Splitright

What 119 real expense-splitting groups look like

Not a survey or a retention study: an aggregate snapshot of 119 groups stored in our own expense-splitting app.

← Back to the Splitright home page

Correction, 4 September 2026: An earlier version described expense-date spans as group lifetimes and inferred user counts and return visits. Those claims were not supported by this snapshot. The article below separates stored records from behaviour we cannot measure.

What this is, and what it is not

On 2 September 2026, Splitright's production database contained 119 groups, 407 member-name slots and 742 expenses. These are aggregate counts from one small app, not a representative sample of how people split costs in general.

This is a snapshot of retained records, not every group or expense ever created. Deleted records are absent. We cannot establish how many unique people created the groups, how often they opened the app, or which groups were tests rather than real arrangements. No group names, codes, member names or expense descriptions are published here.

A member slot is a name, not an installation

In Splitright, someone can add names for everyone sharing a cost without those people installing the app. The same person can also appear in multiple groups. So 407 member slots do not mean 407 users, and two names in a group do not prove that an invitation was sent or accepted.

One organiser can do the bookkeeping, or several people can open a group using its code. Neither the organiser nor the other participants need an account. The member list alone does not tell us which pattern a group follows.

Group size: mean 3.42, median 3

The stored groups are generally small. These figures describe the member-name lists at the snapshot date, not necessarily their size when first created.

Member-name counts across 119 stored groups.
MeasureValue
Mean member slots per group3.42
Median member slots per group3
Groups with 2 or more member slots90 groups (76%)
Groups with exactly 1 member slot22 groups
Groups with no member slots7 groups

56% of stored groups contain an expense

Of 119 groups, 67 (56%) contained at least one expense and 52 (44%) contained none. That is a useful description of the database, but it is not a conversion rate from opening the app to adding an expense, or proof that the empty groups were abandoned.

Stored-expense counts among the 67 groups containing expenses. The 10-or-more row is a subset of the 3-or-more row.
Expense countGroups
1 or 2 expenses19 groups
3 or more expenses48 groups
10 or more expenses19 groups
Maximum stored expenses in one group117 expenses

Twenty-one of the 22 single-member groups contained no expenses. That gives us a reason to test clearer guidance about adding people before splitting costs. It does not prove why those groups are empty, or that changing a button will cause people to use them.

Expense-date spans are not group lifetimes

Among the 54 groups with at least two parseable expense dates, the median span between earliest and latest dates was 2 days, and the maximum was 329 days. In 24 of those groups (44%), the expense dates fell on the same calendar day.

The distinction matters: an expense date is chosen by the person entering it. Someone can enter last week's costs today or return tomorrow to edit them. These figures do not tell us when expenses were entered, how many sessions occurred, how long a group remained in use, or whether anyone returned to the app.

A trip can produce many expenses; a recurring arrangement can produce few. Expense counts and date spans alone cannot classify the purpose of a group or establish retention.

What we can usefully take from this

The snapshot shows a mix of empty groups and groups containing many expenses. It suggests questions for improving the app, but it does not answer why people start or stop using it. We are treating clearer first-step guidance as a hypothesis to evaluate, not as a proven remedy.

For people sharing costs, our practical advice is independent of these statistics:

About Splitright

This is our own app's data. Splitright is a free expense-splitting app for iPhone, iPad and Android, with no account or sign-up required. One person creates a group and can share its six-character code so others can open it from their phones.

Common questions

How many groups contain an expense?
In the 2 September 2026 snapshot, 67 of 119 stored groups (56%) contained at least one expense; 52 (44%) contained none. This does not tell us whether an empty group previously contained expenses that were deleted.
How long does a typical expense-splitting group stay active?
This snapshot cannot establish that. Among 54 groups with at least two parseable expense dates, the median earliest-to-latest date span was 2 days. Expense dates are chosen by users, not visit or entry timestamps, so this is not a measure of group lifetime or retention.
Do 407 member slots mean 407 app users?
No. A member slot is a name someone adds to a group. One person can add multiple names, and the same person can appear in several groups. These records do not establish the number of unique app users.
Do groups with many expenses represent recurring use?
Not necessarily. Nineteen groups contained at least 10 expenses, but expense count alone cannot distinguish a busy trip from a recurring arrangement such as a shared flat.