Zero Share ≠ Zero Sample (with Small Market Data)

So I really love that I am getting data about my FM signal in my ultra-small market. That would never have happened with the legacy people-meter and diary-based rating services.

But there’s an issue I’m learning about in the way the current system calculates share: The system treats a zero-sample day the same as a zero-share day, even though they mean two very different things.

A zero-sample, or “null share”, day means there were no vehicles in my coverage area that were capable of providing a ping.

A zero-share day is different: there were vehicles providing observations, but none of them were tuned to my station.

The problem is that the math treats both situations as 0% share. I don't think a day with no sample should be treated as though it were a day on which my station had zero listening.

For example, suppose I look at a 2 day period. On one day there is no sample at all, while on the other day there is a 50% share. If the first day is a true null-share day, the 2 day result should effectively reflect the 1 observed day: 25% of the market over the period. But if the first day is a genuine zero-share day, the combined result should be less than 25%.

That distinction matters because “we had no observations” and “we had observations and nobody was listening” are not the same.

I've zoomed in on all zero days in the last month, and have identified what I believe are null share days. I'm extremely pleased with the share my station has, as it means it's almost mathematically impossible that any other station is performing better.

With all the above said, I should not get too hung up on share. If one person *hates* my station and is regularly in the area and has a vehicle enabled with this feedback service, it could really send me a false (and frustrating) story about my station.