//
archives

Archive for

Big Data & Creativity – The Uber Case Study

Mercedes-Chauffeur-Hire-S-Class-940-x-400Are we seeing the beginning of the rise of the machines? I recently went to a talk on big data that looked at Uber as a case study example of applying big data to business and this was the first thought that sprang to mind. Essentially Uber is valued so highly not because it provides cool taxis but because it is a big data company that applies big data to automate and optimise the Taxi business. We as agencies can learn alot from Uber in terms of being creative with data, however caution must be taken to avoid big data blindness and ensure proper human analysis is applied to it.

As a core insight, the guys at Uber worked out that the number one predictor of satisfaction with taxi companies wasn’t necessarily how long it actually takes the car to arrive but how long it takes perceptually based on the customers situation. For example you may expect a car to take 5 minutes to get to you in the middle of the day in the city, however 15 minutes if you were out in the suburbs – you adjust your expectations based on your situation. With this in mind they were then able to be creative with data in terms of optimising supply and demand for this key metric with factors such as the car type and quantity, the drivers, weather, traffic lights or sense of time varying depending on location (city vs country). They don’t neccessarily send the geographically closest car but the car that can get their the fastest perceptually to optimise their supply chain. All this happens in the background and creates a seamless experience for the customer, essentially Uber removes the need for a human on the phone as a dispatcher with a more efficient automated system. This is why many taxi drivers are up in arms about Uber as it really takes the human element out of their jobs, dispatchers are no longer required and they are given a predetermined route to drive which essentially turns them into slaves to the computer and can impact on their income.

Data and automation are certainly not a replacement for humans in all situations though, it can help guide us  but it still requires a human to make sense of it and apply it. The reason proposed for this is that ultimately big data is human so it requires a human to combine, interpret and apply it in a meaningful way. Just as we would expect with any form of research, simply because there is lots of it it doesn’t mean that it is statistically accurate without analysis. Take Google Flu Trends as an example, launched in 2008, Google attempted to make accurate predictions about flu activity based on aggregating search queries. Initially the model seemed to predict with accuracy however the cracks started to show when the estimate for the 2011-12 flu season was more than 50 percent higher than the cases reported by the Centers for Disease Control and Prevention. The danger that this highlights, is that it’s easy to fall into the trap of big data blindness where simply because there is lots of it we assume its statistically accurate and make poor assumptions without applying proper analysis of the results. For example, people making flu-related Google searches may not know much about how to tell if they actually have it or not, increasing false positive search queries. This is not to say big data isn’t useful, it just needs to have the same rigour applied to it as any other form of data.

With all this in mind, for us as agencies big data is great as it can help us do away with putting all our emphasis on the typical “Focus Group” as we can interpret other factors of behaviour. It can also help feed our understanding of our market and discovering peoples “secret self” – their unconscious motivations that drive their behaviour in situations which can be difficult to draw out in typical research as people tend to act differently when they are being watched. Much like Uber, we need to be creative with how we approach problems and data as data can help us serve customers better by optimising and automating their experience. Data can also help us understand our customers better then ever before and help us to create messaging that resonates with their unconscious motivations, helping to remove barriers to consumption. The machines still need us, for now at least!

Author: Alex Leece

%d bloggers like this: