The book gives a good general introduction into the concept of Big Data and illustrates it with very interesting examples. The examples show new business approaches possible in the world of Big Data, and how they are different from the traditional strategies.
Example 1: car manufacturer found an improperly working fuel gauge. Normally, this discovery would start from a request to the gauge producer to solve the problem, what could affect new production. However, the car manufacturer armed with data did not immediately informed the gauge producer, but created a software patch to correct the problem and sold it to the gauge producer.
Example 2: organizations use data tools to evaluate potential employees. One organization discovered that graduates from top schools do not perform better than graduates from less prestigious educational institutions, as a result, personnel costs could be reduced with expectation of increased performance. In some jobs, employees with prior criminal record actually outperformed coworkers without criminal record. Employees who used a non-standard browser were found more likely to take independent initiatives.
Important point: all data activities need to start from the strategy based on business need. An average company won’t be able to understand and analyze all existing data – and it is not needed to run a successful business. Only data relevant to a specific goal is required.
The SMART model — start with strategy, measure metrics and data, apply analytics, report results, transform your business — allows you to cut through the chaos.
Instead of starting with the data, start with your business objectives and what you are specifically trying to achieve. This will automatically point you towards questions that you need to answer, which will narrow data requirements into manageable areas.
Once you know what you are trying to achieve and you are clear on what SMART questions need to be answered, then work out how you can access that information so you can measure metrics and data.
The next step is to apply analytics, extracting useful insights from the data that can help you answer strategic questions. The data themselves are meaningless unless they help you to execute your strategy and improve performance.
Example 3: fashion retailer needed to understand who were its best customers. The retailer measured foot traffic passing the store (mobile phone sensor), then combined this information with number of people coming into the store (the same sensor), and then combined it with actual purchases. Result: window display could be optimized more effectively and one store was closed due to insufficient traffic.
Example 4: data aggregated from wearable devices became more valuable than the device itself. Wearable device manufacturer re-imagined its business and became a data company, while still producing the device to collect data.
Important point: analytics and data visualization are part of the same task. An idea that first data is analyzed, then visualized by a different group, and then presented to a decision maker would not provide the most value. Analysis and visualization need to work together and the decision maker should not be expected to use visualization tools – but to receive the insight generated by the analysis.
An interesting site mentioned in the book Information is Beautiful – the site highlights data visualization examples.