What Counts as Evidence at a Distance
Sorting what you know into what you saw, what you were told, what is written and what has no source at all.
three things · one reached them · one with no channel at all · the room side is not counted
The three questions a promotion panel will ask.
Everything a manager believes about somebody came from one of four places: they saw it, somebody told them, it is written down, or it has no source at all. Sorting your beliefs into those four is the single most clarifying thing available before any decision about a person.
The evidence problem in “What Counts as Evidence at a Distance” is easier to discuss when the underlying record can be inspected and corrected. A nonprofit reviewing employee monitoring software for employee monitoring software should use time and project context as one input, explain the collection to staff and volunteers, and keep consequential judgements with an accountable manager.
The fourth category is larger than anybody expects and it is where the damage happens.
For a separate nonprofit perspective on “What Counts as Evidence at a Distance”, see National Council of Nonprofits resources. Use the outside material to test governance, inclusion and implementation questions against the organisation’s real circumstances.
The four sources
Seen: you observed it yourself. Told: a third party reported it. Written: it exists as an artefact somebody else could examine. None: you believe it and cannot say how you came to.
Every belief falls into exactly one.
How to do the sort
Write down what you think about the person, one line per belief, then mark each with its source. Do it before any assessment conversation and keep it to yourself.
It takes fifteen minutes and it is uncomfortable, which is the usual sign of a useful exercise. Most managers find that two or three of their firmest convictions about somebody are in the fourth category, and that the convictions in the fourth category are disproportionately about character rather than about work.
The four sources, and what distance does to each:
| Source | Available in the room | Available at a distance |
|---|---|---|
| Seen | Continuously | Not at all |
| Told | Frequently, from several people | Occasionally, from fewer |
| Written | Rarely needed | Fully, and unaffected |
| None | Accumulates quietly | Accumulates faster |
Why "seen" is weaker than it feels
An observation is one sample, taken at a moment, interpreted by somebody with an existing view. It feels like fact because you were there.
It is still worth more than anything else about somebody you sit with, and it is simply unavailable for somebody you do not.
The structural impossibility
You cannot have seen the work of somebody who works somewhere else. It is not a matter of attention.
Which means for anybody outside the room, the strongest category is empty by construction, and the case has to be built entirely from the other three.
Why "told" needs a name attached
Second-hand reports are legitimate evidence and they are only legitimate when attributable. "People have been saying" is not evidence; "Priya mentioned twice that handovers from him arrive incomplete" is.
The difference matters because the second can be checked and the first cannot. Unattributed reports also have a reliable property: they travel further and faster the less flattering they are, and by the time they reach a manager nobody can say who started them.
Why "written" is the strongest category at a distance
Because it survives without a witness. A document, a piece of analysis, a design, a set of delivery dates: all of these can be examined by anybody, including by somebody disputing the assessment.
This is the one category where distance costs nothing.
The work of reading the work
Reading somebody's actual output carefully takes a couple of hours per person per quarter, and almost no manager does it.
It is the whole remedy. Everything else in this section is commentary on the fact that reading somebody's work properly is available, cheap, and skipped.
The fourth category and what to do with it
Delete it. A belief with no traceable source has no place in any assessment, any conversation about somebody's future, or any written record.
That includes beliefs that turn out to be correct.
Why deleting it is hard
Because it does not feel like a guess. Impressions formed over months feel like knowledge, and some of them are, and there is no way to tell which from the inside.
The rule has to be mechanical rather than discretionary.
Converting a fourth-category belief
Where you believe something and cannot source it, the honest move is to go and look. If you think somebody has disengaged, read three months of their output and ask two people who work with them.
Either you now have evidence or you have discovered that you had none.
Saying "I do not know"
A manager who says they do not have a view on something about somebody they rarely see is being accurate. Almost nobody says it, because it sounds like an admission of failure.
It is a far better answer than an impression dressed as a judgement.
What this sorting cannot supply
Material that does not exist. If somebody's work produces no artefact and nobody who works with them can speak to it, the sorting exercise will correctly report that you have nothing, and that will be the end of it.
That outcome is worth reaching early, because it is a finding about the role rather than about the person: a job whose output leaves no trace and whose practice nobody witnesses cannot be assessed fairly at a distance by anybody. Fixing that means changing what the job produces, which is a real conversation and belongs well before the assessment cycle rather than during it.