Keyhole projective technique – observing brand use
A projective, guided imagery exercise that uses a naïve observer – a curious child looking through a keyhole – to externalise how participants perceive a brand being used. Instead of asking people what they think, the technique asks them to describe what an innocent bystander sees, which bypasses self-censorship and social desirability to surface impressions participants might not otherwise voice.
The qualitative theory
Section titled “The qualitative theory”Projective techniques are among the oldest and most powerful tools in qualitative research. The principle is that people find it easier – and more psychologically safe – to describe what someone else experiences than to articulate their own thoughts and feelings directly. When asked “what do you think?”, people give considered, socially acceptable answers. When asked “what does this person see?”, the filter drops.
The child observer is a specific and deliberate choice. A child is the archetypal naïve observer – curious, uninhibited, and without the brand literacy or social context to construct a “correct” interpretation. Asking participants to describe what a child sees through a keyhole does several things at once: it externalises the experience (reducing defensiveness), it introduces a step-by-step narrative structure (encouraging rich sequential description), and it creates enough psychological distance that participants can surface perceptions they might hesitate to own directly.
The guided imagery framing reinforces this. By asking participants to imagine a scene and describe it as if they could be there – rather than recount an experience – it activates more vivid, sensory, and emotionally honest recall. The result is data about the brand experience as it is actually felt, not as participants think it should be described.
The laddering structure in this methodology is sequential rather than associative. It follows the experience from start to finish: how it begins, what happens next, the emotional dimension, and the behavioural cues that reveal those emotions. This produces a step-by-step phenomenological account of the brand experience – what happened, in what order, and how it felt – which is exactly what makes the technique valuable for journey mapping and CX work.
Configured settings
Section titled “Configured settings”Research framing
Section titled “Research framing”| Field | Value |
|---|---|
| Context & Background | A projective, guided imagery exercise using a naïve observer (a child) to externalise perceptions. The child watches the brand being used, and the respondent describes step by step what the child sees. |
| Research Objectives | Understand how the brand is used, the step-by-step experience, points of difficulty or friction, the emotions involved, and the overall impression the brand leaves. |
| AI Moderator Tone | Friendly and casual |
The friendly and casual tone is important here – the projective setup asks participants to engage imaginatively, which requires a relaxed, non-clinical environment. A formal tone would undermine the guided imagery framing.
Conversation introduction
Section titled “Conversation introduction”Hi, I’m
{{companionName}}, and I’m your survey companion. I’m going to ask you to imagine a scene and describe what you see such that I could imagine being there myself. There are absolutely no right or wrong answers, we’re interested in whatever comes to mind.Imagine a curious child is looking through a keyhole watching someone use BRAND from start to finish. Describe what the child sees, starting with the scene and what’s going on – describe this in detail!
The introduction does three things: it establishes the companion persona, it sets up the projective frame (“imagine a scene”), and it places the child observer at the keyhole. “Such that I could imagine being there myself” is a classic guided imagery prompt – it asks for sensory, descriptive detail rather than abstract evaluation.
The instruction to “describe this in detail” at the end of the introduction is deliberate. It signals that richness and specificity are expected before any laddering begins.
Laddering questions
Section titled “Laddering questions”Five questions at maximum, with Randomise off and Paraphrase off:
| # | Question | Purpose |
|---|---|---|
| 1 | How does the person start using BRAND? | Establishes the beginning of the experience – the entry point into the journey |
| 2 | What happens next – what does the person do? | Advances the sequential narrative through the core usage experience |
| 3 | Then what happens – what does the person do after that? | Continues the step-by-step journey, capturing the middle and late stages of use |
| 4 | How does the person seem to feel while using BRAND? | Shifts from behavioural description to emotional observation |
| 5 | What does the child see or notice that suggests the person feels this way, e.g., anything in how they act, react, or behave? | Asks for behavioural evidence of the emotion – grounds the feeling in observable cues |
Randomise is off because these questions follow a deliberate narrative arc – they’re designed to walk the participant through the experience from start to finish, then shift to the emotional layer. Disrupting the order would break that arc.
Paraphrase is off for the same reason. The specific phrasing – particularly the shift from “what does the person do?” to “how does the person seem to feel?” – is intentional and should not vary.
Questions 4 and 5 work as a pair. Question 4 invites the emotional interpretation; question 5 grounds it in what the child actually observes, not just what the participant infers. This two-step structure is key to getting emotionally rich data that stays anchored to the experience rather than drifting into abstract evaluation.
Default probe
Section titled “Default probe”Tell me more about
{{topic}}?
Paraphrase on. A deliberately simple probe – in this methodology, the guided imagery and laddering are doing the heavy lifting. The probe is just an opening, not a directed follow-up.
Default prompt
Section titled “Default prompt”What about
{{topic}}, can you tell me about that?
Paraphrase on. Open-ended framing consistent with the observational register of the technique – it asks the child to notice, not the participant to evaluate.
Probe & prompt limit
Section titled “Probe & prompt limit”12 – higher than the default to accommodate 6 topics, several of which have both probing and prompting active. With two topics carrying prompt overrides, the AI Moderator needs headroom to cover them without hitting the cap before all topics have been explored.
Topics
Section titled “Topics”| Topic | Description | Probes | Prompts | Prompts override |
|---|---|---|---|---|
| Friction points | Describe moments that appear difficult, confusing or require extra effort during use | 1 | On | What, if anything, does the child notice that seemed difficult, confusing or not quite smooth? |
| Trust and confidence | Understand how assured and certain the user appears while using the brand | 2 | On | How confident does the person seem while using the brand? |
| Emotional journey | Explore how feelings or emotions evolve throughout the usage experience | 1 | Off | – |
| Outcome against expectations | Explore what happens after use and whether the outcome met expectations | 1 | Off | – |
| Product or service interaction | Understanding how the user engages with specific features or elements of the brand | 1 | Off | – |
| Memorability | Identify what stands out most or is most likely to be remembered from the experience | Off | On | What is the child most likely to notice or remember from what they have seen? |
Randomise probe order: off. Randomise prompt order: off.
A few things worth noting:
Friction points and Trust and confidence both have prompt overrides. These are the topics where the researcher most needs data – friction in particular is often not mentioned spontaneously because participants don’t want to criticise, and the child-observer frame is used to soften the prompt (“what did the child notice that seemed not quite smooth?” is less confrontational than “what went wrong?”).
Memorability has probing disabled but prompting on. If the participant mentions something memorable, the AI Moderator won’t follow up on it – but if they haven’t, the prompt will surface it. The specific prompt wording (“what is the child most likely to notice or remember”) uses the projective frame to ask for salience without asking for evaluation.
Emotional journey, Outcome against expectations, and Product or service interaction all have prompts disabled because these should emerge naturally from the guided imagery and laddering. Prompting them risks introducing a research agenda into what should feel like an observational narrative.
Conversation ending
Section titled “Conversation ending”- Closing question: Finally, what lasting impression does the child come away with about BRAND?
- Dynamic ending: On
The closing question returns to the child observer frame one final time, asking for the overall impression left by the experience. “Lasting impression” invites synthesis without demanding evaluation – participants describe what the child carries away, not what they personally conclude. It’s a natural close for the guided imagery arc.
Dynamic ending is on so the AI Moderator can wrap up contextually based on what was discussed, which fits the conversational, imaginative register of this methodology better than a fixed sign-off.
Adapting this methodology
Section titled “Adapting this methodology”When you apply this to a new conversation, these are the fields you’ll typically need to change:
- Replace BRAND in the introduction and all five laddering questions with your brand, product, or service name.
- Context & Background – add context about the specific product, category, or user group you’re researching.
- Research Objectives – update to reflect the specific aspects of the experience you most need to understand.
- Topics – the current list covers a broad brand experience. Add, remove, or retitle topics to match the specific journey or touchpoints you’re investigating. Keep descriptions grounded in what the child observes rather than what the participant evaluates.
- Prompt overrides – the current overrides for Friction points and Trust and confidence are good templates. If you add new topics where prompting is important, write the prompt override in the same child-observer register: “What does the child notice about…?” rather than “Tell me about your experience of…”.
The laddering questions, their order, and the two-question emotional sequence (4 and 5) should be left unchanged wherever possible. They’re the structural core of the technique.
Frequently asked questions
Section titled “Frequently asked questions”Why use a child as the observer rather than an adult?
A child is a naïve observer by definition – they have no brand literacy, no social context, and no reason to give a “correct” answer. This makes the projection easier for participants, and it removes the implicit pressure to evaluate the brand fairly or charitably. Adults asked to describe what another adult sees tend to attribute motives and judgements that reflect their own. A child just describes what’s in front of them.
Why are the laddering questions in a fixed order?
They follow a deliberate narrative arc: entry → action → progression → emotion → behavioural evidence of emotion. This sequence builds a sequential phenomenological account of the experience. Randomising the order would mean asking how the person feels before establishing what they did – the emotional question only makes sense once the experience has been narrated.
Why is prompting disabled for Emotional journey, Outcome against expectations, and Product or service interaction?
These topics should emerge naturally from the guided imagery and laddering. The step-by-step narrative structure is specifically designed to surface the emotional and experiential journey – if a participant doesn’t mention emotions or interactions while narrating the experience, prompting them risks introducing a more analytical, evaluative register that breaks the projective frame.
Can this technique be used for service experiences or digital products, not just physical brands?
Yes – the technique works well for any experience that can be narrated step-by-step. For digital products or services, adapt the introduction slightly: rather than “watching someone use BRAND”, the child might be “looking over someone’s shoulder as they use BRAND”. The core projective principle remains the same.
Why does the closing question return to the child observer frame?
Consistency – maintaining the projective frame throughout prevents a jarring shift in register at the end of the conversation. Switching to a direct question (“what did you think?”) at the closing undermines the psychological distance the technique has established. “What impression does the child come away with?” keeps the participant in the observational mindset while still inviting synthesis.