Latest / 5 Minute UX / Facilitator Prompt Levels (1, 2, 3): A Practical Guide
Transcript
- 0:08This lesson is Facilitator Prompt Levels (1, 2, 3): A
- 0:11Practical Guide. Let's get started.
- 0:20Ever watch a usability test where the facilitator jumps in
- 0:23too soon and accidentally fixes the design flaw for the user?
- 0:26You just lost the only data point that proves your interface
- 0:28is broken. That is exactly why facilitator prompt levels matter
- 0:31so much in every single study you run. These levels act as a
- 0:34critical mechanism to gauge the true difficulty of a task
- 0:37without breaking the integrity of the session. By strictly
- 0:39escalating from open inquiries to direct instructions, you
- 0:42determine precisely where a user's mental model collapses.
- 0:45This structured approach transforms your subjective
- 0:47observations into hard, quantifiable data points.
- 0:49Suddenly, you can distinguish between a task that is merely
- 0:52challenging and one that is fundamentally flawed. You stop
- 0:55guessing and start measuring the exact moment your design fails
- 0:58the user. That clarity is the only way to build products
- 1:00people actually want to use.
- 1:07By the end of this section, you'll be able to describe the
- 1:10pre-session preparation requirements, including defining
- 1:12baseline knowledge and setting scoring limits. Before you ever
- 1:15speak to a participant, you must establish exactly what baseline
- 1:18knowledge they are expected to have. This definition prevents
- 1:21you from misinterpreting a lack of domain expertise as a design
- 1:24failure. You also need to create a decision matrix or scoring
- 1:26sheet to track the specific prompt type and count for every
- 1:29single task. This document turns subjective observations into
- 1:32hard, comparable data points. Most critically, you must set a
- 1:35hard limit on Level two prompts, often capping them at two
- 1:37instances, before marking the task as failed. Without this
- 1:40pre-determined rule, your data becomes inconsistent and
- 1:42unreliable. You cannot afford to let one session's leniency skew
- 1:45the results of your entire study.
- 1:53You start the sequence by asking open-ended questions that
- 1:56encourage self-correction without revealing the solution.
- 1:58This is Level One, where you might simply ask, "What are you
- 2:01thinking right now?" or "Where would you look next?" The goal
- 2:03here is to see if the user can identify the missing step on
- 2:06their own, which proves the interface is actually
- 2:08discoverable. If they self- correct, you stay silent and let
- 2:11them finish the task. When the participant remains stuck, you
- 2:13escalate to Level Two by providing specific hints that
- 2:15narrow the search area. You might say, "Have you checked the
- 2:18top right corner of the screen?" or "You need to find a button
- 2:21that says Submit." This level tests whether the user can solve
- 2:23the problem if given a slight nudge in the right direction,
- 2:26without you explicitly stating the answer. Most protocols limit
- 2:28you to two Level Two prompts before the task is marked as
- 2:31succeeded with difficulty. If the user still cannot proceed
- 2:33after those hints, you must move to Level Three and provide the
- 2:36direct answer. You state the exact action required, such as,
- 2:39"To approve this story, you would click the Submit button."
- 2:41This moment signals that the task has failed in a real-world
- 2:44scenario, so you record the failure and move immediately to
- 2:46the next objective. Giving this direct instruction effectively
- 2:49ends the test for that specific goal. Before you even begin the
- 2:51session, you must define the baseline knowledge expected of
- 2:54the participant and set hard limits on prompt usage. You need
- 2:57a decision matrix or scoring sheet that tracks the number and
- 2:59type of prompts given per task to ensure your data is
- 3:01objective. Without these pre- session preparation
- 3:03requirements, you cannot distinguish between a task that
- 3:06is merely challenging and one that is fundamentally flawed.
- 3:08The most common pitfall is premature escalation, where you
- 3:11jump to Level Three because you feel uncomfortable with the
- 3:13participant's silence. Another error is over-hinting, where a
- 3:16Level Two prompt is so specific it actually reveals the answer,
- 3:18which invalidates your struggle data. You must stick rigidly to
- 3:21the pre-defined script and escalation rules to maintain the
- 3:24integrity of your findings. Remember that your goal is to
- 3:26observe failure, not to prevent it, because that is where the
- 3:29real design insights live. You apply the escalation sequence to
- 3:31a stuck participant while avoiding premature intervention
- 3:34or over-hinting at all costs. By strictly adhering to these
- 3:36steps, you transform subjective observations into quantifiable
- 3:39data points that prove exactly where a design fails to support
- 3:41the user's mental model.
- 3:48Let's say you feel that uncomfortable silence as a
- 3:50participant stares at the screen. Do not jump to Level
- 3:52three just because the quiet makes you uneasy. That premature
- 3:55escalation masks the true usability issue and ruins your
- 3:57data. Here is how over-hinting breaks your test. You might say,
- 4:00"check the top right corner," which sounds like a hint but
- 4:02actually reveals the answer. That turns a Level two prompt
- 4:05into a Level three failure before you even realize it. To
- 4:07fix this, you must stick rigidly to your pre-defined script and
- 4:10limits. Consistent scoring means applying the exact same rules
- 4:12for every single participant. If you deviate, you cannot compare
- 4:15results or trust your findings.
- 4:22Consider your last project and pause to think about the moment
- 4:24a participant froze on a critical task. Did you jump in
- 4:27to help them, or did you let the struggle reveal the true design
- 4:29flaw? This hesitation is exactly where your data gets compromised
- 4:32if you haven't prepared your rules of engagement. Now,
- 4:35define your specific Level two limits and Level three failure
- 4:37criteria in your next test protocol before you ever speak
- 4:39to a user. You must decide if two hints mean a task failed or
- 4:42if it just succeeded with difficulty. Without this pre-
- 4:44determined limit, your scoring becomes subjective and your data
- 4:47becomes useless for comparison. Practice delivering these
- 4:49prompts neutrally to avoid influencing the participant's
- 4:52actions or steering them toward the answer. If you sound eager
- 4:54to solve the problem, you turn a Level two hint into a Level
- 4:57three answer and mask the usability issue. Your voice must
- 4:59remain flat and observational, letting the silence speak louder
- 5:02than your intervention. Finally, train your team on the
- 5:04exact definitions of each level to maintain data consistency
- 5:07across every single session. If one facilitator calls a specific
- 5:10hint a Level two while another calls it a Level three, your
- 5:12entire study collapses. You cannot compare results unless
- 5:14everyone applies the escalation sequence with the same rigid
- 5:17precision. That's the discipline required to turn
- 5:19subjective struggle into quantifiable failure. You've now
- 5:21learned how to measure exactly where the user's mental model
- 5:24breaks down without fixing it for them.
- 5:29Welcome to The Take. I'm joined by a fellow practitioner and
- 5:32together we're going to dig into Facilitator Prompt Levels (1, 2,
- 5:353): A Practical Guide — what it means in practice and what's
- 5:38worth remembering.
- 5:40[HOST] That rule about capping Level Two prompts at two
- 5:43instances is the hardest part to enforce in a live session.
- 5:45[LEARNER] It really is, because the silence feels agonizing when
- 5:49you know exactly where they need to click. [HOST] That discomfort
- 5:52is exactly what you have to tolerate to capture the true
- 5:54failure point. [LEARNER] I often catch myself slipping into a
- 5:57Level Three answer just to relieve that tension before the
- 6:00second hint. [HOST] If you do that, you lose the data point
- 6:03that proves the interface is actually broken. [LEARNER] It's
- 6:06a fine line though, because sometimes a specific hint like
- 6:09"check the top right" accidentally reveals the answer.
- 6:11[HOST] That's the over-hinting trap where a Level Two prompt
- 6:14effectively becomes a Level Three intervention. [LEARNER]
- 6:17I've seen teams use AI to generate script variations, but
- 6:20it can't fix the facilitator's instinct to jump in. [HOST] AI
- 6:23might help draft the prompts, but it can't replace the
- 6:26discipline of staying silent during that struggle. [LEARNER]
- 6:28True, and without that rigid scoring sheet, we end up with
- 6:31inconsistent data across different moderators. [HOST]
- 6:34Consistency is the only way to turn subjective struggle into
- 6:37quantifiable failure metrics. [LEARNER] So the real skill
- 6:40isn't writing the script, it's training the team to resist the
- 6:43urge to help. [HOST] Exactly, because observing the failure is
- 6:46the only way to find the design flaw we need to fix.