Ask What the Last New Hire Got Wrong (Then Blame the Onboarding)

To find the gaps in your onboarding process, do not send a survey. Ask whoever trains your new people one question: was there a recent mistake made by a new hire that you feel a better onboarding would have prevented? The phrasing puts the onboarding in the dock instead of the person, so the answer arrives with no defense attached. In our sessions with operators, that one question turns up more real gaps than an audit of the training material.

One Question Finds More Gaps in Your Onboarding Process Than a Survey

A survey asks people to rate something. This question asks them to remember something, which is easier and more honest. Nobody grades the onboarding out of ten. They tell you about the last time something went wrong.

In one engagement with a registered investment advisory firm, we asked it of the specialist who does most of the training. She gave three answers in about six minutes.

An internal transfer trained on an assumption that turned out to be false. A form that comes back in the mail with the social security numbers missing, every time, from everyone. And cases marked complete before the sub-steps were finished, because only some case types force an action item.

None would surface on a survey. Two would not surface on a process map either, because a map shows the work as designed.

Why does the phrasing do the work?

Because it assigns fault before the person has to. Ask "what mistakes do new hires make" and you get a character assessment. Ask "what mistake would better onboarding have prevented" and you get a process defect.

Watch what that buys you. The specialist had trained an internal transfer who spent a year at the front desk, and she assumed a year in the building meant the person understood the back office. It did not. In her words: "I didn't realize that it was always from like a peripheral view."

Then she said the part most people never say out loud to a consultant: it was easier to train a brand new person than someone who was already there.

Assumed context is the most expensive thing in an onboarding process, because it is the only part nobody can see.

That admission is not available to a question that blames the trainee. Ask about the mistake in a way that makes the process the defendant, and the person will tell you what the process left out. It is the same posture we use to interview a subject matter expert about any work that was never written down.

Ask Again After They Say You Have Covered Everything

The polite end of a good conversation sounds like this: "I feel like we've covered everything." That is not resistance. It is what people say when the obvious material is spent, which happens well before the useful material arrives.

Do not close the call there. Ask one more specific question: is there any other situation from the past that you feel could have been avoided if there was better onboarding? In that session it got an immediate no, then a pause, then two of the best items of the hour, including the mail forms.

The social security numbers get skipped by everybody, and she knew it: "It happens to every single person. It might just be one of those things that it has to happen to you before you're like, oh yeah, I got to remember to do that."

A mistake that happens to every single person is not a personality problem, it is a missing step.

Everyone makes that error once, everyone learns it the hard way, and nobody has written the line that would stop it. More shadowing does not fix that. A checklist line does, and it is exactly the sort of item that disappears when you are staring at years of accumulated training content wondering where to start.

Ask for the Counter Example, Not the Rule

Rules produce generalities. Exceptions produce mechanism. Any description of how a decision gets made comes out cleaner than the last five times it actually happened.

So ask for the time it went the other way. In one session, asking whether anyone had ever been a fit for a move and still not been moved exposed a whole timing problem: staff committed to projects that then slipped by nine months. The stated rule hid it completely.

For onboarding, ask about the new hire who did fine without the step everyone insists is essential, or the one who followed the documented process and still got it wrong. Both answers tell you which parts of your onboarding are load bearing and which are habit.

Ask What Happens If It Does Not Get Done

New hires drop steps because nothing in the training told them what the step was for. A due date with no consequence attached is a preference, and people quietly demote preferences under pressure.

One question fixes that: what happens if that is not done by then, and is there anything you have to escalate? In an asset management session it turned a calendar date into a hard business rule. If those notices are not mailed on time, we cannot raise the rent, and there is no option for failure there. Nobody drops that step once the sentence exists.

Same story with the cases closed early. Only some case types force an action item, and nobody had told a new specialist which was which. Every step that carries a real consequence should state the consequence in the step itself, not in a manager's head. That is half of why training stops working even when the content is correct.

What Would the New Person Not Know?

Vocabulary. Almost always vocabulary, and almost never on anyone's list of onboarding gaps.

Ask it so that not knowing is the expected answer: should a new hire know the definition of every account type, or is there a possibility somebody coming in may not have that context? Phrased that way it reframes the gap as normal rather than as ignorance. The trainer said she would almost say the majority of people coming in do not know all the terminology.

That is a five-minute glossary that saves a new person weeks of nodding along, and it separates material that gets skimmed from training people actually finish.

---

Turn Each Answer Into Exactly One Artifact

The failure mode after a good interview is ambition. You collect four real gaps, decide to rebuild the whole program, and ship nothing. Build one small thing per answer instead.

1. Name the moment, not the topic. "The step where mailed forms come back" beats "documentation accuracy" every time.
2. Write one artifact, not a section. A checklist line, a glossary entry, a two-minute screen recording. One answer, one object.
3. Put it where the mistake happens. File it in the training platform and nobody meets it at the moment the error occurs.

Applied to the three answers above, the output is very small.

Not one of those is a course. They are a checklist edit, a field, and a rule, and together they close three recurring errors a survey would never have found. Slot them into a 90 day onboarding checklist so they land when the new person can use them.

Then give people real work to practice on, rather than whatever walked in the door that day, which is what a training caseload is for.

Make it a standing review. At The Systems Effect we run this question with every client after somebody finishes their first ninety days, then turn the answers into the SOPs and training the onboarding was missing. Ask it this week, of whoever trained your last hire, and write down one thing.

Frequently Asked Questions

How do you find gaps in an onboarding process?

Ask the person who does the training about a recent mistake a new hire made that better onboarding would have prevented. Real gaps live in errors, not opinions, and an error is concrete enough to fix in a single edit. Surveys tell you how onboarding felt, which is far less useful.

What question exposes onboarding problems fastest?

"Was there a recent mistake made by a new hire that you feel a better onboarding would have prevented?" It is the highest-yield question we have observed across engagements. One session produced three specific defects in minutes: an internal transfer trained on assumed context, a form field everyone forgets, and work marked complete before all the steps were finished.

How often should you review onboarding?

Review it every time someone finishes their first ninety days, while the trainer still remembers what went wrong. Waiting a year turns detail into a general impression, and general impressions do not produce artifacts. If the same mistake appears twice, fix the step that week.

How do you ask about mistakes without making people defensive?

Put the blame in the question. Attributing the mistake to the onboarding rather than the person makes the answer a process observation instead of a confession, so nobody has to protect a colleague or themselves. Ask what the new person would not know rather than what they got wrong.