“Good afternoon. This is Domino’s Pizza. What can I get started for you?”
“Hi. Yes. I’d like a thousand assorted large pizzas delivered tonight.”
“Very well. Where would you like them delivered to?
“The Pentagon. Add it to the tab.”
“Will do. Thank you very much.”
Click.
After hanging up the phone, the cashier turned to her co-workers and said, “All hands on deck. Something’s about to go down.”
Maybe it happened that way.
Maybe it didn’t.
I first heard about the “Pizza Index” several years ago when I keynoted a military analytics symposium. Afterward, I met with members of their analytics team. The conversation turned, not surprisingly, to the management of Classified information. I was told that protecting the Classified information was actually the easy part. You know what needs to be protected, and you protect it.
A more challenging situation is when multiple pieces of public data can be integrated to derive Classified information. The most well-known example is the (presumed) phenomenon of increased numbers of pizzas ordered in the days (or hours) leading up to significant, often secret, geopolitical events or crises. During the Cold War, Soviet intelligence operatives monitored food deliveries to assess government activity. They called it Pizzint. Pizza Intelligence. On August 1, 1990, a record number of pizzas were delivered to CIA headquarters. The next day, Iraq invaded Kuwait, marking the beginning of the Gulf War.
The belief that such events are presaged by upticks in pizza orders persists to this day. The Pentagon Pizza Report (@PenPizzaReport) has its own X account with nearly a quarter-million followers.
Interest in the Pizza Index spiked recently as tensions escalated between Israel and Iran, and the United States appeared increasingly likely to become involved in the conflict. (Popular interest in the metric, though, is certainly a lagging indicator of such events.) An item on the news crawl the other day caught my attention and brought back memories: Pizza delivery monitor alerts to secret Israel attack.
Opinions differ regarding the predictive reliability of this metric. A spokesperson for the Pentagon denied any connection telling Newsweek, “There are many pizza options available inside the Pentagon, also sushi, sandwiches, donuts, coffee, etc.” Actually, that seems to be more of a non sequitur statement than a denial. It would be like your parent asking if you ate all the cookies and you replied that you have many dessert options in the pantry.
The Pizza Index and the Waffle House Index (to assess recovery from natural disasters) are examples of Open Source Intelligence, which connect disparate data from public sources to provide detail about someone or something.
At the time of my presentation, teams at a couple different universities were studying the subject, but I haven’t heard anything about the specific research. If I had, I probably wouldn’t be allowed to talk about it anyway.
I understand the appeal. It’s like a puzzle. You’re trying to put the pieces together to uncover something that nobody else is noticing. The critical ingredient, though, is context.
With the Pizza Index, you have a signal, but you need context to interpret it.
More pizzas. OK. What does it mean? What are you supposed to do about it? What other factors are you considering? By itself, the Pizza Index is not a very useful metric, but may have more predictive value when considered in conjunction with other information. But just what information is relevant in each situation?
Generative AI may hold the key.
Generative AI is all about context. About identifying connections and relationships that might not be apparent to perhaps anyone. Integrating the Pizza Index with ChatGPT would be an interesting research project.
Of course, hindsight is 20/20. Easy in retrospect. Not as easy beforehand. How often are you willing to be wrong? After all, there are many false positives.
Or are there???
