Why "Shadow Someone" Training Fails (And What to Build Instead)

Put the new person next to the person who already knows, and let them watch. Job shadowing is not enough for training because it hands over one person's habits from one week of work, skips the decision points where the job actually gets hard, and leaves nothing behind the day the new hire forgets. You do not get a trained employee. You get someone who has watched the job get done.

Shadowing Is a Lottery, Not a System

A shadowing program has no curriculum, no finish line, and no way to tell whether anything transferred. What the new hire learns depends on which week they walked into: a slow week and they never see the escalation that matters, the wrong desk and they inherit somebody's workaround as if it were policy.

We watched an owner make this bet on a dispatcher. He sat with the man for two days, and weeks later was still asking out loud how he could trust the steps got followed when the guy could not reliably check his own calendar. Two days elbow to elbow, and none of it held.

Nothing about those two days was built to survive them.

Shadowing is not training, it is proximity, and proximity teaches whatever happened to be in the room that week.

Why Is Job Shadowing Not Enough for Training? What It Skips

Shadowing does teach real things, and the honest version of this argument says so first: tone, pace, how the room changes when a customer is angry. Video carries some of that. Text carries none.

Here is what it does not carry. For every process we capture three things: the purpose, the decision points where somebody makes a judgment call, and the step by step from the person doing the work. Shadowing gives you a thin version of the third and almost none of the first two, and that gap is usually the real answer when owners try to diagnose why their employee training is not working. The expert narrates the happy path on a Tuesday, and the exceptions are what break the business.

The deeper miss is invisible. A consultant who does this kind of knowledge transfer for a living describes it exactly: two people follow the same steps, one gets it right every time and one cannot get it right to save his life. The difference is never in the steps. It is ten or twelve principles running unconsciously in the good one's head, and nobody narrates what they do not know is there.

Three People, Three Versions of the Process

One symptom we look for in a first conversation: three people do the same process three ways, and nobody in the building knows it.

A nonprofit's largest department finished a wave of SOP work, and the retro surfaced something nobody had scoped. Because the whole team sat in the sessions, they heard each other describe the same process out loud for the first time. What kept coming back was some version of "I did not really know that you were doing that," and that is what made one documented way possible.

Now drop a new hire into that department and tell them to shadow. They do not learn the process. They learn one of three, chosen by seating chart, and inside a month they have invented a fourth.

Shadowing does not spread the best version of a process, it spreads whichever version had a free chair.

The Cost: 6 Months to Competence

The buyer profile this practice was built around is the company where new hires take 6 months to become useful because training is "shadow someone." That is not a training budget of zero. It is the most expensive program in the building, paid in payroll: half a year of salary at partial output, plus the trainer's output cut while they narrate.

Then there is the cost nobody invoices. An owner covered his own payables process for a few weeks after a key employee left, expecting to find it fine. He found 500 to 800 dollars leaking every week from skipped audits, missed part deductions and overpayments, up to 10,000 dollars a month by his own math. Not theft, just a process nobody followed.

In that same company, the departing bookkeeper was training her replacement over screen shares, because none of that work had ever been written down.

A 6 month ramp is not a cheap training program but an expensive one that never leaves an asset behind.

It is the same arithmetic behind the gap between what companies spend on salaries and what they spend on training. The money gets spent either way. Shadowing spends it where nothing accumulates.

---

What Should You Build Instead? Recorded Reality Plus a Checklist

The objection arrives fast: there is no time for this. That is exactly the paradox of handing off work when you have no time to train: there is no time because the last five hires were taught the same way. The replacement is one real run of the work, recorded, with a checklist beside it.

1. Pick the practitioner, not the manager. The person doing the work knows things the person managing it has forgotten or never knew. One person, not a committee: get to a first 80 percent, then let the peer group tear it apart.
2. Record one real run. Not a demo and not a rehearsal: screen capture with narration while the actual work happens, roughly 2 to 4 hours for a process. The standing instruction is to interrupt, and say no, it does not look like that, I am actually doing this.
3. Capture the forks, not the clicks. Stop wherever the right answer depends on context and ask what the right call looks like. Undocumented forks are where processes break.
4. Ship a checklist beside the video. The video teaches the shape of the job. The checklist is what someone holds on their fourth day when they cannot remember step 9.

Notice what the recording replaces: not the expert, the expert's availability. Every hire after this one learns from the best run that person ever did. The next question is sequencing them, which is how to turn your SOPs into a training program.

Keep Shadowing, but Give It a Rubric

Do not delete shadowing. Move it. It is a poor first step and an excellent last one.

Shadowing verifies competence, it does not build it.

Write down what a competent run looks like before the shift: 6 to 10 observable behaviors, each one either done or not done. The trainee works, the observer scores the list, and the gaps become next week's practice. A shadow shift with a rubric produces a number. Without one it produces an opinion.

A staffing branch we work with runs that order before anyone touches a real customer: mock calls scored against a grading rubric first, then real calls pulled from the dialer and reviewed against the same standard. That is the sequence behind coaching reps with mock calls before customers do it for you.

Scored shadowing comes after both, in days 31 to 60, once the documentation and the practice reps are already behind the new hire. That is where it sits in a new hire onboarding checklist for the first 90 days.

Start With One Role This Month

Pick the one role where a resignation letter would hurt most, and the single process inside it that would hurt most on the Monday after.

Recording the practitioner instead of interviewing the manager is the whole premise of the work we do at The Systems Effect, and it is why the documentation matches how a job is really done rather than how somebody remembers it.

So take one afternoon this month. Record the person who really runs that process, doing it for real, and let the next new hire watch the recording before they ever watch the person.

Frequently Asked Questions

Why does job shadowing fail as a training method?

Shadowing fails because it is unstructured observation, not instruction. The new hire sees only what happened during the days they sat there, so the exceptions, the judgment calls and the reasons behind the steps go untaught. It also has no completion point, which is why companies that train this way commonly need 6 months to get someone useful.

What should replace shadow training?

Replace it with a recording of one real run of the work, made by the person who actually performs it, plus a checklist the trainee holds while working. Capture why the process exists, where the judgment calls happen, and the literal step by step. That costs 2 to 4 hours of the expert's time and then trains every hire after.

How long should new hire shadowing last?

A few shifts, not a few months, and only after the role's documentation exists. As verification it works well: the trainee runs the work, an observer scores them against a written list of behaviors, and the misses become the next practice session. If shadowing is still the main way people learn after month one, it is standing in for training that was never built.

How do you standardize training when everyone does the job differently?

Get the versions into the same room before you document anything. When teams describe the same process out loud to each other, they routinely discover colleagues have been running it differently for years, and that discovery is what makes one agreed way possible. Then let one practitioner build the first 80 percent and the rest correct it, so the standard carries their fingerprints instead of arriving as an order.