You might call it day-dreaming; I call it a thought experiment.
This was the thought experiment for today.
When I called into a conference call that hadn’t yet started a friendly voice told me that “your conference will begin momentarily…”
How does the system know the call is going to start at all, let alone momentarily? Is this some form of advanced technology that is tracking the person who is opening the call? Of course it isn’t, it’s just the words that someone decided to use when they developed the system. It got me wondering though, what would it take for the system to know that a call was about to begin?
To start a call the person whose call it is needs to enter a set of numbers into a phone. So the imagined system needs to be able to predict that the person is about to enter the required numbers.
I suppose we need to be clear about what we mean by momentarily at the start. The point of using the word momentarily is, presumably, to encourage someone to stay waiting because something is about to happen. For that inducement to work it needs to be relevant, but it also needs to be far enough in advance for it to catch people who may drop off because they think the call isn’t happening. It take less than 15 seconds to type in the numbers to open the call, but personally I don’t think that is really long enough to be an inducement. Momentarily can’t be used as an inducement from the time that someone starts the typing of the access codes, it needs to be more predictive than that. My view of momentarily is that it really needs to be at least 60 seconds out. Ideally, of course, it really needs to be graduated – “why are you trying to join this call it doesn’t start for 30 minutes“, “you call will start in a short while“, “your call will start momentarily“, “your call will start in two-ticks“.
Sticking to momentarily for now, there is no single defined action that someone does predictably 60 seconds before starting a call, so we need to be able to predict what they are going to do. Predicting what people are going to do is notoriously difficult if we are trying to predict in general, but people are quite predictable in defined situations.
If someone is sat at their desk and looks at their calendar 60 seconds before the call I expect that we can be reasonably confident that they are going to do the next thing on their calendar. We’d need location tracking for starters. We’d also need eye tracking technology and the eye tracking would need to be able to recognise the information that someone is looking at. It would also need to be able to identify that someone has read the right content rather than just glance at it.
There are probably a whole load of other nuanced things that people do that make it predictable that they are going to join a call. What if someone looks at their calendar before starting a journey? What if someone finishes a call 2 minutes before the call starting? What about people who join a call at the same time every day/week/month? Perhaps we could look for patterns in the brain directly? Lots of further thought experiments there.
Our predictions don’t have to be 100% accurate either we are talking about an inducement after all, it’s a nice to have. Just to get something of low accuracy would need advanced and intrusive technology. I can’t image many people being willing to wear eye tracking glasses just so that others could know when their call was about to start.
If you could get predictions to be exact why not take the next step? If you know someone is about to do something why not do it for them? If you predicted that I’m about to open the call, why not open it for me? That would get you into another challenge though; if you are going to undertake an action for someone based on a prediction the predictions can’t just be correct in certain circumstances, they need to be exact in all situations. That takes us back to predictability in general which is remarkably difficult.
Does anyone else enjoy a thought experiment (day-dreaming)?