The Two-Interview Method: Map First, Film a Week Later

The way to record training videos with subject matter experts is to interview them twice: one session to map the process, then a filming session about a week later. The first is a screen share where the expert does the real work while you build the map. The second is the production run, filmed after a week back in the job shows the expert what the first session shook loose. Collapsing both into one afternoon feels efficient, and it is the most expensive shortcut in knowledge capture.

Why is one interview never enough to record training videos with subject matter experts?

The first session with an expert is discovery. You do not yet know the shape of the work, which handoffs exist, or which steps hide a judgment call. Neither, in any useful sense, does the expert: they know how to do the job, not how to present it.

So the first hour produces what discovery produces: backtracking, "oh wait, before that," a tangent about the one customer who is different. Good mapping session. Unusable footage.

Film it anyway and you get both halves badly: a thin map, and an hour of footage nobody will watch. The coach who trained us in this protocol calls collapsing the two interviews grossly inefficient.

Never film on the same day you map. Harvest filming, the middle station of the five-step knowledge transfer wheel we run, interviews the expert twice by design: once to find the work, once to capture it. When you are documenting a process that only lives in someone's head, the map is what you do not have yet, and no camera will produce it.

An expert cannot narrate a map they are still drawing.

Interview one: map the structure and flag the sparks

Run the mapping session live on a screen share with two people on your side: one building the visual map, one taking notes and asking the deep questions. The expert shares their screen and does the real work, narrating. You are drawing the territory, not producing a video.

Give them the instruction we open every session with: we want you to say, no, it does not look like that, I am actually doing this. Frame the hour around improvement, not audit: what breaks, what gets worked around, what frustrates them today.

You are hunting three things: the purpose of the process, the decision points where the right answer depends on context, and the step order as the practitioner performs it. Listen as well for the moments the voice changes, the rule they will defend. Flag those and keep moving.

Expect the map to run long. A departing bookkeeper at one home services company took 85 minutes just to describe her weekly commission run, and the owner, watching, realized he was paying for a three-day job every week. Filming that would have bought 85 minutes of tape and no training. Mapping it told us which eight minutes to film, which is what the questions that make an expert interview work are for.

The engineered one-week break

The gap between the two sessions is not a scheduling accident. It is doing work.

The expert goes back to the job with the inventory fresh and starts seeing the process from the outside. The exception they forgot surfaces on Tuesday. The step they described in one word turns out to contain four. They come back, in the coach's phrase, a richer interview candidate.

Prime the week instead of assigning homework. Send the flagged list, ask them to notice anything that does not match, and tell them not to write anything up. Wait one full cycle of the process, and never less than a few days. A weekly close needs a week; a month-end routine needs the month end.

You need the week too: the map gets cleaned up and the filming outline gets built.

The week is not a delay but the part of the method that works while nobody is in the room.

Interview two: film the what, then ask the why

The filming session runs in three moves, in one sitting.

1. Shoot the what clean. The expert performs the task start to finish against the outline, narrating, uninterrupted. Stay silent and write your questions down. Retakes cost nothing, so restart a segment rather than talk over it.
2. Run the why pass. Now go back section by section and ask why here, why not the other way, what happens downstream if someone skips it. This is where the sparks you flagged last week get spent.
3. Turn stories into principles. Ask it both ways: is it because of this, or because of that? Keep going until the answer stops being a story about one bad Tuesday and becomes a rule that holds in a job you have never seen. That pass is a wisdom interview, and ten or twelve principles will cover processes you never filmed.

The passes stay separate because they produce different assets. The clean run becomes the training video. The why pass becomes the decision points inside the written SOP, and the reason a new hire calls it right in a case the video never showed.

Budget 2 to 4 hours of the expert's time per process across both sessions. The instruction never changes: show up and do the job while we watch.

Scripted or conversation-based?

We ask clients this outright, and the honest answer is almost always conversation-based with a scripted spine. A full script turns your best practitioner into a nervous actor, and you lose what video was supposed to buy: tone, hesitation, the feel of a judgment call in real time.

Script only what has to land word for word.

Outline-led is the working default: 5 or 6 action headers on a second monitor, no full sentences. The expert knows the order and picks their own words, which is why the module sounds like the company instead of a voiceover. When something does need a script, write it from the transcript of what they already said.

One rule from the edit bay applies either way: when the audio comes back bad, diagnose before touching anything, restore the take rather than regenerate it, and never let a patch run longer than the slot it fills. Whether a process deserves video at all is a separate call, weighed in video SOPs versus written SOPs.

---

Turn the film into assets, not archives

A recording is not a deliverable. It is raw material, and the discipline after the camera stops is the product.

The run is mechanical once footage exists: pull the transcript, map segments against the outline from interview one, select frames by sharpness, box the click targets in red, and assemble the SOP. One session should produce a written SOP with annotated screenshots, a trimmed video module, and a looping clip for the step everyone forgets. That conversion is covered in how to turn a 10-minute screen recording into a complete SOP. The recorder matters less than buyers expect, though Loom and Scribe fail in opposite directions.

Then put the assets where the work happens, organized by role. One company paid five figures to have its whole operation documented into a platform nobody ever opened, then commissioned the work again two years later with one condition: I do not want to do work that does not get used.

Interview, wait a week, film, convert: that loop is what we run at The Systems Effect for owners whose most valuable knowledge sits in one person's head. Run one lap yourself this week. Pick the process people interrupt you about most, book an hour with whoever really does it, and before it ends, put next week's filming session on the calendar.

Frequently Asked Questions

Why should you interview an expert twice?

Because mapping and filming are different jobs. The first interview is discovery: you are finding the handoffs and the decision points as the expert backtracks and corrects. The second is production, run against an outline built from the first. Doing both at once buys a thin map and unusable footage.

How long should you wait between a mapping interview and a filming interview?

About a week for most processes, and never less than one full cycle of what you are documenting. A month-end close needs the month end; a task that runs 20 times a day needs a few days. The break is deliberate: the expert goes back to work carrying the inventory, spots exceptions they could not name in the room, and returns a richer interview candidate.

Should training videos be scripted or conversational?

Conversational, with a scripted spine. Put 5 or 6 action headers on a second monitor and let the expert use their own words, so the module keeps the tone and judgment that made it worth recording. Script only what must land word for word, such as compliance wording, and write those lines from the transcript of what the expert already said.

What do you do with the recording after the interview?

Treat it as raw material, not a deliverable. Pull the transcript, map segments against your outline, select the sharpest frames, annotate the click targets, and assemble a written SOP with screenshots plus a trimmed video module. Then load it where the work happens, organized by role, because a recording in a drive folder is an archive, not training.