Lost in the Learning Loop
A Mixed-Methods UX Research Study
Part 1 of 4 · The Research
The Learning Decision is an Identity Decision in Disguise
Research conducted February–March 2026 · Survey: n=20 · Interviews: n=4
1. Why This Research
I started this because I was looking for people like me.
The market has never had more options for self-directed learners. Courses, tutorials, bootcamps, YouTube, certifications, the choices are endless and the noise is loud. Yet the people I knew, analysts, designers, career-switchers, were not thriving in this abundance. They were overwhelmed by it. They were starting things and stopping. Scrolling LinkedIn and feeling behind. Asking everyone around them what to learn next, and still not knowing.
I wanted to understand how career-driven learners make the decision of what to learn and not the ideal version of that process, but the real one.
2. What I Did
This is an exploratory, mixed-methods study combining a quantitative survey with qualitative moderated interviews.
Survey
20 respondents~ designers, analysts, career-switchers, and students across the US, India, and Europe. Distributed via LinkedIn and personal network.
Interviews
4 moderated interviews, 30–40 minutes each, conducted over video call. Participants were recruited from survey respondents who opted in, and from the researcher’s extended network.
Analysis
Thematic analysis and affinity mapping in FigJam. Interview transcripts processed via Otter.ai.
Limitations
This is a small, exploratory study. 20 survey responses and 4 interviews are not statistically significant. Findings are directional, patterns worth investigating at larger scale, not conclusions to generalize broadly. Two participants were personally known to the researcher, which may have affected researcher distance in those interviews.
3. Who I Spoke To
Four participants, recruited from survey opt-ins and the researcher’s extended network:
- P1 : Budding Product Manager/ Student. Career-switching; all learning decisions oriented toward one specific role.
- P2 : Financial Analyst / Career Switching. Targeting Quant Finance and Data Analytics simultaneously; job hunting.
- P3 : Teaching Assistant. Learning French B1→C1 for visa stability and Masters application.
- P4 : Designer. Already practicing; learning to improve an existing skill.
All participant identities are anonymized. Names have been withheld throughout.
4. What the Survey Said
75% of respondents have abandoned at least one course. The top reason was not losing motivation. It was finding something more interesting or urgent, 9 out of 20 respondents said this. The second most common reason was the course feeling less relevant than expected, 7 out of 20.
That distinction matters. Motivation running out is an internal failure. Something better coming along is a decision made in the moment, often unconsciously, in favor of the next shiny thing.
Yet the dominant emotion when starting something new is not ‘doubt’. It is excitement. 9 out of 20 respondents described feeling excited when deciding what to learn next. Only 1 said they felt genuinely lost about where to start; meaning for most, the problem is not confusion.
But excitement does not arrive alone. 8 respondents felt pressured by the job market. 6 felt overwhelmed by too many options. 4 felt anxious about picking the wrong thing. These were not separate groups, many reported all of these at once. This is not contradiction. This is the baseline state of the self-directed learner: wanting everything, pressed by the market, paralyzed by abundance. The problem is not lack of motivation. It is too much of everything at once, with no internal framework to cut through it.
One more finding worth naming: social media influences but does not decide. 12 out of 20 respondents said seeing someone else’s post or project online “sometimes plants ideas” they then research further. Only 1 said social media heavily drives decisions.
18 out of 20 respondents said some version of this:
“I’ll finish this one.”
The confidence of starting and the reality of commitment collide hard. When excitement meets the actual cost of choosing one thing over everything else, 75% of the time, something gives way. That is not a motivation problem. That is a commitment problem. And commitment is hard when you have not yet decided who you are becoming.
5. What the Interviews Revealed
The survey told me what people do. The interviews told me why.
Going in, I had two hypotheses. Both were wrong.
Hypothesis 1: P1 would show an External Permission pattern; needing validation from others before committing to a learning path.
What actually happened:
Get Prachi Gupta’s stories in your inbox
Join Medium for free to get updates from this writer.
P1 was the clearest, most decisive participant in the study. The learning path had been decided long before the interview. “Since the start, I wanted to become a product manager only.” The word ‘become’ not learn, not try ~ said everything.
Hypothesis 2: P3 would show the same blocked pattern as P1; clear goal, external barrier.
What actually happened:
P3 knew the goal, knew the path, and knew exactly what was at stake, yet the learning has been on and off since 2019. P3’s barrier was not external. It was internal ~ a detailed backup plan that made Plan A feel less urgent.
The four participants, mapped across identity clarity and commitment:
P1 : Decided
All learning decisions are oriented toward one specific career role. Abandonment is rare and financially conditional, paid courses get finished; free ones don’t. The entry decision is made carefully. The commitment holds.
P2 : Searching
Targeting an industry but not a role. Learning decisions are made in two layers: observing what successful people in the market look like and researching what the market currently demands. Courses are abandoned not from loss of interest, but from shifting priority. A sticky note system was introduced to manage this; but the system treats the symptoms, not the cause.
P3 : Drifting
Clear goal, uncertain commitment. The discipline is acknowledged but not acted upon. What makes P3 particularly revealing is the backup plan, its detail quietly reduces the urgency of Plan A. P3 is not drifting despite knowing what is at stake. P3 is drifting partly because there is always somewhere else to land.
P4 : Moving
Already practicing the skill. Already has an identity in it. The learning decision is not about becoming something, it is about improving what already exists. A mentor set the direction years ago. Lowest abandonment of the four.
6. The Pattern Across All Four
Four people. Four different situations. One thing separated those who kept going from those who drifted.
Not discipline. Not time. Not access to resources. Clarity about who they were becoming.
P1 said it without hesitation: “Since the start I wanted to become a product manager only.” P4 was already a designer. The identity question was settled. There was no choosing, only improving.
P2 and P3 both knew what they needed to learn. Both could describe exactly what was at stake if they failed to complete it. And both were still drifting. Because knowing what to learn and knowing who you are becoming are not the same question.
Without a fixed identity anchor, learners drift with the loudest opinion or get stuck in the noise of the market; not because they lack intelligence or effort, but because they are navigating an identity question disguised as a learning question.
Zuzanna Lito (Senior UX Researcher, CGI) offered a reframe that sharpened how I read the data: replace ‘what decision is this informing?’ with ‘what changes if I learn X vs Y?’ Participants who can answer this clearly show low abandonment. Those who cannot are already at risk before the first lesson begins.
7. The Real Finding
Self-directed learners do not struggle to start. They struggle to commit.
Starting is easy, especially when it costs nothing. A free YouTube tutorial, a 7-day trial, an open Coursera enrolment. The barrier to entry is almost zero. But commitment means closing doors. And closing doors means choosing one version of yourself over all the others you could still become.
The learners who do not abandon the learnings are not more disciplined by nature. They have fewer real choices, because they already know who they are becoming. The identity question is already answered. Distractions do not disappear. They simply become irrelevant.
“The learning decision is an identity decision in disguise.”
Most platforms are designed for learners who already know what they want. The onboarding asks: “What do you want to learn?” That assumes the decision is already made. For most learners in this study, it is not. The better question is: “Who are you becoming by choosing this over everything else?”
8. What This Means
The problem is not inside the course. The problem is before it.
Current platform design optimizes for learners’ post-decision, better content, smarter recommendations, streak reminders, completion badges. None of this address the moment before enrolment, when the learner is still figuring out whether this course belongs to the person that they are trying to become.
Three things worth changing
- Redesign the onboarding question. Replace “What do you want to learn?” with questions that surface the identity decision underneath: Why do you want to learn this? What does the bigger picture look like for you? Who are you becoming by choosing this over everything else? Learners who can answer these clearly are unlikely to abandon.
- Treat the moment of enrolment as a design opportunity, not just a friction point to eliminate. Platforms currently optimize sign-up for speed; one click, no questions, straight into the course. But a learner who commits in ten seconds has made a ten-second decision. The quality of that entry decision predicts whether they will finish. A single well-placed question at enrolment, what will be different for you when you complete this? creates a micro-commitment that post-enrolment nudges cannot manufacture. Designing for reflection before the first lesson may be more effective than any streak reminder after it.”
- Distinguish between stuck and blocked learners. A learner with no direction (P2) needs a different intervention than a learner with a clear goal but an external barrier (P1). Treating all abandonment as a motivation problem applies the wrong solution to the wrong state.
9. What I Would Do Differently
- Recruit beyond personal network. Two of four participants were personally known to the researcher. Friendship affects researcher distance: P3’s interview showed this clearly.
- Larger survey sample with targeted recruitment. 20 responses are enough to find directional patterns. It is not enough to test them. A sample of 80–100, with deliberate segmentation by field, would allow deeper comparison.
- Add a diary study component. One interview captures a moment. A two-week diary study: brief daily check-ins during active learning would capture the actual sequence of drift in real time, not in retrospect.
- Separate the survey population from the interview population. Interviewing survey respondents creates a biased sample, only the most engaged opt in. A fully independent interview recruitment would provide a cleaner comparison.
