Groups Analytics
Your church management system has the data. You just can't get to it. Here's how AI and a simple pipeline can turn that data into reports that actually help you care for people.
Most churches are glued to Planning Center. And for good reason. It's somehow the only worship team app that actually works and is functional. But here's the thing. The data inside PCO is a gold mine that almost nobody is using.
You know the drill. You want to know how many people are actually connected into a group. You run a list. It takes forever. The filters don't do what you need. You can't merge data across different areas. So you export to a spreadsheet and start manually sorting through rows. Your team splits the alphabet and spends hours on data entry every month.
Meanwhile, you're making decisions about group launches, leader coaching, and follow-up based on gut feeling because the data is locked inside a system that won't let you see it clearly.
It doesn't have to be this way.
What's Possible
These are real reports I run every week. They take seconds to generate. The data below is sample data from real reports I run. Names and places are changed to protect people, but the structure is exactly what's running at our church.
One row per active group. Color-coded capacity, join rate, churn, attendance rate, and a team engagement score that tells you where to focus your next coaching conversation. You can see which groups need help, which ones are ready to multiply, and which leaders need a different kind of support.
| Group | Enrolled | Capacity | Openings | Signed Up | No Shows | Join Rate | Left | Attendance | Team Eng. |
|---|---|---|---|---|---|---|---|---|---|
| Mon @ 7:30pm - Ava L. | 14 | 12 | -2 | 16 | 1 | 94% | 3 | 72% | 66.7% |
| Tue @ 7:00pm - Jordan W. | 15 | 12 | -3 | 17 | 1 | 94% | 2 | 80% | 58.3% |
| Mon @ 7:30pm - Kai N. | 11 | 12 | 1 | 18 | 0 | 100% | 7 | 55% | 66.7% |
| Wed @ 8:00pm - Drew F. | 9 | 12 | 3 | 14 | 5 | 64% | 5 | 34% | 8.3% |
| Mon @ 7:30pm - Tyler M. | 6 | 12 | 6 | 18 | 9 | 50% | 11 | 28% | 33.3% |
| Mon @ 7:30pm - Reese C. | 11 | 12 | 1 | 18 | 0 | 100% | 7 | 54% | 66.7% |
Sample data from real reports. Names and places are changed.
Drew's group has 8% team engagement. Almost no one in the group is connected to a serve team. That's a coaching conversation: "Hey, one thing that would really help your group grow is getting people plugged into a team."
Tyler has 18 sign-ups but 9 never showed. That's a different conversation entirely: "What does your welcome process look like? Are you reaching out before the first meeting?"
Ava and Kai are over capacity but healthy. High attendance, strong team engagement. These are your multiplication candidates. Time to talk about apprenticing a new leader.
Every person who's been in a group over the past three years. Attendance categories, engagement scores, baptism status, team membership, and semester-over-semester trends. Filter by group, leader, or member to see exactly where someone is and what their next step should be.
| Member | Group | Category | Last 4 Wks | Semester | Engagement | Score | Baptized | Team |
|---|---|---|---|---|---|---|---|---|
| Member #1041 | Mon @ 7:30pm - Ava L. | Consistent | 100% | 96% | Highly Engaged | 80 | Yes | Production |
| Member #1042 | Mon @ 7:30pm - Ava L. | Consistent | 75% | 88% | Engaged | 65 | Yes | Guest Services |
| Member #1043 | Mon @ 7:30pm - Ava L. | Regular | 50% | 58% | Moderate | 49 | No | -- |
| Member #1044 | Tue @ 7:00pm - Jordan W. | Rare | 0% | 17% | Low | 12 | No | -- |
| Member #1045 | Tue @ 7:00pm - Jordan W. | Consistent | 100% | 92% | Engaged | 62 | Yes | Band |
Sample data from real reports. Names are changed.
Member #1043 attends regularly but isn't baptized and isn't on a team. That's a next-step conversation a leader can have. The data surfaces it so it doesn't get missed.
Member #1044 has basically stopped coming. That's a care conversation. Without this report, that person quietly disappears. With it, a leader knows to reach out this week.
A report that runs weekly and shows every group event in the last three weeks that's missing an attendance report. Then AI drafts individual emails to each leader with a direct link to the exact event in Planning Center. Copy, paste, send. What used to take an hour takes five minutes.
| Leader | Group | Missing Date | Days Since | Link |
|---|---|---|---|---|
| Blake S. | Thu @ 7:00pm - Blake S. | Jun 12 | 11 days | PCO Event Link |
| Cameron P. | Mon @ 8:00pm - Cameron P. | Jun 9 | 14 days | PCO Event Link |
| Cameron P. | Mon @ 8:00pm - Cameron P. | Jun 16 | 7 days | PCO Event Link |
| Maya R. | Wed @ 7:00pm - Maya R. | Jun 18 | 5 days | PCO Event Link |
Sample data from real reports. Names and places are changed.
AI drafts the follow-up email for you. It sounds like this: "Hey Drew, we're catching up on June's missing attendance. When you get a chance, could you use the link below to let us know who made group that night? If you didn't meet, just put zero."
What used to take an hour now takes five minutes. No more tracking down which leaders owe you a report. The data shows you exactly who, when, and gives you a direct link to share.
See group capacity by location and group type. Our church started in Midtown. As we grew, we expanded into the surrounding neighborhoods. When we first ran this report, it told us something we felt but couldn't prove. Our open capacity was almost entirely in surrounding areas. Midtown, our main target, was full.
Location × Group Type Heatmap (Fill %)
Sample data from real reports. Locations are changed.
Midtown married groups are at 117%. Men's and women's are over 90%. That's our core area, where our church meets, and there's no room. We're either turning people away or overcrowding living rooms. The demand is there. The seats aren't.
The surrounding areas tell a different story. Uptown and West Side have groups running at 42-67% capacity. The open seats aren't where the demand is. Without this report, we would have kept launching groups in those areas because that's where we had "room."
This changed our strategy. Instead of expanding outward where we had capacity, we focused on multiplying leaders in Midtown where people actually wanted to be. The data told us to go deeper, not wider.
Net change by day of the week. Which nights are adding people? Which ones are churning through them? A churn rate over 100% means more people left than are currently on the roster. That's a structural problem, not a leadership problem.
| Day | Groups | Joined | Left | Net Change | Avg Per Group | Churn Rate |
|---|---|---|---|---|---|---|
| Monday | 20 | 89 | 64 | +25 | +1.3 | 42% |
| Tuesday | 14 | 71 | 53 | +18 | +1.3 | 48% |
| Wednesday | 7 | 22 | 49 | -27 | -3.9 | 118% |
| Thursday | 6 | 31 | 24 | +7 | +1.2 | 55% |
| Sunday | 3 | 18 | 11 | +7 | +2.3 | 38% |
Sample data from real reports.
Wednesday had a 118% churn rate. More people left than are currently on the roster. We dug deeper and found it was heavy team rehearsal night. Fewer mature believers were available. The only people in group were newer attendees without deep roots.
The fix wasn't better leaders. It was a structural change. We launched fewer groups on Wednesday and pushed toward Monday and Tuesday. The data told us it was the night, not the people.
When every leader started, when they stopped, and how long they lasted. This report surfaced a pattern we never would have found otherwise: if you can get a leader through months 9 to 15, they stay until they move away. That changed where we focus our coaching energy.
| Leader | Started | Status | Tenure | Semesters | Phase |
|---|---|---|---|---|---|
| Ava L. | Sep 2019 | Active | 6.8 yrs | 14 | Veteran |
| Jordan W. | Sep 2023 | Active | 2.8 yrs | 6 | Established |
| Quinn H. | Sep 2025 | Active | 0.8 yrs | 2 | Critical Window |
| Morgan T. | Jan 2025 | Stopped | 0.5 yrs | 1 | Didn't Return |
| Riley D. | Sep 2024 | Stopped | 0.9 yrs | 2 | Left in Window |
Sample data from real reports. Names and places are changed.
Quinn is in the critical window. Months 6 through 12 is where we focus our coaching energy now. If he gets through this season feeling confident and supported, he'll lead for years.
Morgan and Riley didn't make it. That's not failure. That's a signal. If this pattern keeps showing up, it means you need to invest earlier in the leadership journey, not just react when someone leaves.
The attendance categories, engagement scores, Maturity Index, and report formats behind the data above. No SQL. Just the decisions.
Under the Hood
You don't need to know SQL. You don't need to know how APIs work. You need someone on your team who does, and then AI handles the rest.
Describe the report you want in normal words. Here's what the Group Health Report spec looks like:
"One row per active group. Show me group name, leader, enrolled count, capacity, and openings. Add how many people signed up this semester, how many never showed, and the join rate. Show how many left. Calculate the semester attendance rate. Add a Maturity Index column. Color-code anything below 50% red, 50-74% yellow, 75%+ green."
That's it. That's the spec. But here's where it gets powerful. You get to define your own terms. I created something I call the Maturity Index. It's a simple 3-point score. One point for being in a group. One point for being on a serve team. One point for active giving. A person scoring 3/3 is fully connected. A person scoring 1/3 has a clear next step. Your church management system already tracks these things. You just need to decide what matters and name it.
If writing a spec from scratch sounds intimidating, it doesn't have to be. You can drop the prompt below into any AI tool and it will interview you through the entire decision tree. It asks the questions. You answer in plain language. By the end, you have a complete spec.
You are a church data consultant helping a groups director build a reporting spec for their small group system. Your job is to interview them, one question at a time, and produce a complete spec document by the end. Walk through the following areas in order. Ask one question, wait for their answer, then move to the next. Adapt your follow-up questions based on what they tell you. 1. CHURCH CONTEXT - What church management system do you use? - How many small groups do you have? Roughly how many people are in groups? - What types of groups do you run? (e.g., dinner groups, men's/women's, couples, interest-based, serve teams) - Do you have multiple locations or campuses? - What does your group calendar look like? Semesters, trimesters, year-round? 2. GROUP TAGGING - How do you currently tag or categorize your groups? (day, time, type, location, leader, etc.) - Are there tags you wish you had but don't? - Do you track group capacity? If so, how do you define it? 3. ATTENDANCE - Do your leaders take weekly attendance? - What does "consistent" attendance mean to you? How many times per month or semester? - How would you define these categories for your context: Consistent, Regular, Inconsistent, Rare, Inactive? - Do you track people who signed up but never attended? - Do you track people who left a group mid-semester? 4. ENGAGEMENT SCORING - Beyond attendance, what signals tell you someone is engaged? (serving, giving, event attendance, leadership roles, etc.) - If you could score engagement on a 0-100 scale, what factors would matter most? - How would you weight those factors relative to each other? 5. GROWTH PATH / MATURITY INDEX - What does spiritual growth or connection maturity look like at your church? - If you had to define 3-5 milestones someone hits as they get more connected, what would they be? - Do you currently track serving and giving in your church management system? 6. REPORTS YOU WANT - What questions do you find yourself asking about your groups that you can't easily answer today? - If you could open a spreadsheet every Monday morning and see exactly what you needed, what would be on it? - Who else on your team would use these reports? What would they need to see? 7. RED FLAGS AND ALERTS - What situations do you want to catch early? (leader burnout, group decline, people slipping away, etc.) - What thresholds would trigger a flag for you? (e.g., attendance below 50%, a member missing 3 weeks in a row) 8. DEFINITIONS AND TERMS - Are there any terms unique to your church that should be defined in the spec? (e.g., what you call a small group, how you define "active," what a semester means) After you've gathered all the answers, compile them into a clean spec document with these sections: - Church Context (summary) - Group Tagging Standards - Attendance Categories (with definitions and thresholds) - Engagement Score (factors, weights, and levels) - Maturity Index (milestones and scoring) - Report Formats (what each report contains) - Key Terms (glossary) Present the spec and ask if anything needs adjusting.
Drop your spec into Claude, ChatGPT, or Gemini along with a pre-prompt that maps your church management system's data structure. The AI writes the SQL query for you. No coding required on your end.
You need somewhere to run the code the AI writes. There are a few options. We use Google BigQuery because it's free at our scale, connects directly to Google Sheets, and your church likely already has Google Workspace. But there are other tools that do the same thing. Your technical person will know which fits best.
One click exports to a spreadsheet your whole team can read. Walk to work, wonder something, type it in, get an answer. This is where the reports live and where your team will spend most of their time.
If you're not a spreadsheet person, you don't have to be. Drop your report into any AI tool and ask it to turn the data into charts, dashboards, or whatever format helps you think. Attendance trends over time. Capacity by location. Engagement by group type. You describe what you want to see and the AI builds it.
I'm going to paste a spreadsheet export from my church's small group report. I need you to turn this data into clear, simple visuals that I can share with my team. Create the following: 1. ATTENDANCE OVERVIEW - A bar chart showing the number of members in each attendance category (Consistent, Regular, Inconsistent, Rare, Inactive, etc.) - Use green for Consistent/Regular, yellow for Inconsistent, red for Rare/Inactive - Include the count on each bar 2. GROUP HEALTH SNAPSHOT - A table or scorecard showing each group with their attendance rate, capacity percentage, and any red flags - Color-code attendance rates: green (75%+), yellow (50-74%), red (below 50%) - Sort by attendance rate, lowest first, so I can see who needs attention 3. TRENDS OVER TIME - If the data includes multiple weeks or months, show attendance trends as a line chart - Highlight any groups with a downward trend of 3+ consecutive periods 4. CAPACITY BY LOCATION - A simple bar or heatmap showing capacity utilization by location and group type - Flag anything over 90% (nearly full) or under 50% (underperforming) 5. LEADER DASHBOARD - One summary card per leader showing their group name, size, attendance rate, and how long they've been leading - Flag any leader in their first year or any leader whose group attendance dropped more than 15% Keep the design clean and simple. Use colors intentionally, not decoratively. Label everything clearly. These visuals need to make sense to someone who has never seen the raw data. After you create the visuals, give me a 3-sentence summary of the most important thing the data is telling me.
Getting Started
Planning Center, CCB, Breeze, or any system with an API. The concept is the same regardless of the platform. You're just pulling data from a different source.
The reports are only as good as the data going in. Tag your groups by day, time, type, and location. Take attendance weekly. Cancel events when you don't meet instead of marking zero.
You need a developer, IT volunteer, or tech-savvy staff member who can set up the API connection and get your data flowing into BigQuery. This is the one-time setup piece.
Claude, ChatGPT, or Gemini. Any of them work. Use a paid account. The free tiers hit limits fast and the reasoning is noticeably weaker. Both Claude and ChatGPT offer non-profit pricing. Once the pipeline is set up, you're writing plain-language prompts and getting SQL back.
Define your terms before you write a single query. Attendance categories, engagement scores, what "active" means at your church. See ours as a starting point.
I know how this sounds. Spreadsheets. SQL. Data pipelines. It sounds like the opposite of pastoral care. But here's the thing.
In Luke 15, Jesus tells the story of a shepherd who had 100 sheep and realized one was missing. He left the 99 to go find the one. But here's what we skip over: he knew he had 100. He knew when the count was 99. He had a system for knowing who was there and who wasn't.
The number of churches that don't even know how many sheep they have is concerning. Not because they don't care. They care deeply. They just don't have the capacity to manually track every person across every group every week. And when you're running reports by hand or splitting the alphabet across your team, the people who quietly slip away are the ones you never get to.
When someone attends once and then misses three weeks in a row, the data catches it. When a leader is quietly drowning because half their group stopped showing up, the data shows it. When someone's been faithfully attending for a year but hasn't taken a next step, the data surfaces it.
This isn't big brother. It's the opposite. The more you can take off your plate with automation and smart reporting, the more time you have to actually sit across the table from your leaders and members. And you'll know which conversations are the most impactful because the data already told you where to focus.
I built this for my church and I can help you figure it out for yours. Whether you're starting from scratch or trying to make sense of the data you already have, let's talk through what this could look like in your context.
Book Time with Nick