Caspar Herzberg |
This article first appeared in the Spring 2017 issue of The Record.
The advent of the steam engine in the 18th century was the catalyst that started the industrial revolution, and steam power enabled factory workers to process raw materials into finished goods with a speed and efficiency that was unprecedented. It’s fair to say that steam power changed the face of manufacturing forever.
Today a similar revolution is underway, and the impact that it’s having on businesses is just as profound. This time the catalyst is not steam power but data analytics, and companies that exploit the vast quantities of data that they generate, collect, and store can transform the way they operate – providing a huge competitive advantage and boosting their bottom lines.
This is not just fanciful talk: the benefits of data analytics are real and quantifiable. Research commissioned by Microsoft and carried out by Keystone Strategy in 2016 found that those companies that have developed the most sophisticated data and analytics platforms and applied those capabilities as a regular part of their business are more profitable, and their employees are more productive. How much more? It turns out that they enjoy operating margins that are a full 8% higher than organisations that are lagging behind in data and analytics.
That’s a significant increase, and in raw financial terms this translates into an average of US$100 million in extra operating profits for the more advanced companies (controlling for factors such company size and industry vertical).
How does data analytics provide all this? “Companies that embrace data are simply doing things differently, substituting human judgement with data driven objectivity and automated business processes,” explains Joseph Sirosh, corporate vice president of Microsoft’s Data Group. “If their leadership culture embraces a data mindset then that helps them make the most of their existing assets. These are the organisations that are creating new customer experiences, optimising their revenue growth and profitability with data, creating new industries and disrupting old ones.”
As well as having superior financial results, companies with the most sophisticated data and analytics capabilities also have business processes that are more sophisticated than their peers, Keystone Strategy found. Data analysis can have a beneficial effect on the way a business operates right across the board, from sales and marketing, engineering and operations, to finance, HR and back office.
The bottom line is that the business functions of companies that have embraced the data analytics revolution look dramatically different because of the way they store, process, and use their data to make more effective decisions.
One reason that the data analytics revolution is happening now is because data is everywhere. Fifteen years ago, Rolls Royce built very few sensors into its engines, and these generated a relatively small amount of data. But today each engine is packed with sensors generating thousands of signals, and a single airline’s fleet may generate gigabytes of data every hour.
Another key reason is that even though a company like Rolls Royce may be collecting terabytes of sensor data from its engines, cloud platforms like Microsoft Azure and technology offered by Microsoft’s SQL Server – which can run in the cloud – mean that that vast amounts of data arriving at speed can be collected and analysed almost instantly, and valuable insights drawn from it.
For example, Rolls Royce can analyse detailed data from every engine fuel pump in an airline’s fleet using Microsoft’s Cortana Intelligence Suite. By comparing it to data models and to other pumps in the fleet it can alert the airline that a specific pump might not be performing well and should be replaced, or conversely that a pump’s performance is normal and may not need any attention until the next routine maintenance window. Carrying out work on components based on their actual condition, rather than per an inflexible maintenance schedule, can lead to huge cost savings by minimising both fleet disruption and maintenance costs.
Microsoft’s SQL Server database is ideally suited for this type of application, and here’s why: most advanced analytic applications today take the approach of moving data from databases into the analytics application itself to carry out the data analytics. The problems with this are obvious: the large amounts of data being moved results in high latency – slowness, in other words – and the problem gets worse as data volumes grow. And when it comes to carrying out deep analytics on real-time transactions, the truth is that this is practically impossible.
SQL Server 2016 is revolutionary because it allows sophisticated analytics and machine learning models to be run within the database itself. That can result in valuable business insights being produced more than one hundred times faster than when they are run on data that has been moved out of database. “Microsoft’s goal is to bring intelligence straight into the database with SQL Server,” says Sirosh. “Bringing intelligence into analytics offerings will help organisations embrace data and give them a competitive advantage.”
The RAC, a UK motoring organisation, has developed a black box device that is fitted into a car and real-time data from the vehicle’s management system through the cellular data network to a data analytics system running in Microsoft’s Azure cloud service. Using this data analytics system, patterns and trends can be spotted in the data, and anomalies identified – such as a car with an injector fault that may cause it to fail in the near future. That information can then be used to tell the car owner about the fault so they can get it repaired before it causes expensive damage to the engine.
Although the data analytics revolution is underway, it is still far from complete. That’s because new capabilities such as machine learning and artificial intelligence are still developing. “I expect a vast number of intelligent capabilities to emerge in the cloud,” says Sirosh. These have the capacity to make the effects of data analytics even more profound. “I see the process of data analytics itself being transformed by machine learning,” he says.
“For example, imagine a system that lets any person shape and customise data simply by giving examples of how they want their data to look as a final result. We see automated program generation using machine learning making the entire process of data transformation and analytics dramatically simpler. It will open data analytics to new people, which means new ideas and solutions to the world’s problems,” he concludes.