Five metrics carry almost any conversation about small group health: participation rate, attendance consistency, average size and the leader ratio, time from arrival to first group, and whether group members serve. Planning Center’s Groups reports answer the first three. They cannot answer the last two.
Key points
- Planning Center’s multi-group Reports page gives average group size, unique members, a church participation percentage and a turnover rate. Good trend lines, but the help article names the participation stat without defining its denominator.
- The per-member attendance percentage does not count canceled events or events with no attendance submitted, so a group whose leader skips half the weeks scores higher than one whose leader is diligent.
- Group Roles counts a person twice if they lead one group and belong to another, which Planning Center says plainly on the same page. So leader counts are roles, not people.
- Every group membership carries a
joined_attimestamp, but nothing in Groups compares it to when the person first showed up, so time-to-group is a join you build yourself. - Whether group members serve, give and stay more than non-members is a cross-product question, and a Planning Center report only ever sees its own product’s data. So Groups cannot answer the question groups exist to raise.
Somebody says “we’ve got forty groups running”, and the room takes that as a health report. Forty groups is an inventory count. It says nothing about how many people are in them, whether those people turn up, or whether the person who joined in February is still there in July.
The five below are the ones I’ve seen change a decision. Each has a formula, a place the data actually sits, and, where one exists, an honest word about what “good” looks like. They’re the groups slice of a wider set; the metrics that matter beyond Sunday headcount covers the rest.
Which small group metrics matter?
| Metric | The question it answers | Formula | Where the data sits | In Groups reports? |
|---|---|---|---|---|
| Participation rate | How much of the church is in a group at all? | Unique group members who are also in your adult cohort ÷ adults you count as yours | Groups member export intersected with a People list on Person ID | Partly, as “church participation percentage”, on Groups’ own population |
| Attendance consistency | Do the people in groups actually turn up? | Meetings attended ÷ meetings held, per member | Each group’s Attendance tab | Yes |
| Attendance submission rate | Are leaders recording anything at all? | Events with attendance submitted ÷ events scheduled | Attendance tab; grayed-out columns | No |
| Size and leader ratio | Are groups the right shape, and who is carrying them? | Average members per group; members ÷ leaders | Reports page: Group Size, Group Roles | Yes |
| Time-to-group | How long from first Sunday to first group? | Share of a newcomer cohort with a group within 180 days; median of joined_at − first check-in among joiners |
Groups API and Check-Ins, joined by person | No |
| Groups-to-serving | Does being in a group lead anywhere? | Share of group members who served in the last 90 days | Groups and Services, joined by person | No |
Six rows for five metrics, because attendance consistency has a prerequisite most churches skip.
How do you calculate small group participation rate?
Participation rate is the unique people in a group who are also in the adult population you consider your church, divided by that population, measured over the same window. The intersection matters: a roster carries children, people who moved away and members who haven’t been seen since spring, and counting them above the line but not below it gives a rate that can pass 100%. It’s the single number that tells a board whether groups are a ministry or a programme.
This is the one metric here with a real benchmark behind it. Lifeway Research’s 2022 phone survey of 1,000 Protestant pastors found churches report, on average, 44% of weekend worship attendees involved in a small group, Sunday School or similar group, down from 49% in 2010. A separate 2024 survey of groups ministry leaders landed on the same 44%.
The distribution is the more useful half. Fewer than 1 in 5 churches (19%) said 75% or more of their attendees were in a group, while 24% said fewer than a quarter were, up from 17% in 2008. Knowing which end of that spread you’re on is the point of measuring.
Planning Center gives you a version of the number for free. Filter the multi-group Reports page by group type, campus, status or tags, and the Health Stats section returns the number of groups in the filter, the average group size, unique members, a church participation percentage and a turnover rate.
That percentage is useful as a trend and harder to defend as a figure: the help article names the stat without saying what sits underneath the division, and the denominator is where all the disagreement in the room lives. Your database probably holds every child, every guest who filled in a card in 2019, and every household that moved away without telling anyone. Divide by that and the rate reads low every month in a way nobody can explain.
Pick the denominator on purpose
Three defensible choices, in ascending order of flattery:
Every active profile. Easy to audit, almost always too big. Planning Center lets you define membership types however your church defines membership, so “active” means whatever you set it to mean, and most churches set it generously.
Adults with a check-in in the last ninety days. My preference. It counts people who are actually around, and anyone else on staff can reproduce it.
Average weekly worship attendance. The number your board already knows, which makes the ratio instantly legible, and the closest match to what Lifeway’s 44% is built on. Note their denominator is all weekend attendees, not adults only, so an adults-only rate isn’t directly comparable to that benchmark. The catch is that it’s a headcount while the numerator is a roster, so the two aren’t quite the same population.
Whichever you pick, write it down beside the number and don’t change it. A participation rate that moved because someone redefined the denominator is worse than none at all. That discipline applies to every figure on a church metrics dashboard, and it’s why participation rate earns a line on the monthly board report: the definition is part of the metric.
Count adults on both sides, too. Kids in a children’s ministry group belong in a different number. Mixing them in makes the rate look healthy in exactly the churches whose adults are least connected.
Attendance consistency, and the number that flatters you
Each group’s Attendance tab is the best per-person view Planning Center ships. The percentage column shows each member’s attendance relative to the number of events in the timeframe, with a green check per event attended and a “Visited” tag on anyone on the roster who isn’t a member yet. It exports to CSV, and Print View is capped at roughly eight dates per sheet.
Across groups, the Attendance tab on the Reports page gives you the three columns worth watching: average percentage attended, the median number of typical attendees, and the number of events held. Since March 2025 that graph exports to CSV by email, with member and visitor counts separated so you can see what’s driving the total.
Underneath, each record is simply whether the person attended plus their role at the time: member, leader, visitor or applicant, which is why the visitor half of that CSV is trustworthy.
Now the catch. Planning Center states that canceled events, and events with no attendance submitted, are not counted in the percentage. Read that twice. Take two hypothetical groups of eight that both met twenty times. The first leader submits every week and averages three people present: 38%. The second submits only on the six weeks the room was full: 90%, on a quarter of the data. The diligent leader looks like the problem.
So attendance consistency has a prerequisite metric: the share of scheduled events with attendance actually submitted. On the Attendance tab those are the grayed-out columns: countable by eye for one group, tedious across forty. There’s no standard for this either, but my working line is 80%. Below that, the percentages are measuring leader diligence rather than group health, and no analysis downstream will fix it.
Fixing it is logistics rather than data. Leaders and members you’ve delegated to can take attendance in Church Center from an hour before the event, and the permission can be handed to a member of the group. The groups that report reliably are the ones where the leader isn’t the only person who can.
Having this data at all is less common than you’d think: in Lifeway’s 2024 survey of groups ministry leaders, only 53% track attendance in their adult Bible study groups and 56% keep a roster of who is in each group. If you have a percentage to be annoyed by, you’re ahead of half the field.
Group size and the leader ratio
The Reports page carries both halves. Health Stats gives average group size; the Group Size section lists the total number of people in each group, which is more useful, because an average of nine can be four groups of four and one of twenty-nine.
Group Roles gives leaders and members, with a caveat Planning Center prints right there: someone who leads one group and belongs to another is counted twice, so Group Roles won’t match unique members. Those are counts of roles. For a headcount of humans who lead anything you need the Groups API’s membership records, where each membership is member or leader, deduplicated by person.
There is no benchmark, and you should stop looking for one
The number everyone quotes is twelve, plus or minus four, and Lifeway is careful about how it frames it: in general, Sunday School and small group leaders agree the right size for a group is 12, plus or minus four. That’s practitioner consensus, honestly labelled. Not a survey finding, and not a study tying group size to retention or to anything else measurable. I went looking for that study and didn’t find one I’d put in front of a board. The practice doesn’t follow the advice either: 90% of churches don’t limit the size of their ongoing groups.
The absence is worth saying out loud, because the alternative is quoting a vendor blog that quotes another vendor blog. Build the comparison inside your own church instead:
- Bucket your groups into size bands: 1–5, 6–9, 10–14, 15 and up.
- For each band, take the average attendance percentage over twelve months.
- Look at which bands hold their percentage and which decay.
You’ll usually find one band that holds and one that quietly leaks. That’s your benchmark, it’s about your church, and it’s defensible in a way that a number from a blog post never is.
The leader ratio is the same story: no published standard, and a real constraint. What’s worth watching is the trend and the tail: how many groups have exactly one leader and no apprentice, because those are the groups that end when one family’s circumstances change.
How long does it take a new person to get into a group?
This is the metric I’d fight to keep if a board made me drop four of the five. Somebody can attend for a year, like it, and never be known by anyone. Group membership is the point at which eight people would notice if they stopped coming, which is why who stopped coming and who never got connected are two versions of the same question.
Time-to-group is the gap between someone’s first check-in and their first group membership. Every membership carries a joined_at timestamp; the first-check-in date lives in Check-Ins, and getting it recorded reliably is the same discipline as first-time guest tracking. Nothing in Planning Center subtracts one from the other. When you do, report it the way any conversion metric has to be reported: the share of a newcomer cohort that joined a group at all within a fixed horizon, say 180 days, and then the median days among those who joined. The median on its own flatters the wrong cohort: three of forty joining quickly beats thirty joining inside three months, because the thirty-seven who never joined have no joined_at and drop out of the middle.
What you can do instead is the standing version of the question: who’s around and not in a group. A People list takes conditions from more than one product, and the common Groups lists Planning Center documents cover active group members, people who joined a group in a time frame, people who applied and are still waiting, and members who missed events in a window. Add a Check-Ins condition to a “not in an active group” rule and you have your unconnected-attenders list, refreshed nightly.
The limitation is the one that shows up everywhere in Planning Center reporting: a list takes dated conditions but keeps no history. It tells you who is unconnected today. It won’t tell you whether your average time-to-group improved after you changed how groups launch in September. That needs two dates per person and a subtraction. Our guide to what each Planning Center report returns walks the same boundary product by product.
One thing shapes that data more than people expect. There are three ways someone joins a group: a Church Center link or QR code, an email invitation, or an administrator adding an existing profile. Only the last is immediate. The other two need the person to act and a leader to approve. Unapproved requests are people who tried to connect and got no answer, which makes “applied to join” the list nobody looks at and the one I’d check first.
Do people in groups serve, give and stay?
This is the question that makes the case for your groups ministry, and Groups cannot answer it. Not because of a missing feature. A report in Planning Center reads its own product’s data, and group membership sits in Groups while serving sits in Services and giving sits in Giving.
The national research points one way. Lifeway’s 2025 survey of 2,130 Protestant churchgoers found that people attending a small group five or more times a month score 74.1 on its discipleship assessment against 60.4 for churchgoers in no group at all, and that those same frequent attenders report 9.1 worship services in a typical month against 3.9. Lifeway’s church-growth analysis also puts small group discipleship among four factors predicting whether a church grew its worship attendance.
That’s correlation and self-report, and people in a group five times a month were already the most engaged people in the building. It tells you the relationship exists across a few thousand churchgoers. It doesn’t tell you it holds in yours, which is the only version a board can act on.
Measured locally, the pipeline is groups to serving: the share of group members who served at least once in the last ninety days, against the same share for attenders who aren’t in a group. If group membership does what you believe it does, the first number is meaningfully higher and you have evidence rather than conviction. If the two match, your groups are a social good rather than a discipleship pathway. A finding worth having.
It runs the other way too. Serving is usually the earliest place disengagement shows, because volunteers stop being available before they stop attending, so a group member coming off the schedule is an early signal about the group as much as the person. How to measure volunteer engagement covers that side.
How do you build these from Planning Center exports?
Two of the five need a spreadsheet. Here’s the whole job, tedious parts included. Budget two hours the first time.
- Export the groups list. Since November 2025 admins and group type managers can export it as CSV with the filters and columns they choose. This is your inventory: group, type, campus, size.
- Export multi-group event attendance. On the Reports page, filter, then use CSV Export on the event attendance graph; the file arrives by email, member and visitor counts separated.
- Export the members. The per-group Attendance tab exports one group at a time. For anything larger, build a People list of active group members from the common Groups lists recipes and export that.
- Export your denominator. A People list of adults with a check-in in the last ninety days. Save the list. Next quarter you want to re-run the rule, not rebuild it from memory.
- Join on Person ID, never on name. Two Sarah Johnsons will corrupt every number downstream. Person ID survives export; household ID does not, which is why household roll-ups get rebuilt by hand.
- Compute participation rate: distinct Person IDs that appear in both files, divided by distinct Person IDs in the denominator file. The intersection, not the whole roster. A roster carries people who haven’t checked in for months, and dividing all of them by recent attenders inflates the rate, past 100% in a stale enough group. The roster members who aren’t in the denominator are their own list: people a group claims but the building hasn’t seen. Write the denominator definition into the cell beside it.
- Compute submission rate before you trust anything: events with attendance submitted divided by events scheduled, per group. Sort ascending and read the bottom ten. That’s a leader-support list, not a data problem.
- Bucket groups by size band and average the attendance percentage within each band, as above.
- Date-stamp the file and keep it. One quarter is a number; four quarters is a trend, and a trend is the only thing a board can act on.
If you’d rather have this refresh itself, exporting Planning Center data into Google Sheets covers that route and where it stops working.
What the export route costs
Step 9 is where this dies. Not because the work is hard. It has to happen again next quarter, and the person who built it in March is the person with the least free time in June. A participation rate from two quarters ago is trivia.
And it can’t take a follow-up question. Somebody asks whether the people who left groups also stopped giving, and that’s a fresh export, a fresh join, and next month’s problem. This is the general shape of the reports Planning Center can’t run, not a quirk of Groups.
That’s the specific job Parable does: it reads every Planning Center product nightly into one place where they share a person key, so group membership sits beside attendance, giving and serving for the same household. Time-to-group, groups-to-serving and participation rate become questions you ask rather than files you assemble, and they land on a dashboard that’s current on Monday morning, with SQL and warehouse access underneath.
If your groups ministry is small enough that you know every leader by name, none of this needs software. You’ll notice the group that stopped meeting.
Past a few dozen groups, noticing stops working, and the thing you lose first isn’t attendance data. It’s the person who came for nine months, liked it, never joined anything, and left without anyone having a reason to call.