Data science solves math problems, but it's up to business people to see how the answers to those math problems can also solve business problems. Data-driven cultures can help tap into valuable insights hidden in plain sight.
Quick Links1. Have a vision
2. Locate and organise your data
3. Quantify it
4. Stop, allocate and collaborate
5. Ask yourself
Companies have begun to realise that the data they collect about their customers – and collected by other organisations about their customers – can unlock new ways to earn profit. In some cases, the data itself may be more valuable than the products they sell. Free applications for cell phones, for example, are often paired with data collection software that can be analysed and sold to other firms for a profit.
Gathering massive amounts – terabytes at a time – of consumer data, then querying that data for insights would have been impractical ten years ago. No longer. Modern tools like Apache Hadoop democratised analysis once reserved for companies with billion-dollar research teams and supercomputing resources. And data scientists – experts in machine learning, artificial intelligence, database management and statistics – are rapidly becoming rock stars of the business world.
The essential skill, however, isn’t an IT skill, but a business skill. In fact, it may be best to avoid thinking about data science as a technology at all. A survey of large firms by New Vantage Partners indicated that 95% of managers view cultural problems – inertia and organisational alignment – as the chief impediment to implementing a data science project. Technology is simply a means to an end when leveraged to answering interesting questions.
A data-driven culture is not an experiment. It requires full commitment to the idea that all decisions are driven by data. This data culture vision is one that embraces business objectives and growth. The data stored within your business can help drive the growth of your organisation. Arguably, every company in existence can derive some sort of benefit from the information at their fingertips. Data can be utilised to not only understand your customers, but your employees and your potential hires. The aim is to make data all-pervasive within your organisation. To put it bluntly, data should drive most of your business decisions. However, data is only as effective as the people using it. Management must lead this process, bring in the right people, and ensure everyone in the organisation is on board.
Intermountain Healthcare Inc., a group of hospitals in Utah and Idaho, re-defined their idea of success by comparing their post-surgery infection rates with that of the national average. Management and surgeons determined through data that they would exceed national standards.
The data is there. You just need to find it, clean it and make it useable. A major hurdle for many organisations is trying to grasp the complexity and magnitude of their data, let alone making a move towards data. This is not a simple task.
First, you must organise the data. This can be done in groups which relate to particular areas of interest or business function. The idea is to turn all your data into quantifiable values – that usually means digits. Once the data has been grouped, it can then be systematically categorised and coded so that everything within one category shares some meaningful commonality, and that the differences between one category and the next is distinct. Finally, once the data is coded and categorised, it needs to be displayed and organised so that is can be used and interpreted.
Jeff Liberman, COO of Spanish-language media company Entravision Communications Corp. was perplexed the first time he heard about big data. It wasn’t until he took off his broadcaster’s hat and took on an analytical mindset that he realised the importance of data and information. He needed to reinterpret what was displayed to him for it to make sense.
Once you have achieved this, your organisation will be on its way to quantifying the worth of that data. Treating data like a business asset is a trending field known as infonomics. It has come about because companies recognise that they use data daily to make business decisions, and that means there is value in those numbers. This is an important business and psychological component towards a data-driven culture as it gives management and staff the impetus to take things forward, and to ensure that internal data maintains a high level of quality so that decisions based on the information is sound.
There’s no consensus around how to deploy data scientists in an organisation. Chuong Do, head of data at the online learning platform Udemy, notes that many organisations centralise their data science teams because the principle bottleneck to developing a data science section is recruiting. However, centralised teams too far removed from decision makers may lack enough context or influence within the enterprise to turn their insights into action. “In some cases, this can lead to an unhealthy dynamic where data science is treated as a support function, answering questions from product managers rather than operating as true thought partners and proactively driving conversations from a data-informed perspective.” Says Do.
Often, data science emerges organically in larger organisations, section by section, as individual managers realise that they have valuable information to mine; this too presents a risk of creating siloes within each unit so that none communicate with one another. It is of utmost importance that the organisation encourages interdepartmental collaboration and communication. The goal is to maximise the potential of the data available.
The Amsterdam City Council experienced this first hand. On appointing their first-ever Chief Technology Officer, Ger Baron was faced with the daunting task of organising 12,000 decentralised datasets from different government departments.
More frequently, firms develop insights from their data that their managers cannot convert into action. Data science teams often cannot communicate the value of their finds to decision makers; or, data sets are too “unclean”, filled with transcription errors or fields with incomplete information. Data cleaning is a tedious and expensive task that itself can be hard to automate. Managers would do well to ask some questions of their organisations then, before beginning a data science project.
Do you have clearly defined goals? Are you starting a data science project because everyone else is doing it, or to answer a clear question? What are your expectations of a data science team?
What will my team look like when it’s full? Do you need statisticians? Machine learning experts? Analysts? What composition gets you to your goal?
Who is going to lead this effort? Do you have an expert identified? Do you trust this person? Will you listen to this person?
When and where do they go? Can you recruit effectively and quickly? Is your human resources function up to the task? Will your data scientists integrate directly with your operations or research divisions, or operate as a stand-alone function?
Does your culture match your goal? If your company will ignore the recommendations of a data science team, it may be pointless to try to build one.
Is your data interesting? This is as much a recruiting question as a data science question. Solid analysts are in high demand. If you have no interesting questions to ask, you won’t find people ready to work on your data.
Transforming the culture of your organisation into a data-driven one has become easier in recent years. While a data-driven culture naturally encapsulates data science, it also requires other less digital, and more human factors.
Companies such as DataVLT have been working hard at democratising data analytics for businesses of all sizes. DataVLT exists to help businesses to unlock the power of their data and to take their data analytics to the next level. Find out how your business can gain invaluable decision-making insights here.
DataVLT is an affordable, on-demand analytics platform secured by blockchain technology. It is designed to simplify the complexity of data science. Backed by artificial intelligence and machine learning capabilities, DataVLT empowers enterprises to make meaningful sense of their big data and scale cost efficiently. Essentially, it is an end-to-end data/information management platform.
Learn more at www.datavlt.com