RFM

Do not add the headcounts of the two models into one number.

The 11 segments are under RFM segments on ConsoleAll-platform analyticsBusinessCommerce & revenue. The 10 buckets are on ConsoleAll-platform analyticsUsersSegments & tags, when you create a segment of type RFM.

Metrics & definitions

Both models use quintiles: people are ranked and cut into five bands scored 1 through 5. A more recent purchase scores higher on recency. People with no purchase, or no matching event, are left out.

Where	Data	Split
Commerce	Purchases on the order ledger in the last 365 days	11 segments, including big spenders. Compared with 30 days earlier, with the largest moves listed
Segment	Your event, amount property, and lookback	10 buckets. An empty amount property means counts. Recomputed daily, with moves stored

The commerce segments are champions, loyal, big spenders, new, potential loyalists, promising, need attention, about to sleep, cannot lose, at risk, and hibernating. The chart is recency score by frequency score, with headcounts at the end of the selected range. Someone who was in a segment 30 days ago and has no purchase now ends their move as churned.

The segment’s 10 buckets have no separate big-spender bucket. They compare recency with the average of frequency and monetary scores. Open a bucket to filter members.

How to use it

1. Get purchases onto the order ledger first. See Commerce & revenue. With no orders, the commerce grid is empty.
2. For the 11 segments, read headcount and the change versus 30 days ago, then the moves from one segment to another.
3. To use a bucket in a funnel, create an RFM segment whose event and amount property match what you send. It refreshes daily. See Segments.

Access & limits

Reading the grid needs the same access as commerce. Creating an RFM segment needs an owner, admin, or analyst. Commerce money is the converted base currency on the order. A segment that sums an event property uses the raw value and does not convert it.

FAQ

The two headcounts differ? Commerce only uses purchases on the ledger from the last year, and it has an extra big-spender segment. A segment uses your event and lookback, and the bucket edges differ.

A move ends in churned? That is someone who was in a segment 30 days ago and has no purchase in the current lookback. It is not a segment you can open.

Product engagement score

The score describes how broadly the product is used, how densely people return, and whether people are net new or net lost. It does not replace revenue. Daily actives alone can rise because of one campaign, so read the three parts together.

ConsoleAll-platform analyticsBusinessProduct engagement

Switch between Site, By platform, and By account. Core events, the stickiness ratio, and weekend exclusion are on the PES tab of ConsoleAll-platform analyticsConfigurationBusiness settings.

Metrics & definitions

Part	Label in the console	Formula
Adoption	Adoption	Average count of core events used by active users, divided by how many core events you listed, times 100
Stickiness	Stickiness	Daily over weekly, daily over monthly, or weekly over monthly active users, times 100
Growth	Growth	Quick ratio is (new + recovered) / dropped, then mapped onto 0–100

A quick ratio of 0.5 or below is 0, 1.5 is 50, and 5 or above is 100, linear in between. If nobody dropped and someone is new or recovered, the ratio is treated as 5. The score is the average of the three. A part that cannot be computed is left empty rather than counted as zero.

You can pin up to 10 core events. If the list is empty, TapCub infers in this order: required or core events from an installed industry pack, then live events in business categories (at most six from either step). If neither finds any, it takes the 4 events with the most people in the last 30 days. Auth, navigation, error, and system events are never treated as core. The page says whether the list is pinned or inferred.

By account, TapCub takes the larger accounts and reads the event table instead of the daily person rollup. Excluding weekends affects daily actives inside stickiness only, which suits products that are naturally quiet on weekends.

How to use it

1. List events that mean someone reached value. Do not list pageviews.
2. If weekends are naturally quiet, turn on weekend exclusion and pick daily over weekly or daily over monthly.
3. Read the weekly or monthly trend. When the score falls, look at the three parts separately.

Lifecycle states are in Lifecycle. The growth board is AARRR.

Access & limits

Reading matches other business reports. Changing core events, the stickiness ratio, or weekend exclusion needs an owner or admin. Site and platform scores use the daily person rollup, so they depend on identity resolving to a person. Fully anonymous traffic still lands in that rollup under its subject at the time, but has no person profile.

FAQ

Adoption stays low? The core list is long, or it includes events few people hit. Replace an inferred list with a fixed one in settings.

Growth is empty? The period has no new, recovered, or dropped people, so the quick ratio cannot be computed.