Why reference checks get rushed, and what it costs

In SHRM’s benchmarking survey of 1,118 organisations, 54% said they use references to assess non-management hires. Which means the rest make those hires without one.

That number usually gets read as carelessness. It is not. Nobody decides references do not matter. What happens is smaller and more ordinary than that.

The step with the least resistance

Look at where a reference check sits. The candidate is already a yes. The hiring manager wants them to start. The offer is drafted. Everyone in the process has already formed a view, and the reference is the last thing between that view and the thing everyone wants.

Now add a busy week. Something has to give, and the reference is the obvious candidate, because:

• It is chaseable, so it can always be chased later. Unlike an interview, it has no scheduled slot to miss.
• It is expected to be positive. The candidate picked who you would call. Everyone knows this, which makes the exercise feel ceremonial.
• Nobody notices when it does not happen. A skipped interview is visible. A reference that was “started” and never finished looks identical to one in progress.
• It is somebody’s afternoon. Three references, each a voicemail, a callback at a bad time, and a follow-up nobody owns.

So it does not get skipped in a decision. It gets rushed, or half-done, or left open until the person starts and then quietly closed.

What the rushed version costs

The cost is not that you hire a fraud. That is rare, and a reference the candidate chose was never likely to catch it.

The cost is that you lose the one part of the process where someone who actually worked with this person tells you what the work was like. Everything else in hiring is the candidate describing themselves: a résumé they wrote, answers they prepared, a conversation they are performing in. The reference is the only input that comes from outside that.

What you lose specifically:

The thing they are genuinely better at than other people. This almost never comes out in an interview, because candidates lead with what they think you want. It comes out immediately when you ask a former manager to compare them to people they have managed.

What the job will have to teach them. Not a weakness, a starting point. You will find this out in month two either way. Finding it out before the offer means you can plan for it instead of being surprised by it.

The tone of the yes. “Absolutely, tomorrow” and “yes, I think so” are the same word and different answers. A rushed reference does not hear the difference, because a rushed reference is a form somebody filled in.

And when a hire does go wrong, the cost is not confined to the hire. A role that turns over at ninety days costs the team that covered for them, the manager who ran the process twice, and the people who trained someone who left.

Why “just do them properly” does not work

Because the advice is aimed at the wrong thing. Nobody needs convincing that references are valuable. The SHRM number is not a belief problem, it is a capacity problem.

Telling a recruiter who is running forty open roles to be more diligent about references is telling them to find time that does not exist. The week that squeezed out the reference check will squeeze it out again next week.

The only interventions that hold are the ones that remove the work rather than reprioritise it.

What actually makes it stop happening

Let the reference answer the way that suits them. The reason references go quiet is not reluctance, it is that a call is a scheduled commitment and a text is not. Given the choice, plenty of people will answer three questions by text at nine in the evening who would never have made a fifteen-minute call work.

Start it before the offer, not after. Once the offer is out, the reference has no ability to change anything and everyone knows it. Starting earlier is the difference between an input and a formality.

Ask the same questions every time. Not for fairness alone, though that matters. Because improvised reference calls get shaped by how much you already like the candidate: sharp questions for the ones you doubt, easy ones for the ones you want. That is backwards, and it is invisible while you are doing it.

Make the chase somebody else’s job. Reminders, retries, the follow-up nobody owns. This is the part that fails, and it is the part that does not need a person.

Decide in advance what an unresponsive reference means. Otherwise the answer is always “well, we liked them anyway.”

The honest version of the trade

If a reference check costs you a week of chasing, you will eventually stop doing them on the candidates you feel good about. That is not a character flaw, it is arithmetic.

Which means the question is not whether references are worth it. It is whether you can get one without it costing a week. If you can, the 54% number stops being a statement about diligence and starts being a statement about tooling.

---

Reference Agent texts or calls each reference, asks your questions, and sends back a scored reference. It is free while in beta.