# Usability study protocol v2: compact day view for repeated logging

Version 2, 2026-10-07. Revised after a synthetic rehearsal of v1. Every change is listed with its reason in
`CHANGES.md` (C1–C10). Case 3 of the UX research portfolio. It follows case 2,
`../MyFitnessPal Redesign Study`.

## 1. Brief

**Background.** In case 2, public reviews of MyFitnessPal's April 2026 diary redesign pointed to
design costs in repeated logging, not a learning curve. Reviewers described four costs: more
navigation for common tasks, losing the day at a glance, removed copy, move, and bulk-edit
operations, and a daily budget that is harder to inspect. Accidental "same as yesterday" copies
persisted for months.

**Decision this study supports.** Should the team build an optional compact day view for
established loggers?

**What reviews could not answer.** Whether the costs are real for established users when measured
on tasks, and whether a compact view reduces them without hurting comprehension.

## 2. Research questions and hypotheses

| RQ | Question | Hypothesis | Measures |
|---|---|---|---|
| RQ1 | Does the compact view make reviewing the day faster? | H1: fewer taps and less time to report a meal's items and total | taps, time, success |
| RQ2 | Can users find calories remaining after exercise? | H2: higher accuracy in compact (budget shown in one line) | answer accuracy, taps |
| RQ3 | Are copy and move operations easier? | H3: fewer taps and higher success in compact | taps, success, errors |
| RQ4 | Do accidental copies decrease? | H4: fewer unintended copies in compact | accidental copy events |
| RQ5 | Does density hurt comprehension? (guardrail) | No hypothesis. A drop in accuracy or ease would block the change | comprehension questions, ease (SEQ) |

## 3. Design

- Moderated, remote, 60 minutes, on the participant's own phone through a prototype link.
- Within subjects: each participant uses both conditions. The order is counterbalanced (A→B, B→A),
  with half the participants in each.
- **Condition A, baseline.** The current Today structure: a calorie card without exercise, a macros card,
  and a Diary below other cards with collapsed meal cards. Per-meal detail opens on tap. Copying is a
  "Same as yesterday" shortcut on each meal card. There is no multi-select, and exercise appears only
  in the full-day view.
- **Condition B, compact.** Every meal expanded with items and per-meal totals, the budget in one line
  (goal − food + exercise = remaining), multi-select with copy, move, and delete, and copying yesterday
  moved to the meal menu with undo.
- The same seeded food log is used in both conditions. The day structure is matched, so participants do
  not answer from memory.
- Sample: 12 participants. This is enough to see large differences in task time and errors and to
  surface usability issues. It is not enough for precise effect sizes.

## 4. Participants and screener

**Include:**
- Uses MyFitnessPal (any tier) for 2 or more years.
- Logs food 5 or more days a week in the past month.
- Uses the mobile app as the main logging device.

**Exclude:**
- Works in UX, product, or market research.
- Works for MyFitnessPal or a competitor.
- Took part in a study in the past 3 months.

**Mix:** iOS and Android; at least 4 who log 4 or more meals a day; at least 4 who use copy features weekly.

Screener questions:
1. How long have you used MyFitnessPal? (less than 1 year / 1–2 / 2–5 / 5+)
2. In the past month, on how many days a week did you log food? (0–2 / 3–4 / 5–7)
3. Where do you log most often? (phone app / website / both)
4. Which phone do you use? (iPhone / Android / other)
5. How many meals or snacks do you usually log a day? (1–2 / 3 / 4+)
6. How often do you copy meals from a previous day? (never / monthly / weekly / daily)
7. Do you or anyone in your household work in UX, product, market research, or for a nutrition or
   fitness app company? (yes → exclude)

## 5. Tasks

Each task begins from the condition's home screen. Ease is rated (SEQ, 1–7) after each task. The answer keys
are per dataset (C5).

| # | Scenario (read aloud) | Success | Codes from case 2 |
|---|---|---|---|
| T1 | "Tell me everything you've logged today, and how many calories each meal came to." (C1) | All items and every meal total correct | glance, navigation |
| T2 | "How many calories do you have left today?" Then the probe: "Does that include your workout?" (C3) | Correct remaining including exercise. Unprompted and prompted correction coded separately | budget |
| T3 | "Yesterday's breakfast was the same as today's will be. Copy it into today." | Today's breakfast equals yesterday's in the final state. Verification and confidence coded (C6) | removed functions |
| T4 | "The [item] at lunch was actually a snack. Fix the log so it's only under Snacks." | The item appears once, under Snacks, with its original serving, and not in Lunch. "Success with workaround" coded (C4) | removed functions |
| T5 | "Check what you had for dinner yesterday. Don't change anything." | Names yesterday's dinner items, and the log is unchanged | accidental copies |
| T6 | "You just copied the wrong meal into today. Put today back the way it was." (C2) | Today's log matches its state before the wrong copy | accidental copies, removed functions |

## 6. Moderator guide

1. **Welcome (3 min).** Purpose, recording consent, "we're testing the design, not you", think aloud. "For
   today, imagine it is Thursday, October 8." (C7)
2. **Warm-up (5 min).** "Walk me through how you logged food yesterday." "What do you look at first when
   you open the app?" "What, if anything, has changed for you in the app this year?"
3. **Condition 1 tasks (17 min).** Read each scenario. Prompts: "What are you looking for?", "What did you
   expect to happen?" Do not help. If stuck for 2 minutes, mark failed and move on. SEQ after each. After T2,
   probe "Does that include your workout?" (C3). Observation codes: hesitation near data-changing controls,
   tap and undo, unnoticed copy (C8), and whether the participant verifies a copy (C6).
4. **Condition 2 tasks (17 min).** Same tasks, matched data.
5. **Comparison (6 min).** "Which would you rather use every day, and why?" "Was anything missing or harder
   to understand in the second version?" Comprehension check on the compact view: "What does the number
   in blue at the top mean?"
6. **Wrap (2 min).** Thanks and incentive.

## 7. Measures and logging

The prototype logs every tap with a timestamp, the task, and the condition. It derives the following
automatically:

- **Effort per task:** taps + scrolls + back navigations, from task start to answer (C9).
- **Time on task:** from scenario read to answer.
- **Success:** per the criteria in section 5.
- **Errors:** actions that change the log in a way the task did not ask for.
- **Accidental copy:** any copy action outside T3.
- **Ease:** SEQ, 1–7.
- **Moderator codes:** unprompted versus prompted correction (T2), verification and confidence (T3),
  success with workaround (T4), and accidental-copy observations (C6, C8).

## 8. Before fielding (C10)

Verify the baseline condition against the live app on the study date: the home summary's formula (does
"left" include exercise?), meal-card collapse, and the copy shortcut's placement. Document this in the
condition spec. If production differs, change condition A to match before recruiting.

## 9. Analysis plan

- Per task, paired comparison of A and B: median difference in effort and time, success counts, and
  accidental copies. Report descriptively with exact counts. With n = 12, no significance claims are made.
- Think-aloud and comparison answers are coded with the case 2 categories (glance, navigation, budget,
  removed functions, accidental copies) plus a comprehension category. Each issue is rated for severity
  0–4 (Nielsen).
- **Decision rule.** Recommend building the compact view if B reduces taps or time on T1, T3, and T4,
  improves T2 accuracy, and lowers accidental copies, with no drop in comprehension or ease. Mixed
  results go to a targeted follow-up, not a build.