Skip to main content

Payer segments

Open Journal -> Advanced Analytics -> Segments. The report groups paying accounts by payment recency, frequency and total value. This is threshold-based RFM analysis, not a measure of visits, VIP status or predicted churn. The examples use a training payment history.

Period and currency​

FieldSelection
ProjectAll accessible projects or one project. Accounts in different projects are counted separately.
Game serverAll servers in a project or one server. Select a project first.
PeriodToday, last 7, 30 or 90 days, current or previous year, or custom dates. The default is the last year through today.
From / Through (UTC)Both dates are inclusive. Payments outside this period contribute neither frequency nor value and are not considered the last payment.
Payment currencyOne currency and precision. Automatic selection uses the first available currency sorted by code, then decimal places.

The observation date is the last day of the period. For a period still in progress, only payments before the current moment count. Recency uses UTC calendar days: a payment on the observation date has recency 0.

Refresh applies the conditions. Reset restores the initial dates and thresholds. Applied values remain in the page address and survive a reload. Background refresh does not change the result while fields are being edited.

The example contains 108 paying accounts and 255 USD payments. The high-value monetary threshold was changed to 50; all six groups are populated.

Segment thresholds​

FieldDefault and meaning
Recent payment, up to days30. Maximum recency for recent, regular and high-value payers.
At risk, up to days60. Upper boundary of the next range; strictly greater than the recent-payment boundary.
Dormant, up to days90. Upper dormant boundary; strictly greater than the at-risk boundary. Beyond it, an account is long inactive.
Regular payer, minimum payments2. Frequency within the selected period, at least two payments.
High-value payer, minimum payments5. Cannot be lower than the regular-payer threshold.
High-value payer, minimum receipts1000 in the selected currency. A positive amount; use a decimal point for fractions.

The monetary threshold supports up to 20 whole and 18 fractional digits. Recency thresholds are bounded by the supported calendar span, not a ten-year history limit. Changes affect this report only, not balances, prices, discounts, player permissions or loyalty rules.

SegmentRule with default thresholds
High-value payersLast payment at most 30 days ago, at least 5 payments and at least 1000 in receipts.
Regular payersAt most 30 days, at least 2 payments, but not all high-value conditions are met.
Recent payersAt most 30 days, frequency below the regular-payer threshold.
At riskFrom 31 to 60 days without payment.
DormantFrom 61 to 90 days without payment.
Long inactiveMore than 90 days without payment.

Each account belongs to exactly one group. A large historical amount does not keep a long-inactive account in the high-value group. “At risk” and “Long inactive” describe payment history, not proof that a player has left the project.

Raising the monetary threshold empties the high-value group. Overall receipts and account counts stay unchanged; only their distribution changes.

Project and server​

Project-wide wallet payments without a server belong to the project report but not a single-server selection. An account paying on several servers still counts once in the overall project report. Do not simply add the account counts of individual servers.

The server filter retains payments attributed to that server. Frequency, value and recency are recalculated within that scope.

Measures​

MeasureCalculation
Paying accountsAccounts with a qualifying payment in the selected currency and period.
PaymentsDistinct payments. Multiple settled parts of one payment increase its amount, not its frequency.
Gross receiptsThe full paid amount before fees and refunds. Not a wallet balance or net revenue.
Average paymentOverall gross receipts divided by the payment count.
Account shareSegment accounts divided by all paying accounts, as a percentage.
Receipts per accountSegment gross receipts divided by its account count.
Payments per accountSegment payment count divided by its account count.
Days since payment, averageAverage recency of the last qualifying payment for accounts in the segment.
Guest payments and receiptsPayments without an account. Shown separately and excluded from the cards and segments.

Different currencies and precisions are never added together. For example, USD with two and six decimal places are separate choices. The same account can appear in multiple currency selections, so their account counts are not additive either.

Selecting EUR includes only EUR payments. USD amounts are neither converted nor added.

Only positive settled top-ups and purchases count. Test payments, bonus grants and administrative adjustments are not payments themselves. Repeated delivery of the same payment information does not increase the measures. A refund does not erase the payment or reduce this report's gross amount; use payment statistics for financial totals including refunds.

Historical periods​

Selecting the previous year measures recency on its last day, not today. Multi-year ranges are available. The result always contains six groups and needs no pagination; the wide table scrolls horizontally.

The training example for 2025 includes 36 accounts and 104 payments. Later payments in 2026 do not enter this selection.

Use payer retention to study returns month by month. Compare segmentation results using consistent currency, period and thresholds.

Empty results and refresh​

Without qualifying accounts, counters are zero and averages and shares show “No data”. Guest receipts may still exist. When the selected currency is absent, the report does not silently switch to another currency.

No segment contains an account in this historical period. Missing observations are not presented as a zero average payment.

New information appears after processing; delayed payments can refine past periods. The report cannot reconstruct missing history. If a selection is too large, narrow the dates or select a project or server. During an outage, stale numbers are removed and an error is shown; failure is not presented as a zero result.