Case Study: Weak Signal on Every Carrier at Two Senior Residential Campuses

The Challenge

The operator needed to know how poor indoor coverage really was across both campuses, whether one carrier was to blame, and what kind of fix to plan for, backed by measurements rather than complaints.

What We Surveyed

• Mid-Atlantic campus: two multi-story residential buildings, from basement to roof. Nearly 4 million positioned readings.
• Gulf Coast campus: one large mid-rise building with a broad multi-wing base. About 7.5 million positioned readings.
• Every occupied level, plus a rooftop control on each campus.
• Six carriers: AT&T, Verizon, and T-Mobile, plus the prepaid brands Dark Star (on AT&T's network), US Mobile (on Verizon's), and Mint (on T-Mobile's).
• Measurements: signal strength (RSRP), signal quality (RSRQ), 5G signal and signal-to-interference, the serving band, and Wi-Fi signal.

What We Found

1. The walls were the problem, not the network

On both campuses, rooftop readings were about 21 dB stronger than indoors overall, and every carrier measured on the roof was stronger there. The towers delivered usable signal to the site, and the buildings absorbed it.

2. No carrier escaped

Indoors, every carrier's typical reading sat between −114 and −117 dBm at the Mid-Atlantic campus and between −107 and −113 dBm at the Gulf Coast campus. No carrier was materially better than the others, so switching carriers would not have solved anything. The fix has to be carrier-neutral.

3. Low band was what reached the deepest rooms

The bands each network served on tell the same story. At the Mid-Atlantic campus, the 700 MHz low bands carried the largest share of serving readings, and those readings were the weakest, around −115 to −118 dBm: low band was what still got through where the higher bands had faded. The mid-band capacity layers, Band 2 (1900 MHz) and Band 66 (AWS, 1700/2100 MHz), were present but weak.

4. Mid-band 5G barely got inside

At the Gulf Coast campus, Verizon's C-band 5G (n77, around 3.7 GHz) showed up in only about 1,700 serving readings, against roughly a million Verizon LTE readings. Standalone 5G quality also swung sharply by floor: on one level, every standalone 5G reading fell below 0 dB signal-to-interference. A 5G indicator says little about usable 5G indoors.

5. Height helped, basements hurt

In one building, the share of floor area below −110 dBm fell from over 90% in the basement to under 20% on the top floor. On the Gulf Coast, it fell from 72% on the lowest level to 30% on the highest.

6. Wi-Fi was not a safety net

Wi-Fi was marginal too, at around −78 to −79 dBm, so Wi-Fi calling could not be counted on as a fallback.

7. Some areas were below −110 dBm on every carrier

Those are places where an emergency call could fail, and they were flagged as the first priority in the problem-area inventory.

8. Strong rooftop signal was not clean signal

Signal quality on the roof was no better than indoors: several cells arrived at similar strength and none dominated. That matters when deciding where a donor antenna can go.

What the Client Received

• An executive summary and a report for each building.
• A problem-area inventory, rated Critical, Major, Moderate, or Minor by size, depth of the deficit, and how many carriers are affected.
• Heatmaps for every level, metric, and carrier, with per-level statistics and the underlying data.
• An interactive 2D dashboard, filterable by level, metric, carrier, and threshold, and a 3D stacked-floor viewer, including an offline version. See custom signal dashboards.

The Outcome

The operator now has a measured baseline for every floor and every carrier: the evidence needed to scope a carrier-neutral in-building system, to decide where a donor antenna can and cannot go, and to verify the result against the same grid after installation.

Cell Surveys is measurement-only. We do not design, sell, or install coverage equipment, so the findings carry no product to justify. Learn about our onsite signal surveys, or contact us to scope a survey of your buildings.