The form went out on a Monday. Name three people who can speak to your work this cycle. Due Friday.
I filled it in on Thursday night, in about ten minutes, with Slack open in the other tab. I picked the three names I could think of. That is the honest description of my process.
All three of them wrote something nice. None of it was worth much.
Not because they were lazy. Because the most recent thing any of them had actually watched me do was small. One wrote about a code review I gave. One wrote that I'm easy to work with. The third wrote four sentences that could have been about anyone on the team who isn't actively on fire.
The person who should have been on that list sat in a different org. We spent three weeks in February untangling a schema migration that would have quietly broken two of his services. Four calls, one of which got heated, and at the end we shipped a plan neither team hated. He would have written the paragraph that mattered, because he was the only person outside my team who had watched me do the thing my rubric calls cross-org influence.
I didn't leave him off because I forgot he existed. I left him off because it was October, and our project was back in February.
That box is one of the very few parts of a performance review I completely control.
My manager writes what he writes. Calibration happens in a room I'm not in. The rating comes back with reasons attached that were settled somewhere else. But nobody chooses my peer reviewers except me, and I was choosing them based on who I happened to have talked to that month.
It gets worse when you look at what recall is actually selecting for. The people I talk to most are the people I work with daily and share a manager with. Their testimony carries the least weight, because of course your own team likes you. The word that actually moves a promotion case comes from two orgs away, and those are exactly the people I work with in bursts, months apart, on the hardest problems. The work most worth vouching for gets done with the people I am least likely to think of.
It isn't really a popularity question either. Different rubric rows need different witnesses. The person who watched me handle an incident at 2am can't say whether I mentor anyone. The person I mentor can't say how I handle an incident. I was picking three names as though any three would cover the whole rubric, when what I needed was one name per thing I was trying to prove.
And I do this to other people constantly. A request lands in my inbox and I sit there trying to remember what the two of us actually did together. Usually what comes back is the feeling that it went well and none of the specifics. So I write something warm and general, submit it, and move on. I have certainly been the useless third reviewer in somebody else's packet.
Everyone is doing this to everyone, and then we all wonder why the feedback comes out thin.
I brought it up with a friend who had just gone up for staff. She told me she keeps a list of names in her notes app and adds to it whenever somebody watches her do something hard. I asked how long she had been doing that. Nine years. She started after the first time she went up and didn't get it.
That isn't a personality trait. That is a workaround for a tool that doesn't exist.
So we built it into Perfolio.
When you capture a piece of work, it captures who was there with you. Not a contact list, a record of presence: this person was on that migration, in that design review, on that incident at 2am. When the nomination form shows up, you can start from the rubric row you're trying to cover and see who actually watched you cover it, with the artifact still attached so you can remind them what it was. It also notices when someone who used to appear in your record hasn't shown up in six months, which is usually the last moment reaching out still feels normal.
None of that writes the feedback for you, and it shouldn't. Your reviewer still says whatever they want to say. The only thing that changes is that you're choosing from what actually happened instead of from what you can dredge up on a Thursday night.
If you fill that box from memory, you aren't nominating the people who saw your best work. You're nominating the people you saw last.