Korva is a scroll of the internet’s best ideas.
Nobody remembers last week's listening
Survey limits
Surveys are one of the two standard ways to measure podcast audiences, and they come with two limits. They are expensive to run, so scaling one is costly, and they ask people to recall what they listened to in the past week, which humans are bad at.
Music's switch
Music went through this already. Ratings there used to rely on surveys, the same as radio, and when measurement moved to monitoring what people really listened to, the ratings changed completely. People can't correctly identify and report back their own actions, and flawed data can only be cleaned so far.
IP matching
The other standard method is IP-based matching, which uses an IP address as an ID for the listener. At best an IP identifies a household, which fits demographic data poorly, because demographics describe individuals.
Why does an IP address only match a household?
Most homes connect through one router, and every phone, laptop, and television inside shares the single public IP address the internet provider assigns to it. Mobile carriers go further, putting many subscribers behind a single shared address. Many providers also rotate addresses among customers, so the same address can point to a different home next month.
How did Billboard charts change when measurement did?
Music charts went through a similar switch from reports to counts in 1991. Until then, Billboard's album chart worked from reports phoned in by record stores. In May of that year it moved to Nielsen SoundScan, which counts sales from barcode scans at the register. Albums by Garth Brooks and N.W.A. reached number one within months of the switch.
More from PodBiz | Where's The Money In Podcasting?
