Evan Dunkel

Medication is a daily tracking concept built around one question: “Have I recorded this already?” I separated the planned reminder from the recorded action and made correction part of the normal flow. The intended outcome is a history someone can understand when they return to an interrupted task.

Role: Independent product designer. Team: Solo concept. Timeline: October 2026. Platform: Mobile. Status: Concept; not shipped.

I owned the problem framing, the comparison of directions, and the interaction specification. This is a personal tracking concept, not a clinically validated product; no adherence or health outcomes are claimed. The photos in this case study are placeholders.

The scenario begins when someone dismisses a medication reminder while busy, then returns later unsure what they recorded. A calendar shows the schedule, but the immediate question concerns today’s history. Treating the dismissed reminder as a completed action would erase an important distinction.

The starting brief was to build a more engaging reminder. I reframed it as helping someone distinguish “scheduled” from “recorded.” More notifications could increase activity without resolving uncertainty about the last action.

This scenario has not been validated through research. I would first observe how people revisit a log after an interruption and ask them to explain the difference between a reminder and an entry. The product question is whether the record reduces ambiguity, not whether it encourages more notification opens.

Constraints: Personal tracking only; no dosing recommendations; intermittent connectivity. Principles: Clarity over engagement; correction over perfect streaks.

I compared three directions based on how quickly someone could find the last entry and recover from a mistake.

Weekly calendar: Useful for reviewing a pattern over time, but gives the immediate question less prominence. It asks someone to locate today before they can inspect what was recorded.

Habit streak: The more ambitious engagement direction. It rewards consistency, but risks making a correction or a missed entry feel like a failure rather than an ordinary part of keeping an accurate log.

Today and history: My chosen direction. The reminder and latest record have separate places, so a scheduled event does not appear to be a confirmed entry. The trade-off is a narrower first view, with longer-term history taking a secondary role.

I would test that ordering before expanding the calendar. More information should earn its place by helping someone answer the immediate question.

I gave the latest recorded action a timestamp and a plain-language status. The schedule remains nearby, but dismissing a reminder does not change the log. Recording an action needs an explicit step that produces an entry someone can revisit.

This is less automatic than treating a notification response as completion. I accepted that friction because a convenient shortcut would otherwise leave the person unsure which action the app had recorded.

The wording must stay specific: a record describes what someone entered, not an independent confirmation of what they took. Before a pilot, I would review that distinction with a clinical adviser and test it with interrupted tasks.

Every entry has an edit action, with immediate undo as an additional shortcut. I rejected swipe-only correction because it hides a basic recovery path. Someone should not need to discover a gesture to fix a mistaken record.

An amended entry would show its updated time and an “Edited” label while retaining the original value in history. That introduces more detail than simply replacing the timestamp, but helps explain why the record changed.

The difficult case is a correction made on one device before another has synced. I would resolve that behavior with engineering before promising a seamless history across devices. Accuracy of the record matters more than making every save appear instantaneous.

Relay