A while ago I referred to Beth Kanter’s post on why “Data Informed” decision making is so much better than the “Data Driven” dito touted by the fans and proponents of “Big Data”.
Here’s an example on the topic, drawn from my summer job as a tourist coach driver in Stockholm:
First, some info on what being a tourist coach driver in Stockholm entails, in this context:
Stockholm resides between Sweden’s third largest fresh water lake, “Mälaren”, and the Baltic Sea, with several locks regulating the flow of water. Around 30% of the city area is made up of waterways, and another 30% is parks and green areas. The city is built on top of 14 islands. From a driving/traffic perspective, these characteristics makes driving in the city “interesting”, since with so much water between the parts of the city, the connection points, particularly the bridges, become serious congestion points for the traffic. To this, add the fact that large parts of the city are quite old, with streets architected for pedestrians and horse carriges, not for cars, and particularly not for large and heavy vechiles.
So, with this brief background over and done with, let’s get to the distinction between being “data driven” vs being “data informed” when driving – particularly when driving heavy vechiles – in Stockholm:
A “data driven” approach would rely upon use of the routing guidande mechanism available in any freeware or commercial GPS: you’d simply plug in your destination address, follow to the letter the instructions given by the nice sounding (often female!) artificial guidance voice, and voila’, without any hassle you’d get to your destination, in the time indicated when you entered your destination info. Thus, as a driver in the data driven approach, the only requirement on you – over and above the knowledge and skills of manuvering a vechile – would be to follow the instructions of the digital voice in your GPS unit, you wouldn’t need any particular local knowledge about city layout, street traffic directions, or typical traffic patterns. Nor would you need any decision making capability, the only thing required of you would be the ability to obey orders (sounds like nirvana to the empty suit-types!) Basically, a human (or even a non-human) automation with driving skills could do it.
As it happens, I see a lot of these “human automations” blindly following the advice of their GPS units in city, and inevitably, they end up in all kinds of trouble, not least because the traffic situation and patterns change in realtime, leading to that a route that worked ok yesterday, does not work today. (As a side note, I see similar phenomena in my “normal” daytime business where mgmt increasingly believes that “if we’d only have the perfect process (business processes) identified and defined, we’d be able to run our shop with low pay moron’s!”)
Enter crowdsourcing GPS:es, e.g. apps like Waze (which btw is absolutely great, if used in the right way, as well as free!): these thingies do not rely solely upon static info (street grid etc) but also overlay actual traffic patterns in realtime, i.e. these crowd sourcing GPS:s are aware of traffic volumes on the various roads, and use this info in their routing algorithms. So, equipped with Waze, shouldn’t you as the driver be able to input your destination address and then blindly follow the instructions…?
Turns out that the answer is “NO!”: for the fun of it, over the past 3 weeks that I’ve been driving tourist coach in Stockholm, I’ve had Waze running in guidance mode next to me – without necessarily obeying its instructions! – and it turns out that more often than not, the instructions given by the “system” would have gotten me – with my 25 tons heavy and 14 meters long coach – into all kinds of trouble, either serious trouble such as entering a cul-de-sac, or less serious but very annoying trouble such as ending up in a lot of traffic.
However, I still use Waze while driving, and I think it’s a great help, but I’m not foolishly and blindly obeying (actually not even using!) its routing/guidance info, instead I use its crowd sourcing capability to inform me about current traffic patterns and volumes in the city, which I then take into account while doing my own routing planning.
Bottom line: we are still far away from a situation where computers, regardless of how much “Big Data” we feed them with, and how much “Business Analythics” we employ on that data, can outperform a skilled human.