Church giving analytics turns donation records into eight numbers your offering total can’t give you: participation rate, median annual giving, concentration, first-time givers, retention, recurring share, fund mix and year-over-year change. The total can hold steady while the base under it shrinks.
Key points
- Eight metrics carry almost every giving conversation: participation rate, median annual giving, concentration, first-time givers, retention, recurring share, fund mix and year over year.
- Congregations take 81% of their funds from direct individual contributions, and 40% of them receive essentially their entire annual revenue that way, so participation and retention carry the most budget risk.
- The research literature on congregational giving consistently finds roughly 80% of funds coming from about 20% of participants. Old studies, but the right shape to check yourself against.
- Giving to religion rose 2.4% in 2025 and fell 0.2% after inflation, so a church whose total is up two points is flat, not growing.
- Planning Center Giving totals each donor for one date range at a time, so retention, year over year, and any comparison against the people you count as active all need two exports and a join done outside Giving. The single-period numbers come from one file.
The giving question a church asks itself is nearly always the same one: are we on budget. Fair question, four seconds to answer.
The trouble is that the answer is a single number attached to nobody, and almost everything worth knowing about generosity is a shape rather than a total. This post extends finding lapsed givers in Planning Center, which works one of these eight end to end. Here is the rest of the set.
Why doesn’t the offering total tell you what’s happening?
Because a total is two numbers moving independently, and it can’t tell you which one moved.
Say giving is up three percent. That can mean more households giving, which is what you were preaching for. It can equally mean the same households giving more, or a chunk of them giving nothing while your top ten cover the gap. Three churches, one identical line on the budget report.
The national picture is that ambiguity at scale. The share of Americans who give to charity at all fell from 50.9% in 2018 to 46.9% in 2020, while the average given by the households still giving rose 11.6%. Hartford’s congregational finance research shows median per capita income rising from $2,000 in 2020 to $2,222 in 2023, alongside a pattern clergy described in interviews: the number of total donors is down, but per-person donation amounts are up. Treat that second half as what practitioners report rather than a measured national trend.
Fewer donors, larger gifts, totals that look reassuring the whole way down. It’s the giving version of the problem a flat attendance total hides: departures and arrivals cancel, and the number never mentions it.
The eight giving metrics worth tracking
Each is a formula, a data source and a decision it changes. All eight can be built from Planning Center exports; the section after next covers how.
1. Giving participation rate
What it is. The share of your giving units that gave at least once in the period.
Formula. Active units with at least one gift ÷ units you consider active. The numerator is the intersection, not every donor in the export, or a non-attending giver sits above the line and not below it and the rate has no ceiling. The denominator is the whole argument: define it once, write it down, and don’t quietly change it between board meetings.
Where it comes from. The denominator is a People list of active adults or households. That’s a decision, not a lookup. The numerator is the By donor export’s keys looked up against that list, counting the matches.
The catch is the unit. Giving records a donation against a person; households are a People concept, and the only household-shaped feature here is the option to combine donors who want their giving on the same statement, set by hand, one pair at a time. A per-person rate is easy; a per-household one is manual. Pick one and say which.
2. Median annual giving
What it is. The middle of the distribution of what each donor gave over the year. Not the median gift: that is a different number, the middle of the individual donation rows, and it comes from the By donation export rather than By donor. Report annual giving per donor; the per-gift figure mostly tells you how people split their giving, not how much they give.
Formula. Median of the per-donor totals for the period. Compute the mean beside it.
Where it comes from. The By donor export, which gives information about donors who gave during a particular date range and sorts by donation totals and number of donations.
Report the median, not the average. The average is dragged up by a handful of households and describes nobody. The gap between mean and median is your concentration in one line, which makes the pair more useful than either alone.
One trap: a “donated more than $100” list rule catches single gifts over $100 and will not show people who gave multiple donations totaling more than $100. The household giving $40 a week is invisible to it.
3. Concentration
What it is. How much of the budget rests on how few households: the top ten, and the top twenty percent.
Formula. Aggregate donor totals to households first. A couple with two donor records is one household, and left as two rows they land lower in the order and split one share across two places. Then sort the household totals descending, add a running cumulative sum, divide by the grand total, and read the result at row 10 and at the row covering the top fifth of households.
Where it comes from. One By donor export and two spreadsheet columns.
There is a real finding to compare against, with an age warning attached. A University of Notre Dame review of the religious giving literature reports that studies of both large and small congregations consistently show that approximately 80 percent of the funds are given by 20 percent of the participants, citing Hoge in 1994 and Chaves and Miller in 1999, and notes the imbalance is smaller in conservative congregations.
That is a literature from the 1990s and 2000s, not a live measurement, so treat it as a shape rather than a target. What matters is your own series: whether your top decile covered 55% of the budget last year and 62% this year.
4. First-time givers
What it is. People whose first-ever gift landed in this period.
Formula. Count of donors with a first gift date inside the window.
Where it comes from. A People list, not a Giving report. Planning Center documents a donor list rule for a person’s first donation inside a date range, with the caveat that only fully processed donations count: “failed, in process, or uncommitted donations will not appear in donor lists,” which is the usual reason a list and a report disagree by a few thousand dollars.
Count them monthly. A first gift is the clearest evidence you have that somebody moved from attending your church to belonging to it.
5. Retention and lapse
What it is. The share of last period’s donors who gave again this period. Its complement is your lapsed rate.
Formula. Donors present in both windows ÷ donors in the prior window.
Where it comes from. Two By donor exports, joined on donor ID outside Giving.
Track first-year givers separately. A household whose first gift was last March and one that has given for fifteen years are not the same commitment, and a number averaged across both moves for reasons you can’t read. The full method for the lapsed half is written up separately, including the large share of any raw lapsed list that turns out to be false positives.
6. Recurring share
What it is. The proportion of giving arriving on a schedule rather than by a decision each week.
Formula. Recurring dollars ÷ total dollars. Do it by dollars and by donor count; the answers differ.
Where it comes from. Processed donations, not the forecast. Planning Center forecasts what active recurring gifts should produce, with a breakdown by fund and a CSV export, and each donor’s schedule sits on the Recurring tab of their profile. But a forecast counts schedules that were created last week, cancelled yesterday or failed on the day, none of which is money that arrived. For the share that arrived, export the donations for the window and split the rows that came from a recurring schedule from the rest; check your export for the column that marks it, since Planning Center doesn’t publish the Giving export’s column list. Quote the forecast only as a forecast, against a forecast of total giving.
This metric has the best sector evidence, and it is unflattering. The National Study of Congregations’ Economic Practices found that 55% of congregations offer recurring contributions, and 15% of those receive no gifts at all through them as of its 2017 data. The rail exists and carries nothing, which is why the share is worth measuring rather than assuming.
Adoption has moved since. Online giving went from 31% of churches in 2015 to 67% in 2023, and per capita annual giving ran $1,809 in congregations without online giving against $2,428 in those using it a lot. A correlation rather than a lever, but the strongest published argument for making a recurring gift easy to set up.
One caution: a failed recurring gift looks exactly like a decision to stop. Planning Center does not retry a failed donation on the date it failed; it attempts the next scheduled one. An expired card therefore reads as a lapsed giver for a month, and chasing those is the fastest money on this list.
7. Fund mix
What it is. The split of giving across funds, and how that split is moving.
Formula. Dollars per fund ÷ total dollars, against the same window last year.
Where it comes from. The By donation export, filtered or grouped by fund. Funds exist to track the intent of a donor’s donation, which is why a shift in mix is a shift in what people believe they are funding.
The common trap is reading a general-fund decline as a giving decline when a building campaign launched in March. Re-run without the fund filter first. And in-kind gifts sit outside ordinary donation totals in their own report, so they never appear in the numbers you take to the board unless you go looking. Planning Center added CSV export and filtering for in-kind donations in April 2025.
8. Year over year
What it is. This window against the same window last year, in dollars and in donors.
Formula. Two matched windows, two measures each. Never month against previous month.
Where it comes from. Two By donor exports with identical date boundaries.
Report it in real terms, because nominal growth in church giving has been mostly inflation for years. Hartford’s congregational finance report puts it bluntly: the median congregation’s income in 2010 was $150,000, and to have the same purchasing power in 2023 it would need over $209,000, about $45,000 above the median actually reported. The median actually reported was $165,000, the highest nominal figure in nearly fifteen years. Nationally, giving to religion reached $151.58 billion in 2025, up 2.4% in current dollars and down 0.2% after inflation, and religion’s share of American charitable giving is 23% across 2024 and 2025, against a 2011–2015 average of 33%.
Against that, a three percent nominal increase is not a win and a two percent dip is not a crisis. It’s the weather.
| Metric | Formula | Where the data is | What breaks it |
|---|---|---|---|
| Participation rate | Active units with a gift ÷ active units | By donor export keys matched against the active People list | An undeclared or shifting denominator; counting donors outside it |
| Median annual giving | Median of per-donor totals | By donor export | Reporting the mean, or calling it the median gift |
| Concentration | Cumulative % of total at top 10 and top 20% of households | By donor export, aggregated to households, then sorted descending | Treating a decades-old 80:20 finding as a target |
| First-time givers | Count with first gift in window | People list, first-donation rule | Failed and uncommitted gifts are excluded |
| Retention | Donors in both windows ÷ prior window | Two By donor exports, joined on donor ID | Mixing first-year and long-standing givers |
| Recurring share | Recurring dollars ÷ total dollars, both from processed donations in the window | Donation export, split on the recurring marker; forecast only as a forecast | Using the forecast as the numerator; failed cards reading as decisions |
| Fund mix | Dollars per fund ÷ total | By donation export, grouped by fund | A campaign launch mistaken for decline |
| Year over year | Matched windows, dollars and donors | Two By donor exports | Month-to-month comparisons; ignoring inflation |
Which of these has a benchmark you can trust?
Fewer than you would like, and saying so beats a confident number with no page behind it.
Reliable. Congregations receive 81% of funds from direct individual contributions, 78% of it given during a worship service as of that study’s 2017 data. Worth rereading next to the online-giving numbers above. Median congregational income and per capita giving are published and specific, as is the national trend for religious giving. Lifeway Research puts self-reported tithing at 51% of Protestant churchgoers giving 10% or more, with the share giving exactly a tithe down from 37% in 2017 to 31%. That is self-reported, by people who already attend monthly, so read it as direction rather than as your participation rate.
Not reliable, because nobody has published it. There is no defensible national figure for congregational giving participation rate; I looked hard, and the gap is real rather than a search failure. Concentration has the 80:20 literature, but it is decades old. Church-specific donor retention simply isn’t published.
Reliable, but about somebody else. The Fundraising Effectiveness Project reports sector-wide retention. In the first quarter of 2026, 7.1% of new donors and 25.8% of repeat donors retained year to date. Nonprofit-wide, year-to-date, and describing organisations most donors see once a year. Holding a congregation to them is a category error dressed up as a benchmark.
So build your own baseline. Compute these this quarter, write down every definition, compare against yourself next quarter. A church with two years of its own honest numbers is better governed than one quoting an average it can’t source. That’s the discipline the rest of the church metrics that matter need.
How do you build a participation and concentration report this week?
Half a day, a spreadsheet, no new software. In this order.
- Export By donor for the last twelve months. Open Donations in Giving, select the By donor tab, set the date range precisely, export to CSV. One row per donor, with a total and a gift count.
- Export By donor for the twelve months before that. Same tab, same filters, shifted back exactly one year. You need it for retention and year over year, and it is easier to pull now than to remember later.
- Export your denominator from People. Active adults, with Person ID included. Person ID is unique for each profile and it is the only join key worth trusting. Email breaks on couples sharing an address, names break on marriages and nicknames.
- Decide your giving unit and write it in cell A1. Person or household. If household, know what you are taking on: the Household ID is not saved to the profile and only exists in the current import, so the grouping is manual. You fill in a household column by surname and address, on the donor file and on the People file.
- If the unit is household, aggregate before you sort. Pivot each By donor file on the household column (SUM of total, SUM of gift count) and pivot the People file the same way, so every later step runs on one row per household. Skip this and every number below is per donor whatever A1 says: a couple with two donor records holds two rows in the concentration table and two places in the denominator.
- Sort the current-year file by total, descending. Add a cumulative-sum column, then a cumulative-percentage column dividing it by the grand total.
- Read concentration off two rows. Row 10 is your top-ten share, in whatever unit A1 says; the row covering the first fifth of donors is the top-20% share. Record both with the date.
- Compute median and mean of the totals column, side by side. The distance between them is the story.
- Compute participation. Look each current-year donor key up in the step 3 file and count only the matches, ÷ the count from step 3, both in the same unit. The raw donor row count won’t do: a donor who isn’t on the active list, a non-attending giver or someone archived since, sits in the numerator but not the denominator, and the rate drifts upward with no ceiling at 100%. The non-matches are worth a separate column of their own: people giving to a church that doesn’t count them as active. Put the denominator definition on the slide, not in a footnote.
- Compute retention with one lookup. Look each prior-year donor (or household) key up in the current-year file. Matches ÷ prior-year donor count is your retention rate; the non-matches are your lapsed list, a large share of it false positives until cleaned.
- Save the definitions beside the file. Windows, denominator, giving unit, filters. Next quarter’s number is worthless if it was computed differently, and it will be, unless the rules sit next to the numbers.
Steps 1 to 3 take twenty minutes. The rest is spreadsheet work, and if you would rather it landed somewhere live, the export-to-Google-Sheets route covers keeping it refreshed.
What can’t Planning Center Giving answer on its own?
Three things, and none of them is a missing menu item.
A household’s trend. The By donor report totals each donor for one date range, so four years side by side is four exports and a join you perform yourself. The dashboard shows giving trends and recent donations for the church as a whole. Direction without names.
A rate over time. You can list who gave for the first time in Q1, and a two-condition list will name last year’s first-time givers who gave again. What no list returns is that share tracked across years. Lists answer as of now.
Giving against attendance. The wall, and it’s architectural: Giving is a separate application from Check-Ins, so a report inside Giving can’t see whether a lapsed household is still in the building. The full inventory of what each report returns is in Planning Center giving reports.
A fourth limit is organisational. Giving’s permission levels are deliberately strict. The Reviewer role, meant for pastoral staff and board members, sees the dashboard but not reports, donation history or statements. Right default for donor privacy; it also means the people asking these questions can’t answer them, so everything routes through one bookkeeper and arrives a week late.
What belongs in front of a board, and what doesn’t
Not all eight. A board governs monthly and can’t act on detail, so give it participation rate, concentration, retention and year over year in real terms. Four numbers, each with last year beside it. Median annual giving, first-time givers, recurring share and fund mix are staff numbers: they change what you do next month rather than how the church is governed. Which metrics earn a monthly page is worked through in the health metrics your board should see.
One rule regardless of audience. If a number moves and you can’t name the households that moved it, it’s a figure to read aloud, not something to lead with. That test separates a church metrics dashboard from a wall of gauges.
Where the export route runs out
Not on the first question. An afternoon buys real answers, and for an annual budget conversation that’s a fair trade.
It runs out on the second question, and on the calendar. Somebody asks whether the fifteen households that lapsed are still attending, and you aren’t answering from the file you built. You’re starting another one. Then it’s April, and January’s numbers describe a church that has moved on.
That’s the gap Parable closes. It reads Giving beside Check-Ins, Groups and Services every night into one place where they share a person key, so participation, concentration and retention are standing numbers rather than a project, each clickable down to the households inside it. There’s SQL and warehouse access underneath for anyone who’d rather write the query.
Some giving questions genuinely fit inside Giving: how much came in, to which fund, are we on budget. Use Giving for those. It answers them well, and a second login to re-answer them would be a step backwards.
But the eight numbers above are all questions about people, and giving records are the one place in your church where people quietly tell you what they think of it every single week. Worth reading as more than a total.