Nobody was there

I spent nine minutes talking to a small dark circle at the top of a laptop.

That was the first of four interviews for a senior role. Not a video call with someone on the other end. A camera, a countdown, and three questions I hadn't seen until that day.

Here's how it works. The invitation arrives by email with a deadline attached, and the questions aren't in it. You click to start the process, and only then do you see them. Three questions, three minutes for each answer. You record yourself, you get one retake per question, and then you submit. Nobody watches live. Whoever eventually watches does so at a time that suits them, quite possibly at one and a half speed.

In theory you could open the process, read the questions, close it again and take a few days to think. With a deadline sitting over it, I didn't. I read them, worked out what each one was really asking, picked my examples and started recording. I could have had a fortnight with those questions and it wouldn't have changed much. Knowing what to say was never the difficulty.

The eye line goes first. Your own image sits on the screen below the lens, so looking at yourself means looking down. Look at the lens instead and you have no idea how you're coming across. You choose between appearing engaged and knowing what you look like.

Then the clock. These weren't warm-up questions. Ask me to take you through the most significant piece of organisational design I've led and we'd normally spend twenty minutes on it, because you'd want the business case, who resisted it, what changed afterwards and what I'd do differently now. Three minutes buys the situation, what I did and the result. That's the whole budget. It leaves nothing for the judgement inside any of it, and it removes the follow-up question, which is where most of what I'm any good at actually shows up. Nobody is there to say "and how did that land with the partners?"

So you end up in situation, action, result, because nothing else fits. That's the shape assessors are trained to look for in graduate hiring, and it works well there. At senior level I think it flattens you. What separates people at this level is the messy part: the trade-off you took knowingly, the option you rejected, the thing you got wrong and put right six months later. All of that lives in the middle of the story, and the middle is the first thing to go when you're watching a clock. What you submit is the version of your own work that would sit comfortably on a CV, which is the version they've already read.

The retake is a strange gift. You can watch your first attempt back before deciding, which is an ordeal of its own. On one question I lost my thread and waffled, so using it was easy to justify. What came out second time was sharper and had more in it than three minutes could hold. You spend your one retake fixing the first problem and arrive at a different one.

Here's the part I can't get past. I didn't read from a script. I wanted to sound like someone having a conversation, so I kept my eyes on the lens and talked. I doubt everybody does. You can't read off a second screen without it showing, but nobody needs to. Bullet points stuck just under the lens, four practice runs against a stopwatch, every answer trimmed to land at two minutes fifty. None of that is cheating and all of it is available to anyone who treats the exercise as a recording rather than an interview. That candidate lands on time every time, with no hesitation and no lost thread, and on any structured rubric they score better than I did.

Which means the format can reward preparation as performance. It measures how well you deliver under conditions nobody actually works in, and it favours whoever treats it as a recording rather than an interview. If you run one of these, that's the group you're advancing, and they won't show up in your metrics as a problem.

The stage isn't pass or fail, either. Benchmarks published by the vendors themselves put a healthy conversion rate for one-way video screening at 18 to 25 per cent of submissions advancing, and treat anything above 35 per cent as a sign the stage isn't screening hard enough. One vendor case study describes 68 candidates going forward out of 340. Three in four are cut here, deliberately. You aren't clearing a bar, you're being ranked against everyone else who recorded themselves that week, on a rubric you will never see. Being perfectly good isn't enough when they're taking one in five, and nothing in the invitation tells you that's the game.

I had the questions, I had a retake, and I still had no real idea what I was being measured on. Structure? Clarity? Whether I used the full three minutes or came in under? I don't even know whether a person watched the recordings. In a conversation you pick up signals as you go and adjust. Here you record into silence.

I didn't progress, and I won't know why. That happens.

None of which makes the tool worthless. For volume hiring it earns its place easily. Graduate intakes, entry-level roles, customer service teams, where you're screening hundreds of people against much the same criteria and the realistic alternative is nobody looking at them properly at all. Everyone gets the same questions under the same conditions, and that consistency is worth something.

Every firm I've worked in has used something like this, and used it exactly there. What surprised me was meeting it at senior manager level, as the first of four stages, before anyone on the hiring side had spoken to me at all.

If you own one of these processes, here's the part your candidates won't tell you. They still want the job, so they record their answers, submit them, thank the recruiter and say nothing about having no idea what you were looking for. The silence isn't satisfaction.

One change would fix most of it and it costs nothing. Tell candidates what good looks like. Say whether you want one example or three, the strategic view or the mechanics, and roughly what depth three minutes is meant to buy you. Everyone is guessing at that, which means you're partly measuring who guessed the same way you would have.

I'd do another one. But I'd do it knowing what it is: not a conversation, and not really an interview. A filter with a camera pointed at you, which is a reasonable thing for a firm to run, as long as everyone involved is straight about that being what it is.