Ask a pastor how many people came on Sunday and you’ll get a number in about four seconds.
Ask which families stopped coming since Christmas and you’ll get a pause, then a name or two, then something like “I’d have to look into that.”
That pause is the whole problem. Not indifference — nobody in ministry is indifferent to people leaving. The number is easy to produce and the list of names is not, so the number gets reported every week and the list never gets built.
Why the total hides the thing you need to see
Say four hundred people come on a typical Sunday. Over three months, eleven families drift out. Some weeks a few of them are still around, some weeks not.
Meanwhile a handful of new families start attending. Someone’s cousin visits twice. A small group brings friends.
Your total holds steady at four hundred.
Nothing in that number tells you eleven families are gone. The arithmetic quietly cancels them out, and the report that would have shown it — the one that lists people rather than counting them — is the report nobody has time to build.
This is why “attendance is fine” and “something feels off” can both be true at once. They’re describing different data.
Drift is a pattern, not an event
Almost nobody leaves a church on a specific Sunday. What actually happens looks like this:
A family that came three weeks out of four comes twice a month. Then once. Then they’re at a soccer tournament, then a long weekend, then it’s been six weeks and coming back would mean explaining where they’ve been.
At every step it looks like a normal gap. A vacation. A sick kid. Life. Any single absence is unremarkable — that’s exactly what makes the pattern invisible while it’s forming. You can only see it by looking at one household’s attendance over months, and no dashboard is arranged that way.
By the time it’s obvious, the conversation you needed to have three weeks in has become an awkward one four months later.
What to actually look for
If you’re going to hunt for this by hand, these are the patterns worth the effort:
Frequency change, not absence. Don’t look for people who missed last Sunday — that’s mostly noise, and the list is enormous. Look for people whose rhythm changed. Someone who came weekly and now comes monthly is a much stronger signal than someone who missed once.
A window long enough to see a trend. Compare a recent stretch against an earlier one — the last two months against the two months before, or this autumn against last autumn. A single week tells you about the weather.
Serving before attendance. Volunteers usually disengage in a specific order: they stop signing up for shifts, then they block out availability, then they stop attending. If someone quietly came off the schedule two months ago, that’s often the earliest thing you can see.
Giving as confirmation, not warning. A cancelled recurring gift is real, but it’s slow. By the time it shows up, the decision was made a while ago. Useful for confirming a pattern, too late to be the alarm.
Group membership. Someone in a small group is known by eight people who will notice. Someone attending for a year with no group can disappear without anyone having a reason to call.
Doing it by hand
You can do this with what you already have. It’s just tedious.
Export check-in records for the last six months. Pivot by person. Bucket each household by how often they attended in the first three months versus the last three. Sort by the biggest drop. Cross-reference the top of that list against your giving export and your serving schedule, because a family that stopped attending and stopped serving is a different conversation from one that’s traveling.
That’s a real afternoon of work, and the honest problem isn’t the afternoon — it’s that the answer is stale in three weeks and you have to do it again. Most churches do it once, find it valuable, and never do it a second time.
What to do with the list
Here’s where it goes wrong even when the work gets done: someone builds the list, hands it to a pastor, and the pastor doesn’t call anyone. Not from neglect — because a bare list of names with no context is genuinely hard to act on. What do you say?
Context is what makes the call possible. Before reaching out, you want to know how long they’ve been around, whether they’re in a group, whether they serve, whether anything changed in the household. A family that stopped coming when their kid aged out of the children’s ministry needs a different conversation than one that stopped after a move across town.
And then the call itself isn’t about attendance. Nobody wants to hear that a system noticed they’ve been gone. The version that works is the ordinary one — I’ve been thinking about you, how are you doing — asked by someone who actually knows them, early enough that it’s a normal question rather than an intervention.
That’s the whole point of finding the list early. Not to chase people down. To have a normal conversation at a moment when it’s still normal.
The part that’s worth automating
There’s nothing wrong with doing this by hand. The problem is that it only works if someone keeps doing it, and the person who builds the spreadsheet is usually the person with the least free time.
This is the specific job Parable does: it reads your Planning Center data every night — Check-Ins, Giving, Groups, Services — and lets you ask for that list in plain English. Who came every week last autumn and hasn’t checked in since Easter? You get the households, with each one’s giving, groups and serving history beside them, so the context arrives with the name.
Under about two hundred people you probably don’t need software for this. You’ll notice. The maths changes when there are enough people that noticing stops being possible, which is usually somewhere between two and four hundred.
But whether you use a tool or a spreadsheet, the discipline is the same: look at people, not totals; look at rhythm, not attendance; and look early enough that a phone call is still just a phone call.
Your attendance number will tell you the room is full. It will never tell you who’s missing from it.