[Editors note – I was in Ohio last weekend and got feedback that my blog can be hard reading for folks without a prior knowledge of advertising, so I included an introductory paragraph before jumping into the main topic.  If you already know the advertising industry just skip to paragraph two.]

Since the emergence of mass media, advertisers have used two statistics to measure the value of advertising – reach and frequency.  Reach is the number of unique people who see an advertisement.  In the online world this is also called unique viewers.  Frequency is the number of times a person saw an advertisement on average.  There has been a lot of research done over the last 50 years to determine how many times someone needs to see an advertisement in order to remember it, and most advertisers now target a frequency of at least three (if not much more).  That explains why you see the same ads over and over during football games.  If you have a national brand, you are typically trying to get both high reach AND high frequency.  The combination of the two is very valuable and very hard to get, for reasons I will mention below.

When I was a kid, my family used to sit around with popcorn and watch The Cosby Show.  Cosby was on Thursday nights at 8pm, and every Thursday night, across the entire eastern time zone of the United States, millions of people did the exact same thing at the exact same time.

There are shows today that have a bigger audience than The Cosby Show, but most people record television and watch it on their own schedule.  I did a little survey on my own Tivo (technically my wife’s TIVO since she is the only person allowed to use it) and the average length of time a show has been sitting on our machine without being watched is 2 months.  There was even one show from last September.  I should point out that two months is actually an improvement for us – before we had Netflix we used to record a lot of movies for a rainy day and those would sit on the machine for years.

Time shifting presents a challenge for advertisers.  Fashion, consumer electronics, cars, movies, politics – these all come with some expiration date.  In traditional media the advertiser is paying for spots that may not be relevant by the time they are watched.

The second challenge, which is a recurring theme of this blog, is that audiences have moved online.  The Internet offers more exact targeting than television – you can get exactly 100,000,002 unique individuals to watch a commercial.  The hard part is that these users are completely free from a programming schedule.  I may go to Hulu at 4am to watch an episode of the Office from four seasons ago.  A few minutes later I go to Yahoo to check my stocks.  A minute after that I go to Facebook.  Not only do I have the attention span of a hummingbird, but my media consumption is unpredictable.  Good luck finding me three times in a 30 minute window.

What I am getting at is that a new element has emerged in the advertising equation: TIME.  The old definition of “frequency” only considered the duration of a television show.  The assumption was that everyone on the planet was watching the show at the same time.  In recent years the definition of “frequency” changed, without much debate, and it now refers to the number of impressions over the life of an advertising campaign.  This means that a frequency of three, which used to mean watching the same advertising spot three times in a 30 minute period, now means watching it three times over a three month advertising campaign.  People are clearly not measuring the same thing anymore.  Further, new technology has driven advertisers towards more performance-oriented campaigns, which means that a campaign may last three hours instead of three months.   This creates a real problem – for the first time, media partners have to be measured for throughput.

Throughput is generally not used outside of manufacturing or communications, but in this instance it means ad frequency divided by time.  In other words, a throughput of three ads per hour is much more powerful than a throughput of three ads per year.

Throughput is the secret of good media buying – it often distinguishes effective advertising campaigns from ineffective campaigns.  The New York Times was an early leader here with “Follow Me Ads” – you might have noticed that when you click on a few pages in a row the ads tend to be the same.  That is a great solution for getting high throughput.  One of my investment theses for adtech investing is that throughput will be a savior for big ad networks too.  Agencies can get reach using licensed software, and they can buy media on exchanges, but it is hard to find the same person over and over without proprietary media relationships.  That is why, if you hold the price of media constant, agencies buying across exchanges generally perform worse than large ad networks.  This is an inherent problem and I don’t believe it will be solved for many years to come.