Just-in-time manufacturing has earned companies billions of incremental dollars of profit. Driven by now-CEO Tim Cook shutting down its own manufacturing and outsourcing to more agile suppliers, Apple reduced its inventory on hand from 90 days to five days over several years by switching to "just-in-time" manufacturing via partners.
You might notice your iPad from Apple being shipped to you straight from the factory in China. Less inventory on hand means higher profits and better cash flow, which is a big reason why "just-in-time" lean manufacturing took off.
Just-in-time manufacturing originally referred to manufacturing goods to exactly match demand, but more recently it has also come to mean producing with minimum waste. Advertising is by its nature, a probabilistic, risk-based exercise -- the marketer is making a bet that commandeering your attention at a certain time and in a certain place is worth it for the chance to influence your subsequent purchasing decision.
- Continuous improvement (using data to stop doing the same crap over and over again)
- Eliminating waste (stop showing your stuff to people who "don't care or aren't there")
- Ability to produce things with smaller "batch sizes" or production run sizes by reducing set-up time (ability to create more dynamic messages since you don't have to, for example, physically print each one, which has a HUGE set-up time)
- Automation (take manual button pushing and replace it with algorithms and rules)
Digital marketing has made significant progress on these fronts and soon all advertising will have no choice but to follow this lead. This is not about the iris-scanning personalized marketing dystopia depicted in the Tom Cruise film, The Minority Report, but even simple recognition of the real-time state of foot traffic past an electronic billboard when combined with current weather and time measurements could provide an order-of-magnitude improvement in the applicability of ads and hence brand recall or other performance metrics for advertisers.
Advertisers could bid for billboard placement based on how many people are actually walking by, whether it's lunch time or leaving-work time ("John Anderton! You could use a Guinness right about now!"), and whether its about to rain or not. The price of screens and displays continues to drop making these types of new ad delivery environments inevitable, but...
The bigger long-term questions are how and how quickly (a) a reliable, dynamic, real-time software infrastructure will arise to support these opportunities, and even more importantly (b) if marketers will be able to adjust the way they budget, plan and buy media to be more immediate and opportunistic.
Just-in-time advertising is, in essence, not making your mind up ahead of time about where and how much money you're going to spend on advertising your products, and in some contexts, not having the ad be selected until context, time or location data can be factored into the decision. Just like its manufacturing predecessor, it will take a lot of work and coordination effort in order to set up the systems to realize what will be billions of dollars in incremental profits and value as the old ways change.
Creative production time and the logistics of just getting creative to the be ready to show as an ad today limits whether a lot of more personalized versions of ads can be kept ready to show to users in various media, but the more electronic these media become the less that is a concern. Ditto the current lack of dynamic pricing environments -- both of these will soon be broadly possible, but the path to changing budgeting and buying behavior will need to be catalyzed, just as automotive manufacturers and other firms started discovering in the 1990s the benefits Toyota had reaped from its just-in-time/lean manufacturing movement (books like James Womack's The Machine That Changed the World (1990) showed what was possible then with lean manufacturing).
Measurement and targeting should be the same thing, and the word for that is "addressability." Television advertising is slowly changing, with advertisers and broadcasters tired of weak measurement services and a reliance on undersized, sub par "samples" of viewer data. Instead real-time, truly addressable set-top box data could be the difference between knowing an estimate of the number of 18 to 49 year-olds that were watching Desperate Housewives last night, and being able to bid to buy a 30-second spot in the first commercial break based on knowing how many people were watching in Los Angeles within the first 120 seconds of the show.
Finally, for just-in-time advertising to become reality we'll need a working feedback loop. It's hard to get better when you don't know what works, and what people actually prefer. By creating various programming and advertising options in a controlled way and measuring consumer responses to those options as they happen (whether or not that's adding a like/dislike button to TV ads, who knows?), we can stop guessing and adjust quickly before the opportunity is lost.
Immediacy is addictive not just for consumers, but also for marketers. Once the CEO has seen a chart that updates her company-wide sales every 30 seconds, I guarantee she's not ever going back to those weekly spreadsheets again.