The false positive is feeling thorough. Check: time your rehearsals and stop at 90 seconds.
• Burying the answer. Context-first answers die under a 1.5x skim and score weak on transcript relevance. Check: can a reader get your point from sentence one alone?
• Over-rehearsed delivery. Reading a script flattens the transcript and looks robotic; filler-heavy answers also produce messy transcripts that are harder to score. Check: record from bullets, not a word-for-word script.
• Prep-countdown panic. The timer auto-starts recording and the candidate freezes mid-compose. Check: pre-select a story during the prep seconds instead of writing a new answer.
• Chasing cosmetic retakes. A limited budget spent on tone leaves nothing for a genuinely ruined take. Check: re-record only for substance or disruption.
• Believing the facial-analysis myth. The candidate over-manages expressions on pre-2021 lore. The false positive is "I must smile for the AI." Check: scoring is transcript-based now; direct the energy into clear words.
• No backup device. A mid-record tech failure with no fallback ends the session. Check: keep a second device and the interviewer's contact details open before you start.

The abandonment figure is worth sitting with. The highest-leverage move in this whole procedure is not a sharper story. It is testing your camera, microphone, and internet on the actual platform, not just any video app, both the day before and again an hour before, so that when the first prompt loads, setup is already solved.

Who is on the other side of the skim, and why it matters

The reviewer is usually a recruiter or talent acquisition specialist moving fast through a stack of recordings, and in leaner markets they lean even harder on the sped-up skim and transcript ranking. Knowing that shapes how directly you answer.

In Refolk's index of professional profiles, there are 44,125 US profiles titled Technical Recruiter or Talent Acquisition Specialist, against just 3,028 equivalent profiles in the UK, a 14.6x gap when you divide the two counts. Thinner review teams do not watch at 1x. They skim at 1.5x or 2x and rank on transcript, which is exactly why front-loading the answer matters more in a thin market than a thick one.

Those counts and employers come from Refolk's index; the multiple is derived from the two figures. A separate Director-versus-Senior seniority comparison was queried in Refolk's index and returned no matching profiles, so that breakdown is not established here and I will not assert it.

If you want to see who tends to review these assessments at the companies you are targeting, Refolk can surface the recruiters and talent acquisition leads behind a given hiring team from public profiles, so you know whether you are talking to a large campus team or a single early-careers recruiter screening at 2x.

Refolk also writes and tailors the resume that gets you to the video stage in the first place, scoring how well your history fits each posting, which is the upstream half of the same job this guide finishes.

Final checks before you submit

Before you call the interview done, verify the recording session against the format's real constraints, not just whether you spoke well. Run this list after your last take and before the submit confirmation.

Keeping this current

Two things in this procedure drift over time, so re-check them per invite rather than trusting a remembered value. First, the platform's numbers: prep window, cap, questions, and retakes are employer-configured and change between roles, so always read the fresh spec screen. Second, the scoring model: HireVue retired facial analysis and moved to transcript scoring, but platforms evolve, so if a platform's current candidate FAQ describes what it analyses, read it and let that, not old lore, set your rehearsal target. The constant across all of it is the discipline: answer first, keep it under 90 seconds, treat the first take as the real one, and solve setup a day early.