Churn Prevention Intelligence · Brief 01
Churn-score spike · Brooklyn & Manhattan
Where risk is accelerating, and in which plans
The one line you configure
Where is churn risk accelerating fastest today rather than simply sitting high — and does anything in our own care, usage or billing data move with it?
In a nutshell
Average churn score across New York and Manhattan rose 38% in a day, and it is not spread — it sits in Postpaid Gold and Platinum on the Aurex 12 Pro handsets, which is the opposite of the usual churn profile. Two of our own signals move with it: data-speed complaints in those districts and a step down in 5G session quality. The score told us who; the connected read tells us why, and it is a network experience problem, not a price one.
At a glance
WHAT MOVED · Average churn score against its own baseline
The book average never moves off 34. Two districts carry the whole rise, and the steepening starts three days ago — while port-out requests are still flat, so the window is open.
WHERE IT IS · Churn-score movement by plan × device age
The gradient runs the wrong way for a price story: risk rises with plan value and device recency. That is an expectations problem, not an affordability one.
WHY IT MOVED · 5G session quality against churn-score movement
The relationship is clean across the city: the worse the 5G session quality, the faster the score moves. New York sits alone in the corner.
What the connected view adds
The scored table alone gives a rising number and a subscriber list. Joined to care and usage on the same district slice, the cause appears — and it is actionable by a different team than retention.
So what
This is a network-experience problem presenting as churn risk, in the most valuable cohort on the newest handsets — the subscribers least tolerant of a speed regression. A retention offer would spend money without touching the cause. Route the district pair to network operations with the session-quality evidence, and hold retention until the experience fix lands; then re-read the score.
The points that matter
Risk rises with value
Postpaid Gold and Platinum on the Aurex 12 Pro handsets move fastest; Prepaid barely moves. The usual churn gradient is inverted.
+44%Platinum · newestOur own data names the cause
Data-speed complaints up 2.4× and 5G session quality down 11% in the same districts, same window.
−11%session qualityNot a price problem
Price complaints and port-out requests are both flat, so a retention discount would miss.
flatprice signalsQuestions it already answers
Is any other district on the same trajectory but earlier?
Ranks every district by score movement against session quality, so the next New York is visible while it is still a watch item.
What can this not tell us?
Not why session quality dropped — that is a RAN question for network operations. It establishes that it did, where, and what it is costing in risk.