Early in my career I was an RF engineer at T-Mobile. When a cell site underperforms, the instinct is to reach for capital: a new site, more spectrum, more equipment. All of that takes money, permits, and months.
So we started somewhere less glamorous: a site audit. Instead of looking at one bad site at a time, we lined up performance data across the market and asked a different question. What do the underperformers have in common?
Same sites, same spectrum, twice the downlink
Relative downlink performance on the affected radio designs
The pattern
The pattern turned out to be specific radio designs with a particular cabling configuration. On paper those sites had everything they needed. In practice, the way they were cabled was holding the downlink back.
Changing the configuration didn't require new equipment or construction. It was a non-capital project: a crew, a plan, and the right sites in the right order.
The result
Downlink performance on those radio designs roughly doubled. Because the same designs were used across the region, the fix spread too, and spectral efficiency (how much data you squeeze out of the spectrum you already own) improved region-wide.
The fastest upgrade is the one that's already installed and configured wrong.
What I took from it
- Look across, not just down. One bad site is a ticket. Fifty sites with the same bad pattern is a story.
- Price the cheap fix first. Before asking for capital, prove you've used what you already have.
- Tell it so people act. "These radio designs underperform because of cabling, and fixing them costs almost nothing" gets approved a lot faster than a spreadsheet of site IDs.