Context

Product Hypothesis

Research Strategy

Study 1: Think-Aloud Usability Testing

Study 2: A/B Experiment

Key Findings

Tradeoff Analysis

Product Decision

Long-Term Validation (60 and 120 Days)

Reflection

Research Summary

Kardder is a social discovery platform designed to help people connect within real-world communities such as universities, clubs, and local events. Users primarily discovered one another by browsing nearby profiles and deciding whether to connect.

To encourage active participation, browsing required users to either send a connection request or permanently dismiss each profile before continuing. The assumption was that requiring a decision for every profile would increase the number of users who connected with one another.

On university campuses, however, people frequently encountered classmates, acquaintances, or mutual friends. Requiring an immediate accept-or-reject decision introduced social pressure that risked making the experience feel uncomfortable rather than empowering.

Should browsing optimize for immediate engagement or for relationships that users actually wanted to form?

The product team believed that requiring users to decide whether to connect or reject before viewing another profile would increase connection attempts and, in turn, improve relationship outcomes.

However, informal user feedback suggested that forcing decisions might discourage exploration, reduce engagement, and create social friction rather than helping people build meaningful relationships.

To evaluate these competing hypotheses, I designed a sequential mixed-methods research strategy that examined both behavioral and experiential outcomes before recommending a product decision.

To de-risk the product decision, I structured the research in three stages:

1. Qualitative usability testing (think-aloud)

I began with moderated think-aloud sessions to understand how users experienced the forced interaction model in real time. This revealed confusion, social pressure, and decision discomfort that would not have been visible through product metrics alone.

2. Controlled A/B experiment

Next, I ran a randomized experiment comparing forced and non-forced interaction models to measure behavioral outcomes at scale, including engagement, connection requests, and subsequent user interactions.

3. Longitudinal telemetry analysis

Finally, I analyzed behavioral telemetry over time to evaluate how each interaction model influenced sustained engagement and relationship formation beyond the initial decision.

Together, these studies ensured the product decision was informed by both user experience and long-term behavioral outcomes, rather than short-term engagement metrics alone.

To understand how users experienced the forced interaction model in real time, I conducted moderated think-aloud usability sessions with 11 UCLA students using a version of the app that required them to either send a connection request or permanently dismiss each profile before continuing.

Participants completed a realistic browsing task, narrated their thoughts aloud, and then took part in a short follow-up interview. This allowed me to observe not only what users did, but why they made those decisions, revealing how the forced interaction model influenced feelings of control, comfort, and social pressure.

Session recordings were systematically coded for recurring themes and independently reviewed by a second researcher to improve consistency, identifying patterns across user behaviors, verbal feedback, and post-task reflections.

The most consistent pattern was not decisiveness, but discomfort. Users often reacted with surprise when they realized they could not leave a profile without acting, describing the experience as confusing, rushed, and socially uncomfortable.

Five themes emerged consistently across participants:

Disliking forced interaction. Most participants openly disliked being required to act on a profile before leaving it.

Feeling unprepared. Users wanted more time to compare profiles, sit with uncertainty, or revisit people later instead of deciding immediately.

Fear of missed opportunities. Permanently dismissing a profile felt unusually consequential because users worried they might want to connect with that person later.

Real-world social awkwardness. On a university campus, rejecting someone did not feel anonymous. Participants worried about future in-person encounters with people they had dismissed in the app.

Avoidance behavior. Some participants said they would become less likely to browse profiles at all if every interaction required an immediate decision.

To evaluate the behavioral impact at scale, I designed a randomized between-subjects A/B experiment with 120 undergraduate participants. New users were randomly assigned to either the forced or non-forced interaction model and were unaware they were participating in an experiment, allowing behavior to remain as natural as possible.

Over a 14-day period, I measured profile views, session frequency, session duration, connection requests, and accepted connections. Results were analyzed using bootstrapped confidence intervals to estimate differences between groups and improve the robustness of the findings.

Users in the non-forced condition explored substantially more than those in the forced condition. They viewed 2.6× more profiles per day (t(26) = -2.86, p = .008, 95% CI: 2.1–3.2×) and initiated 2.1× more sessions per day (t(26) = -3.40, p = .002, 95% CI: 1.7–2.8×).

Session duration did not differ significantly between conditions, indicating that the increase reflected more frequent voluntary engagement rather than longer browsing sessions.

Together, these findings suggest that requiring users to act before continuing discouraged exploration, while the non-forced interaction model encouraged more frequent, self-directed browsing.

Removing the forced interaction requirement did not reduce connection requests. Despite exploring significantly more profiles and returning more frequently, users in the non-forced condition sent connection requests at the same overall rate as those in the forced condition.

The forced condition accounted for 48.13% of all connection requests (95% CI: 30.37%–66.18%), compared with 51.87% in the non-forced condition (95% CI: 33.82%–69.63%). The difference was not statistically significant (t(26) = -0.20, p = .85).

These findings suggest that forcing immediate decisions did not increase users' willingness to connect. Instead, the non-forced interaction model allowed users to explore first and decide when they felt ready, without reducing overall connection-seeking behavior.

Users in the non-forced condition evaluated substantially more profiles before sending a connection request. On average, they viewed 9.04 profiles per request (95% CI: 6.56–11.67), compared with 3.23 profiles in the forced condition (95% CI: 2.46–4.24). This difference was highly significant (t(26) = -4.03, p = .0004).

These findings suggest that removing the forced interaction requirement allowed users to compare more potential connections before making a decision, whereas the forced model encouraged immediate responses with less exploration.

This introduced an important product tradeoff: users became more selective in who they chose to connect with, but they also evaluated more profiles before taking action. The following analyses examine whether this shift influenced longer-term connection outcomes.

Extending this analysis to realized outcomes, both conditions also produced a comparable number of accepted connections. The forced condition averaged 21.00 accepted requests per day (95% CI: 14.50–29.50), while the non-forced condition averaged 16.38 (95% CI: 9.21–24.79). These distributions overlapped substantially and did not differ significantly (t(26) = 1.76, p = 0.096), indicating no reliable advantage for either interaction model in terms of successful connections.

This result is critical because accepted connections represent the most meaningful product outcome. Despite constraining user behavior, the forced model did not increase connection success. Instead, users in both conditions converged on similar outcomes, suggesting that connection-seeking behavior self-regulates even in the absence of enforced decisions.

The forced interaction model produced higher acceptance rates per connection request. Users in the forced condition achieved an average acceptance rate of 93% (95% CI: 0.89–0.97), compared with 84% in the non-forced condition (95% CI: 0.76–0.90). This difference was statistically significant (t(26) = 2.16, p = .040, Cohen's d = 0.77).

These findings suggest that forcing immediate decisions increased efficiency per connection request, even though both interaction models ultimately produced a similar number of accepted connections.

This revealed an important product tradeoff: the non-forced interaction model encouraged greater exploration and user autonomy, while the forced model produced more efficient individual decisions. Determining which approach better served the product required weighing efficiency against the broader user experience.

The research revealed a clear tradeoff between efficiency and user experience. The forced interaction model produced a higher acceptance rate per connection request, but it also introduced social pressure, discouraged exploration, and reduced voluntary engagement.

By contrast, the non-forced interaction model encouraged greater exploration, more frequent voluntary engagement, and greater emotional comfort, while ultimately producing the same number of accepted connections.

The key insight was that acceptance rate alone was a misleading success metric. Although the forced model appeared more efficient on a per-request basis, both interaction models ultimately produced comparable connection outcomes through fundamentally different user behaviors.

Although leadership initially favored the forced interaction model because of its higher acceptance rate per request, the broader evidence told a different story.

I reframed the discussion around three product considerations:

1. Total accepted connections, rather than acceptance rate alone.

2. User exploration and autonomy, rather than forced decision-making.

3. Long-term engagement, rather than short-term efficiency.

Together, these findings supported transitioning to a fully non-forced browsing experience. Implementing the change required updates to browse interaction logic, notification behavior, reversibility states, and interface microcopy, and was rolled out across the product.

To assess durability, I analyzed four months of post-rollout telemetry following the transition to non-forced browsing. Chat conversion increased by approximately 22%, while profile visits per session rose by roughly 18%. Importantly, acceptance rates remained stable over the same period.

These gains persisted well beyond initial rollout, indicating sustained behavioral change rather than short-term novelty effects. The decision scaled.

Replacing forced browsing with an autonomy-based interaction model increased engagement without reducing meaningful connection outcomes. Users explored more, returned more frequently, and formed the same number of successful connections, while chat conversion increased by 22% after rollout.

Beyond improving product metrics, the research changed how Kardder approached interaction design. Reversibility, clarity, and emotional comfort became guiding principles across onboarding, events, clubs, and future product decisions.

Like most early-stage product research, this work involved practical tradeoffs. Although the experiment included 120 participants and used bootstrapped confidence intervals to improve statistical robustness, the evaluation window was intentionally limited. Because the forced interaction model permanently removed profiles after each decision, extending the experiment would have introduced a confound unrelated to the interaction model itself.

The post-launch telemetry also reflected aggregate platform behavior rather than individual user trajectories, limiting the ability to determine whether engagement gains were broadly distributed across users or driven by a smaller subset of highly active participants.

Future work would benefit from longer observation periods and the ability to track individual user behavior over time. Even with these constraints, the consistency of findings across qualitative research, controlled experimentation, and long-term telemetry increased confidence in the product decision.

• Think-aloud usability testing (n = 11) with moderated interviews

• Randomized between-subjects A/B experiment (n = 120)

• Bootstrapped statistical analysis to compare behavioral outcomes

• Longitudinal telemetry analysis over a 120-day post-launch period