What this is, and what it is not
This page describes the production data behind Splitright, our own expense-splitting app, read on 2 September 2026: 119 groups, 407 member slots, 742 expenses, built up by roughly 30 real users since the app launched. That is a small number. It is not a representative sample of how people split expenses in general, it is not a market survey, and nobody should quote these figures as if they described the whole category. Treat everything below as a set of observations about one app's early users, not a statistic about the world.
We are publishing it anyway because the shape of the data is more interesting, and more useful to anyone who splits costs with other people, than the sales pitch we would otherwise be writing. Every number here is aggregate: there are no group names, codes, member names or expense descriptions in what follows, and none have been invented.
A “member” is a name, not a person
The single most important structural fact about this data is one that is easy to misread. In Splitright, a group member is a name that somebody typed in when adding a person to a group — it does not mean that person installed the app, opened it, or even knows a group exists for them. The 407 member slots in our data correspond to roughly 30 actual users.
So the ordinary shape of a Splitright group is one person, the organiser, doing the bookkeeping for three or four other people who never touch the app at all. That is not a flaw in the data; it is exactly how the app is meant to work — no account needed, no sign-up required of anyone but the organiser — but it means every “members per group” figure below counts named participants, not installs.
Group size: mean 3.42, median 3
Across all 119 groups, the mean number of members is 3.42 and the median is 3. Most groups are a small handful of people, which fits the app's obvious use cases: a dinner, a weekend away, a shared flat.
| Measure | Value |
|---|---|
| Mean members per group | 3.42 |
| Median members per group | 3 |
| Groups with 2 or more members | 76% |
| Groups with exactly 1 member | 22 groups |
| Groups with no members at all | 7 groups |
The 22 single-member groups and 7 empty ones are worth naming honestly: they are groups that were created and then left, sometimes before anyone else was even added. That is a normal thing to happen when creating a group costs nothing and takes a few seconds — some fraction of any free tool's usage is people trying it once and moving on.
Whether a group ever records an expense is close to a coin toss
Of the 119 groups, 52 (44%) never recorded a single expense. The remaining 67 (56%) recorded at least one. So slightly more groups than not do get used for their actual purpose, but a large minority — more than two in five — are created and then abandoned before a single cost goes in.
Among the 67 groups that did record at least one expense, usage is uneven rather than evenly spread:
| Expense count | Groups |
|---|---|
| Only 1 or 2 expenses, ever | 19 groups |
| 3 or more expenses | 48 groups |
| 10 or more expenses | 19 groups |
| Largest single group | 117 expenses |
Groups with at least one expense average 4.34 members, noticeably larger than the 3.42 average across all groups. Groups that never record anything tend to be smaller, which is consistent with the ordinary explanation: a group of one or two people is more likely to be a test, or a plan that never happened, than a group of four or five who had an actual event to split.
Most active groups live for about two days
Of the 54 active groups where at least two expenses carry dates, 24 of them (44%) logged every single expense on one calendar day. The median time between a group's first and last expense, across all 54, is 2 days.
This is the clearest pattern in the whole dataset: a Splitright group is usually built around a single event, not an ongoing arrangement. A dinner, a trip, a weekend — the group exists for as long as the event does, the expenses go in during or just after it, and then the group is never touched again because there is nothing left to track. That is not the app failing to retain people; it is the group succeeding at the one thing it was created for and then correctly having no further use.
Recurring use — the kind you would expect from a shared flat paying rent and bills every month — is rarer than the marketing image of an expense-splitting app suggests. It exists: the 19 groups with 10 or more expenses, and especially the one group with 117, look like ongoing arrangements rather than single events. But they are a minority, and even they will end eventually when the flatshare or the arrangement ends.
Volume over time, without dressing it up
Monthly expense volume peaked at 151 in October 2024 and has run at roughly 5–7 a month through 2026. We are not presenting the October 2024 figure as growth, because it is not: it is a single month's peak, almost certainly driven by a handful of active groups logging many expenses at once, and the months since have settled to a much lower, steadier rate.
One fact from 2026 is worth stating plainly because it argues against reading anything optimistic into month-to-month numbers: all 8 groups that were active in 2026 were created in 2026. None of the groups from 2024 or 2025 came back. Whatever those earlier groups were for, it finished, and the people in them did not return to start something new. Each month's activity is coming from a fresh, small set of groups, not from an accumulating base of returning users.
What this actually tells you
Put the pieces together and a consistent, unglamorous picture emerges. Most groups that get created do eventually get used for a real expense — 56% is a majority, if a slim one. But a substantial minority, 44%, die before that ever happens. Of the groups that do work, most are short-lived, not because the app loses them but because the thing they were tracking — a trip, a dinner, a weekend — had a natural end, and the group's job finished when the event did. Recurring, ongoing use is real but uncommon: most groups are not a flatshare paying rent every month, they are one occasion, logged over a day or two, then left alone.
Two practical takeaways follow from that, for anyone splitting costs in a group, whichever tool they use:
- Log the first expense immediately, not later. Given how many groups in this data never record anything at all, the gap between creating a group and actually using it is where most groups die. If you create the group, add the first cost there and then, before the momentum of “we should really track this” disappears.
- If one person is doing the bookkeeping, make sure everyone else actually sees the result. Given that a typical group here is one organiser and several people who never open the app, the running balance only does its job if it is shared somewhere everyone reads — screenshotted into the group chat, read aloud, whatever works — because the people who are meant to pay are, structurally, the ones least likely to be looking at the app themselves.
- Agree the splitting method before the event, not after. This is not something the data proves on its own, but it follows from how compressed these groups are: with a median lifespan of two days and 44% of active groups logging everything on a single day, there is no time to renegotiate who owes what once the event is under way. Decide up front.
Where this data came from
This is our own production data, from Splitright, a free app for iPhone, iPad and Android that lets a group split costs with no account and no sign-up: one person creates a group, shares a six-character code, and anyone with the code can open it from their own phone. It is a small app with roughly 30 users so far, which is exactly why we think the honest version of this data is more useful published than kept to ourselves — there is no story here about explosive growth, just a reasonably clear picture of how a handful of real groups actually behave.
Get Splitright for iPhone and iPad · Get Splitright for Android
Common questions
- Do most expense-splitting groups actually record an expense?
- In our data, 56% of groups recorded at least one expense and 44% never did. So a majority get used at least once, but a large minority are created and then abandoned before a single cost goes in.
- How long does a typical expense-splitting group stay active?
- Briefly. Across the 54 groups with at least two dated expenses, the median time between the first and last expense was 2 days, and 44% of them logged every expense on a single day. Most groups are built around one event — a trip, a dinner, a weekend — and stop the moment that event ends.
- How many people are actually using the app versus just being added to a group?
- Far fewer than the member count suggests. 407 member slots in our data correspond to roughly 30 real users, because a member is a name typed into a group, not a person who installed anything. A typical group is one person doing the bookkeeping for three or four people who never touch the app.
- Is recurring group use, like a shared flat, common?
- No, it is the exception. Most groups that work are short-lived because the thing they were tracking — a trip, a dinner — ended. Only a minority of groups, the ones with ten or more expenses, look like ongoing arrangements such as a flatshare, and even those eventually stop.