Definition

The fraction of arriving users whose sorting hat classification routes them to OSFA rather than to a built path.

What the rate measures

The rate falls out of normal routing operation. Each user who arrives is classified by the sorting hat into a branch; some branches have differentiated paths built for them, others terminate at the OSFA default. OSFA-routing rate is the share of arrivals that land in the second group.

The rate climbs when users land in branches the sorting hat has but the offering hasn't built paths for. Either the team built the classification structure faster than path investment caught up — expected, since classification is cheap and path investment is gated by margin economics — or the user mix has shifted such that more arrivals now resolve to under-invested branches.

Movement over time

The rate is interpreted longitudinally. Its level at a moment isn't itself diagnostic — a sorting hat with five branches and two built paths runs at a baseline OSFA-rate determined by the population distribution across those branches, and that baseline can be perfectly stable. What the audit reads is movement.

A stable rate, even a high one, indicates the offering's intent-distribution theory still matches the user base. A moving rate indicates something underneath has changed.

Pairing with sequence-retention

The rate is one of two axes intent detection produces as audit signal; the other is sequence-retention rate. The combinations resolve to different territory shifts:

• OSFA-rate climbing, retention stable: intent-category drift. The same proportion of users complete the sequence, but more of them route to OSFA. Either the theorized categories no longer carve the actual user base, or acquisition has shifted to bring in users whose intents the sorting hat doesn't have branches for.
• OSFA-rate stable, retention climbing: sequence degradation. The intent mix hasn't shifted, but the instrument has lost discriminating power — questions becoming easier to answer in any direction, branches collapsing to defaults that catch most users.
• Both climbing: a more fundamental shift. Sequence-discriminating-power and category-fit are moving simultaneously, typically because offering-surface drift, intent-mix drift, or compounding sequence edits have eroded both at once.

A single axis moving runs the corresponding diagnostic; both axes moving requires deciding which to investigate first based on which more likely produced the other.

Reading a climbing rate

The predictable misread is treating a climbing OSFA-rate as evidence that desire triangulation itself is broken, then reworking the sequence to push the rate back down or removing the instrument entirely. Both moves forfeit the audit signal that would have surfaced the underlying drift. Run the diagnostic first. The most common cause is territory shift — the categories no longer fit the user base — which is what the signal is built to produce.

Example

A B2B SaaS deploys a sorting hat with five intent branches and built paths for two of them. At launch, OSFA-routing rate sits at 62% and sequence-retention at 81%. Six months later, OSFA-rate has climbed to 78% while retention is essentially flat at 79%. The pairing reads as intent-category drift — the same fraction of users complete the sorting hat, but more of them now land in the three unbuilt branches. Investigation reveals recent marketing campaigns have brought in users from an adjacent segment whose intent doesn't map to the original two built paths. The move is either building a path for the new segment or adjusting acquisition; reworking the sorting hat to push the OSFA-rate back down would suppress the visibility into the underlying shift.

See also

• Intent-mix fit
• Intent clustering
• Sorting hat design
• OSFA-as-fallback