A position opens. Within a week three people have suggested somebody, one of them is a neighbour, and the owner already has a view about two of the likely applicants because they have watched them work elsewhere for a decade.
This is a considerably better information environment than any structured process produces, and it carries a specific set of hazards.
The information advantage is real
An employer in a small community can find out things about a candidate that no interview surfaces.
Whether they finish what they start. How they behaved when something went wrong. Whether they are the same person at seven in the morning as at four in the afternoon. Whether their previous employer would take them back.
That is behavioural evidence accumulated over years, which is exactly the material the capability entry identifies as predictive and which structured hiring processes struggle to obtain.
An employer who declines to use it, on the reasoning that it is not a formal source, has discarded their principal advantage.
The same network distorts
Three ways, and each operates without anybody intending it.
A reputation formed once persists. Somebody who handled something badly at twenty-three is still carrying it at thirty-eight, and the account has been repeated by people who were not present. The person who has actually changed cannot outrun it locally, which is one reason capable people leave small places.
The recommendation carries an obligation. A candidate suggested by a significant customer, a board member, or a longstanding friend arrives with a cost attached to declining them, and that cost is being weighed alongside their suitability whether or not the employer acknowledges it.
The pool narrows to the known. Hiring from the network repeatedly produces a workforce drawn from one set of families and connections. Not through any exclusionary intent, and the effect is identical.
Separating the reference from the referrer
The most useful discipline available, and it costs one question.
When somebody recommends a candidate, ask what they have actually seen them do. Not whether they are a good person, which is a statement about the relationship. What work have you watched them perform, and what happened?
A recommendation that survives that question is genuine evidence. One that does not is a favour being asked, which is legitimate as a social matter and is not information about a candidate.
Being able to tell them apart is what makes it possible to decline a referral without damaging the relationship, because you can be specific about why.
Declining somebody’s candidate
The situation owners most often handle badly, usually by going silent.
Silence is read as either a rejection of the person or a rejection of the referrer, and in a small community both interpretations circulate. The candidate hears about it before the referrer does.
Go back to the person who referred them, promptly, and say what happened without disparaging the candidate. We went with somebody who had done the estimating side before. I appreciated you sending him and I would want to hear from you next time.
That takes a minute, it preserves the referral relationship, and it prevents the account being constructed by other people.
Hiring somebody’s relative
Unavoidable in a small workforce and manageable if the terms are established at the start.
The failure is not the hire. It is the arrangement where a supervisor is also a parent or an in-law and nobody has said what happens if performance becomes an issue. Colleagues will assume the standard does not apply, and their assumption is usually correct.
What resolves it is structural: the person does not report to the relative, performance is assessed by somebody else, and that arrangement is visible rather than internal.
Where the firm is too small for that to be possible, say so plainly to the whole workforce and be conspicuously exact about applying the standard, because everybody is watching that specific thing.
Widening the pool without abandoning the network
The correction is not to stop using local knowledge. It is to ensure the network is not the only source.
Post the position properly even when a candidate is already in mind, so that somebody outside the usual circles has a route in. Ask the people who left the community and might return, using the approach from the workforce chapter. Ask a supervisor who in their own area would surprise you, which surfaces internal candidates whose reputation understates them.
Each of these takes an hour and each reaches somebody the referral network would not have produced.
The reference you give
The other half of this, and it matters in a small market more than anywhere.
An employer’s word about a former employee travels, and employers who give honest references — including the unfavourable parts, stated carefully — become the people other employers actually call.
What may be said about a former employee is constrained, and the constraints are worth establishing with counsel rather than assumed. Within them, the useful discipline is the one from the feedback entry: describe behaviour and specific instances rather than offering characterisations.
An employer known to give a straight account is doing something for the whole local labour market, and is also the person whose call gets returned when they need something.
Edited by Patrick J. Wolf, PhD