Background
In Spring 2026, two Google Forms surveys were distributed to pilot participants — one for parents, one for staff. The goal, as far as we can determine, was to gauge general satisfaction with the NHA App during its mid-pilot phase and identify areas for improvement before a full network rollout.
Both surveys were anonymous. Response rates were low. And the data that came back, while well-intentioned, did not give us anything specific enough to prioritize development work, identify root causes of user frustration, or follow up with individuals who reported problems.
Below is a complete reconstruction of each survey, followed by our analysis.
Parent Mid-Pilot Experience Survey
The parent survey had 11 questions covering onboarding, feature ratings, SPARK, in-app payments, general satisfaction, and three open-ended prompts. An optional contact field was included at the end.
| 1 | 2 | 3 | 4 | 5 | N/A | |
|---|---|---|---|---|---|---|
| Messaging your child's teacher | ||||||
| Reading school announcements/posts | ||||||
| Viewing the school calendar |
| Strongly Disagree | Disagree | Neutral | Agree | Strongly Agree | |
|---|---|---|---|---|---|
| The app is easy to navigate. | |||||
| I can quickly find the information I need. | |||||
| Notifications are timely and helpful. | |||||
| The app feels simpler than SchoolConnect. |
Analysis: Parent Survey
No clear objective. The survey tries to cover everything — onboarding, feature satisfaction, SPARK, payments, navigation, notifications, and SchoolConnect replacement readiness — in 10 questions. Without a defined goal, the data scatters across too many dimensions to draw meaningful conclusions from any one of them.
Questions 7, 8, and 9 ask the same thing three ways. "Best thing," "one thing we should change," and "one thing that would most improve" — these overlap heavily. Open-ended questions are expensive (low response quality, high analysis effort). Using three of them with subtle wording differences dilutes the signal from all three.
Question 5: "Notifications are timely and helpful." This bundles two distinct concepts. A notification can be timely but unhelpful (irrelevant content), or helpful but delayed. When a parent rates this a 2, we don't know which problem they're reporting.
Question 6: "I would feel..." Asking parents whether they'd feel "comfortable" replacing SchoolConnect measures anxiety, not product quality. A parent could be uncomfortable with change itself regardless of whether the NHA App is better. This tells us about change management, not about the app.
Questions 3 and 4 (SPARK, Payments): These are feature-awareness checks embedded in a satisfaction survey. They don't help us improve either feature — they just tell us whether people found them. Discovery and usability are different research questions requiring different instruments.
The one useful signal from this survey is Question 2 — the experience rating matrix for messaging, posts, and calendar. Even here, the data is limited because a low rating doesn't explain why the experience was poor. "I rated messaging a 2" could mean the feature is broken, the teacher doesn't respond, or the parent couldn't find it.
Staff Mid-Pilot Pulse Survey
The staff survey had 12 questions covering general sentiment, feature ratings, parent adoption, training, support, and readiness to replace SchoolConnect.
| Poor | Fair | Good | Very Good | Excellent | |
|---|---|---|---|---|---|
| Direct Messaging | |||||
| Posting to your class/school | |||||
| SPARK | |||||
| Calendar |
| Definitely Not | Kinda | Neutral | Probably | Definitely | |
|---|---|---|---|---|---|
| I felt adequately trained before go live. | |||||
| Training materials are easy to access. | |||||
| I know where to go for help. |
Analysis: Staff Survey
Question 1 provides a quick temperature check. "Love it / Getting there / Frustrated" captures overall sentiment but doesn't explain the cause. A teacher who selected "Getting there" because messaging is unreliable and one who selected it because they haven't had time to explore need different interventions. The follow-up ("Why?") is where the real insight lives, but open-ended answers from anonymous respondents are harder to act on at scale.
The training section (Q5) uses an informal scale. "Definitely Not / Kinda / Neutral / Probably / Definitely" is approachable language, but it makes it difficult to benchmark against standard survey instruments or compare results across future administrations.
Question 4 asks staff to assess parent behavior. A teacher's perception of whether parents have "switched" is filtered through their own experience. Some teachers may not notice because they don't monitor SchoolConnect anymore. Others may overestimate the problem because a few vocal parents are the ones they hear from. This is secondhand data presented as a direct measurement.
Question 6 missing an option: "Yes, very responsive" / "They respond, but the issue is still there" / "I haven't needed support yet" — there is no option for "I contacted support and didn't get a response." The question assumes support always responds. Respondents who were ignored have no accurate choice.
Questions 9-10 repeat the SchoolConnect readiness theme from the parent survey. Asking whether the app is "ready" is a policy question, not a product question. The answer depends on the school, the staff member's role, which features they rely on, and what "ready" means to them. It produces a number that feels useful but isn't.
Question 7 ("#1 technical issue") and Question 8 ("change ONE thing") are the most useful questions on this survey. They're direct, specific, and constrained to one item. If this survey had been non-anonymous with 10x the respondents, these two questions alone would have been more valuable than the rest combined.
Methodology Problems
Beyond the individual question issues, the survey program had structural problems that limited the value of any data it could produce.
- Anonymous by default. Both surveys were anonymous Google Forms. When someone reports "messaging doesn't work," we cannot follow up to ask what device they're on, what they tried, or what happened. We cannot close the loop by telling them when the issue is fixed. This limits our ability to resolve individual issues.
- Low response rate, unknown sample. Response counts were low. We do not know whether respondents represent the broader pilot population or a self-selected group of the most engaged (or most frustrated) users. Statistical confidence is not possible with the data we have.
- No segmentation. We cannot break results down by school, grade level, device type, or how long the parent has been using the app. A parent who downloaded the app yesterday and a parent who has been active for four weeks are weighted equally.
- No baseline. Neither survey was administered before the pilot began. Without a pre-pilot measurement, we cannot determine whether satisfaction improved, declined, or stayed the same over the pilot period.
- Unclear goal. Were we trying to measure satisfaction? Identify bugs? Assess readiness for rollout? Evaluate training effectiveness? All of the above? A survey that tries to answer every question answers none of them well.
- No path from data to action. Even with perfect response rates, most of these questions don't map to a specific development decision. "The app is easy to navigate: Agree/Disagree" — if 40% disagree, what do we build? The question doesn't tell us what's hard to navigate or for whom.
What the Data Showed
Across both surveys, 114 responses came back — 47 from parents and 67 from staff — representing three pilot schools. Even with the structural limitations described above, the numbers tell a story.
Despite the methodology issues, the surveys surfaced genuine problems. The top issues — broken session persistence, missing push notifications, inability to search parents by student name — were specific and actionable. These came primarily from the constrained open-ended questions and the issue checklists, not from the satisfaction scales.
The Likert ratings produced numbers (e.g., 47% negative on notifications) but didn't explain the cause. The prioritized bug lists were the most valuable output of the entire exercise.
These findings directly shaped our new survey design. The three functions that broke down most visibly — messaging, posts, and notifications — became the sole focus of the new instrument. And the inability to follow up with the 47% who reported missing notifications, or the 15% who couldn't log in without help, reinforced the value of authenticated responses — being able to follow up with specific users when issues are identified or resolved.
The Verdict
The first surveys gave us a useful starting point and surfaced real issues that shaped development priorities. Building on those lessons, the next round is designed to go deeper — focused on the three core communication functions, authenticated so we can follow up with respondents directly, and structured to produce data that maps to specific development decisions.