Value Packs / Telecom / Customer Care Analytics / Brief 02

Customer Care Analytics · Brief 02

Network complaints · New York

Not dropped calls — and that changes the owner

Care & complaints + network usage · by category × district × plan · runs on your own infrastructure

The one line you configure

Which complaint categories are growing fastest this week against their own baseline — and for each, is the driver the volume of contacts or how they close?

In a nutshell

Not the usual cause. Network complaints rose 28%, and the increase is not dropped calls — those are flat. It is data-speed complaints, concentrated in 5G subscribers, and it lines up with a step down in 5G session quality in the same districts. A coverage-driven read would have sent this to the RAN coverage team; the actual issue is throughput on an existing footprint.

At a glance

Network complaints+28%New York
Data speed+61%the whole rise
Dropped calls+1%not coverage
5G→4G fallback+2.7×the mechanism

WHAT MOVED · Network complaints against dropped calls, by week

150112.57537.50W-5W-4W-3W-2W-1this week
Data-speed complaints (indexed)
Dropped-call complaints (indexed)

This is the discriminator. Dropped calls — the usual cause, and the one a coverage team owns — never move. The entire rise is data speed.

WHERE IT IS · Complaint rate per 1,000 subscribers, by district

Manhattan5G-dense8.4
New York5G-dense7.9
New York5G-dense6.2
the Bronxmixed2.1
Newark4G-led1.4

Normalised per 1,000 subscribers, so this is not a size effect. The gradient follows 5G density, not population.

WHY IT MOVED · 5G fallback against data-speed complaints

005050100100complaint normfallback normManhattanNew YorkNew Yorkthe BronxNewark5G→4G fallback ratespeed complaints / 1k

Complaints track fallback almost exactly. Subscribers are not reporting no service — they are reporting that 5G keeps dropping them to 4G.

Why the weekly view matters

Complaint categories are noisy day to day. Rolled weekly and split by category, the divergence is clean: one category carries the whole rise while the rest sit still.

So what

This is a 5G throughput and fallback problem presenting as a care volume problem, and it affects only 5G subscribers — the cohort paying most for the experience. Route the affected districts to the RAN performance team with the fallback rate and session-quality evidence, not to coverage planning. Care should expect the volume to persist until that lands, so the resolution path needs publishing in the meantime.

The points that matter

The rise is one sub-category

Data-speed complaints up 61% while dropped calls are flat — so the usual coverage explanation does not apply.

+61%data speed

The mechanism is fallback

5G→4G fallback up 2.7× in the same districts, and complaints track it almost exactly.

+2.7×fallback

It hits the paying cohort

4G-only subscribers are unaffected; the complaints come from subscribers on 5G plans.

5Gonly

Questions it already answers

Is the fallback a capacity issue or a configuration one?

Correlates fallback against cell load and alarm history on the same sites, which separates congestion from a parameter problem.

What can this not tell us?

Not the RAN fix itself. It localises the districts and sites and proves the subscriber impact, so the performance team starts with the answer rather than the search.