Year
2025-26
Role
Sole designer-researcher
Context
TCS Research x IIT Hyderabad
Duration
Mar 2025 to Aug 2026
Team
Sylvan Lobo (TCS Research), Prof. Delwyn Remedios (IIT Hyderabad)
Status
Accepted at ICoRD 2027 (Springer), first author
The problem
Every user knows more than they can say, and user research only hears what they say.
The outcome
M3 Allegory-Driven Elicitation: a new knowledge elicitation method that reads what people say against what they do, and uses a story to surface where their know-how comes from.
3
layers of knowledge captured in one session
ICoRD 2027 paper
4
gaps in today's methods it answers
7 designer interviews
7
unspoken patterns in the validation study
ICoRD 2027 paper
Overview
The goal: develop a new knowledge elicitation method for the research phase of any project, in any field, product or service, one that also reaches the unspoken know-how people can't put into words. This knowledge shapes how users actually use things, so missing it means designing for what people say instead of what they do. The catch: it sits below words, so it can't simply be asked for.
DISCOVER
Challenge
The problem is well documented. Philosopher Michael Polanyi put it as “we can know more than we can tell” (1966). Psychologists Nisbett and Wilson found that when people explain their own choices, they often give confident reasons that are wrong (1977).
I interviewed designers about how they capture this knowledge in their own research. They hit the same walls: people can't put it into words, what they say doesn't match what they do, work culture stops them sharing, and the session depends on the researcher improvising. Every standard method, from interviews to field visits, asks people to explain. So the most valuable knowledge is exactly what they all miss.
How often each gap was observed across standard research methods.
Let’s understand this with a scenario
“You just feel it”
Ask a batsman how they read a fast ball. They'll try. The wrist, the seam, something about the shoulder. Then they give up: “You just feel it.” The skill is real and took years to build, yet they can't tell you what it is. Now picture a researcher with a notebook, waiting for an answer that will never come in words.
Asked how he bowls, the expert explains in fragments, then stops.
DEFINE
How might we catch what people do, not just what they say, without losing what they say?
The know-how lives in the gap between the two. So the job isn't getting people to explain better. It's putting their words next to their actions and asking about the difference.
What people say covers only part of what they do.
DEVELOP
Approach
I built M3 through research through design (Zimmerman, Forlizzi and Evenson, 2007): design it, test it, redesign it. Existing methods each reach part of the picture. Think-aloud (Ericsson and Simon) captures what people can say. Contextual inquiry (Beyer and Holtzblatt) captures what they do. Cultural probes (Gaver, Dunne and Pacenti) and metaphor (Lakoff and Johnson) reach further but are hard to interpret. M3 brings these into one session, in three layers: people explain a task, then do it while I watch in silence, then reflect on it through a story.
It took three versions. The first let people describe their work through a story instead of directly. The second ran over several weeks and let people choose their own story; I dropped it because choosing got in the way of answering. The first validation session, with an IT professional, couldn’t go ahead: confidentiality meant he couldn’t show his organisation’s internal workings or software, and real workplace problems couldn’t be acted out at home or anywhere outside work. So validation moved to an everyday task anyone can perform anywhere: cooking.
From the validation study: asked whether he follows the recipe, the participant admits he doesn't really.
DELIVER
Result
Both research questions answered: yes, standard methods miss tacit knowledge, and yes, an allegory-based method can reach it without losing what people say. M3 gives researchers three things they didn't have:
one session that captures what people say, what they do and where their know-how comes from
a clear signal to look for: wherever what people say and what they do don’t match, that mismatch points to knowledge they couldn’t put into words
a story layer that reaches what observation alone can't, like who taught them and why the practice matters to them
The findings come from thematic analysis (Braun and Clarke, 2006), and a second coder is now checking they hold.
Key takeaway
What people leave out often matters more than what they say. User research should be built to catch what is missing, not just record the answers.
Reflection
The validation was small: five people, one everyday task. It shows the method works, not yet that it works everywhere. M3 is still being developed. Next: extend it to confidential, site-bound work like the IT case, and run it in product and service research side by side with a standard interview, to show it finds more.
Grounded in
Polanyi, The Tacit Dimension (1966)
Collins, Tacit and Explicit Knowledge (2010)
Nonaka and Takeuchi, The Knowledge-Creating Company (1995)
Ericsson and Simon, Protocol Analysis (1993)
Lakoff and Johnson, Metaphors We Live By (1980)
Gaver, Dunne and Pacenti, Cultural Probes (1999)
Zimmerman, Forlizzi and Evenson, Research through Design in HCI (2007)
Full list: 26 references in the ICoRD 2027 paper.
