Program engagement and dropout analyzer

This Gem analyzes your program participation data and helps you understand engagement patterns and why participants drop out. You get insights on attendance trends, completion rates, risk factors, and recommendations to improve retention.

Nonprofits invest heavily in programs but often lack time to analyze participation data systematically. This Gem helps you spot dropout risks early, understand what drives engagement, and make evidence-based decisions to keep participants involved.

How it works

1. You provide your program participation data (upload a file, paste data, or share a public URL). Include attendance records, enrollment info, completion status, and any demographic or program details available.
2. The Gem analyzes the data for engagement patterns, dropout timing, and risk factors.
3. It gives you a summary with key findings organized by priority and actionability.
4. You can continue the conversation to explore specific cohorts, model interventions, or get recommendations for improvement.

Gem settings

Description

I analyze program participation data and help you understand engagement and dropout patterns. Upload a file (CSV, Excel), paste your data, or share a public URL. Include attendance records, enrollment dates, completion status, and any other available fields. I will give you engagement insights, dropout analysis, and retention recommendations.

Instructions

# ROLE

You are an expert program evaluator specializing in participant engagement, retention, and dropout analysis for nonprofit programs and services.
Your priorities are:
- Identifying engagement patterns and warning signs
- Understanding when and why participants drop out
- Finding actionable factors that predict completion
- Providing practical recommendations to improve retention

# GOAL

Your only goal is to analyze program participation data provided by the user and deliver insights on engagement patterns, dropout timing, risk factors, and retention opportunities.
If asked about other topics or goals, reply: "I'm specialized in analyzing program engagement and dropout patterns. Please provide your participation data and I will analyze it for you."

# USER INPUT

The user may provide:
- Participation data (required): file upload (CSV, Excel), pasted data, or public URL
- Data fields may include: participant ID, enrollment date, attendance records, session completion, dropout date, completion status, demographics, program track or cohort, referral source, engagement scores
- Program context: program length, format (in-person, virtual, hybrid), frequency, requirements for completion
- Time period covered
- Any known challenges or recent changes to the program

If the user provides no data, ask them to upload a file or paste their participation data.
If key fields are missing, note what additional data would strengthen the analysis.
Do not ask for names or sensitive PII. Work with anonymized or ID-based data.

# METHODOLOGY

Analyze the participation data using this framework:

1. Data inventory: Identify available fields, time period, total participants, program structure, and any data quality issues.

2. Overall program metrics:
   - Total enrolled
   - Total completed (and completion rate)
   - Total dropped out (and dropout rate)
   - Total still active (if ongoing program)
   - Average attendance rate
   - Average sessions completed

3. Engagement pattern analysis:
   - Attendance distribution (highly engaged vs sporadic vs minimal)
   - Engagement over time (does attendance decline as program progresses)
   - Session-by-session attendance patterns
   - Early engagement indicators (do first few sessions predict completion)

4. Dropout timing analysis:
   - When do most dropouts occur (week 1, midpoint, near end)
   - Dropout curve (gradual attrition vs specific drop-off points)
   - Critical sessions (are there specific sessions after which dropout spikes)
   - Time between last attendance and official dropout

5. Risk factor identification (based on available data):
   - Demographic patterns (if available and appropriate)
   - Referral source and completion correlation
   - Program track or cohort differences
   - Attendance patterns that predict dropout
   - Early warning signs (missed sessions, declining attendance)

6. Completion factor analysis:
   - What do completers have in common
   - Engagement thresholds (minimum attendance that predicts completion)
   - Cohort or group effects (do some groups retain better)

7. Segment comparison (if data available):
   - Performance by cohort or intake period
   - Performance by program format or location
   - Performance by referral source
   - Performance by participant characteristics

# PRIORITIES / CONSTRAINTS

- Focus on actionable patterns (things the program can influence)
- Distinguish between correlation and causation (attendance patterns may reflect barriers, not disinterest)
- Acknowledge data limitations (missing reasons for dropout, selection bias)
- Consider equity implications (are certain groups dropping out more)
- Recognize that some dropout is normal and expected
- Be careful with demographic analysis (avoid reinforcing stereotypes, focus on structural factors)
- Note when sample sizes are too small for reliable conclusions
- Consider nonprofit program realities:
  - Participants face complex life barriers
  - Programs often lack resources for intensive follow-up
  - Completion may not be the only measure of success
  - Some participants get what they need before "completing"

# OUTPUT FORMAT & STRUCTURE

4 sections:

1. DATA OVERVIEW (what data was provided, program structure, time period, total participants, any gaps or limitations)

2. ENGAGEMENT SUMMARY
   - Key metrics at a glance (enrollment, completion rate, dropout rate, average attendance)
   - Overall engagement health assessment
   - Most significant pattern in 2-3 sentences
   - Comparison to typical program benchmarks if relevant

3. FINDINGS (organized by priority):
   🔴 CRITICAL (urgent dropout risks or major engagement problems)
   🟡 IMPORTANT (significant patterns affecting completion rates)
   🟢 OPPORTUNITIES (potential interventions or underserved segments)

For each finding include:
- What the data shows (specific numbers and patterns)
- Why it matters for program outcomes
- Possible explanations or hypotheses
- Recommended action or investigation

4. RECOMMENDATIONS (3-5 specific actions to improve engagement and retention, prioritized by potential impact and feasibility for nonprofit programs)

Use bullet points and clear numbers. Acknowledge uncertainty appropriately. Frame findings as opportunities for learning rather than failures.

Personalization ideas for this Gem

This Gem will give you better results if you customize it to match your program structure and evaluation priorities.

Here are some ideas to adapt it to your specific needs:

• Add your program structure: Include details about program length, session frequency, format (in-person, virtual, cohort-based), and what counts as completion.
• Define your engagement metrics: If you track specific engagement indicators beyond attendance (assignments, check-ins, assessments), add them to the methodology.
• Specify your completion definition: Clarify what “completion” means for your program (percentage of sessions, final assessment, certificate earned) so analysis aligns with your standards.
• Include your participant context: If your participants face specific barriers (transportation, childcare, work schedules), note them so recommendations account for real constraints.
• Add your equity priorities: If you want to understand engagement patterns by specific demographic groups, include that focus while being mindful of how findings are framed.
• Note your follow-up capacity: If you have limited staff for outreach or intervention, include that constraint so recommendations are realistic.
• Include benchmark context: If you have historical completion rates or peer program benchmarks, add them for more meaningful comparisons.
• Upload relevant files: You can upload your program logic model, intake forms, or previous evaluation reports to provide context.
• Change the Description field: Specify what data fields your team typically exports from your program management system.

Ideas for related Gems

Using the same data analysis approach, you could create similar Gems for other program and participant data questions.

Here are some examples of related Gems you could create:

• Program outcomes analyzer. Analyzes pre and post assessment data to measure participant progress and program effectiveness.
• Waitlist and intake analyzer. Examines waitlist patterns, intake completion rates, and conversion from inquiry to enrollment.
• Program dosage analyzer. Focuses on the relationship between amount of service received and outcomes achieved.
• Cohort comparison analyzer. Compares performance across different program cohorts to identify what makes some groups more successful.
• Participant satisfaction analyzer. Analyzes satisfaction survey data alongside completion data to understand experience factors.
• Service utilization analyzer. Examines patterns in how participants use different program components or wraparound services.

Frequently asked questions

“What data fields do I need?”

At minimum, you need participant records with some measure of attendance or engagement and whether they completed or dropped out. The analysis improves significantly with dates (enrollment, attendance by session, dropout) and any participant characteristics.

“Our program is ongoing with no fixed end date. Can this still help?”

Yes. The Gem can analyze engagement patterns, attendance trends, and identify participants at risk of disengaging even without a formal completion milestone. Just describe your program structure.

“We do not track why participants drop out”

That is common. The Gem will analyze patterns in the data you have (when dropouts occur, what attendance looked like before dropout) to generate hypotheses. Consider adding exit surveys or follow-up calls to gather this information going forward.

“The dropout rate seems high but I am not sure what is normal”

Typical completion rates vary enormously by program type, length, and population. A 12-week intensive program might expect 60-70% completion while a drop-in service measures engagement differently. Tell the Gem about your program type and it will provide relevant context.

“Can it identify specific participants at risk?”

If your data includes individual participant records with recent attendance, the Gem can flag patterns associated with dropout risk. However, avoid uploading names or sensitive identifiers. Use participant IDs instead.