No tags yet.
Thinking of leveraging data analytics and data science as part of your digital transformation? An automated, back-to-basics approach and data-driven culture may be key.
By Thor Olavsurd
Data analytics is a domain in constant motion. With organizations continuing to invest heavily in analytics to support digital transformations, keeping on top of the latest trends is essential to ensuring your organization is adopting the analytics strategies and tactics required for success in the months and years ahead.
In February of this year, research firm Market Research Future forecast the global data analytics market will achieve a CAGR of 30.8 percent through 2023, reaching a market value of $77.64 billion by the end of the period. At the heart of this increasing investment in data analytics is a drive to become a digital enterprise, says David Schatsky, managing director of the lead trend-sensing program within Deloitte's central innovation team and co-author of Deloitte's recent Pivoting to Digital Maturity report.
"A digital enterprise is one that's continually evolving and always seeking to apply and maximize the value of digital technologies to reinvent itself, to reinvent what it offers to the marketplace, how it delivers those things to the marketplace, and how it operates as a business. A digital enterprise is something we view as in constant evolution enabled by effective use of digital technologies and data," Schatsky says.
For organizations seeking to transform their business through data, the following four analytics trends are worth watching in the months ahead.
Rise of the data citizen
As organizations transform to become more data-driven, most experts and industry-watchers agree that the technology, while not easy, is the simplest element. Changing the culture and organizational mindset around data and its effective use is often the most challenging.
"The most important part of data is people," says Caroline Carruthers, director at consulting firm Carruthers and Jackson, former chief data officer of Network Rail, and co-author of The Chief Data Officer's Playbook and Data-Driven Business Transformation: How to Disrupt, Innovate and Stay Ahead of the Competition. "If you have your whole organization all understanding what you want to do around data and information and marching to the same beat and the same drum, that is so much more powerful than hiding 10 data scientists somewhere in an ivory tower."
Rita Sallam, research vice president in Gartner's business analytics team, agrees.
"Culture is always a huge challenge with any technology: change management, literacy. Do we have a literate enough workforce that can actually take the insights that we're now making available to them through these new technologies and be able to act upon them," Sallam says.
Deloitte's Schatsky adds that to achieve data mastery, organizations must inculcate a mindset among the lines of business that whenever the business is facing a decision or taking an action, the people responsible should consider whether there's data that can help the business do it smarter or better.
"It requires a change in mindset," Schatsky says. "For it to be pervasive, it requires a change in leadership. Leadership must be focused and drive it through the organization."
As a result, they believe that the enterprise will increasingly focus on driving a data-driven mindset and fluency in basic data concepts throughout the organization.
However, Meta S. Brown, president of business analytics consulting firm A4A Brown and author of Data Mining for Dummies, warns against taking it too far.
"I think there are very real limits in how far you can expect executive managers to become analytics experts," Brown says. "A little familiarity with terminology might be reasonable, but I really do mean a little."
By way of example, Brown points to attorneys. Executive management teams are expected to know about the most important laws they must comply with and should be able to read a contract, but they're not expected to advise on complex legal problems. The same idea holds true when it comes to analytics, she says.
"I don’t think it’s reasonable to add onto an executive's load classes in data analysis, for example, or doing data analysis for themselves," she says. "It is my professional opinion that what we should expect of them in terms of learning about data analysis is very minimal. It is our job as analytics experts to more greatly work with them all through the process to get the business information out of them, and our responsibility to translate what they’re telling us about business into analytics terms."
Back to basics in analytics
For the past several years, the conversation around analytics has increasingly focused on cutting-edge technologies such as machine learning, deep learning, neural networks, and other elements of artificial intelligence. While those areas will continue to garner a lot of attention in the coming years, many organizations will get back to basics and extract more utility out of less advanced analytics, experts say.
"It's very shocking to me that I've been to so many presentations and read so many articles where people are talking about some of the most complex mathematics one could possibly imagine, and they're talking about it for businesses that haven't really made much use of much simpler mathematics," Brown says. "If every business in the United States would just thoroughly use what's in Statistics 101, our economy would boom."
Carruthers agrees.
"We almost took our eye off the analytics ball because a lot of people got excited about machine learning and AI and suddenly went, 'Ooh, we have to do all these wonderfully whizzy-bang things,'" she says. "We forgot that actually there is a tremendous amount of value organizations get from analytics."
Carruthers believes that as organizations put analytics into production and seek to really derive value from their analytics efforts, they will again place more emphasis on what can be achieved through more basic analytics and reporting capabilities.
Automation is the name of the game
Still, the increasing complexity of data and what is required to process and analyze it means automation will become more important over the coming years.
"The increased complexity of data in terms of type and the analysis required for success is really pushing the limits of current manual approaches," Sallam says. "As a result, virtually every aspect of data management and analytics content development, up and down the stack, the whole stack is leveraging itself to automate analytic processes, automate the way we get information from those systems to act in an optimized way."
Schatsky agrees, noting that much of the drive toward automation is a result of the scarcity of data scientists, data engineers, and other members of data science teams.
"A lot of data scientists will tell you that they spend about 80 percent of their time on tedious and repetitive tasks like data preparation, feature engineering, selecting algorithms: A lot of things that we've found can be automated to some degree," Schatsky says. "This doesn't mean that data scientists will be put out of work because they've been replaced by machines. What it means is they can be five times more productive, which means a company that has two data scientists can start acting like they have 10 data scientists if they make use of automation."
Organizations will increasingly leverage third-party data
Schatsky says that companies with a more advanced analytic program are going to be making more use out of third-party data in the coming years.
"Being able to make effective use of third-party data goes beyond the normal technical hurdles of data integration, data cleansing, and all of that stuff," Schatsky says. "It includes more market-facing competencies that companies need to develop: continually scanning the landscape to understand what data sources may be available, fostering streamlined processes for evaluating these data sources for producing data, including contracting and legal issues and risk review. This is a competency that companies should really cultivate and invest in."
By way of example, Schatsky points to supply chain management. In the past, he says, companies that operated a supply chain may have managed risk by keeping track of the behavior of their suppliers: how often they were late on delivery, how often the products they delivered met or failed to meet quality standards, etc.
"That's sort of kindergarten-level supply chain risk management," Schatsky says.
More sophisticated organizations may have implemented prior-generation third-party data, like credit rating information, to help determine the risk of working with a supplier or customer.
Today, he says, organizations with more sophisticated analytics capabilities are looking to include third-party data like weather forecasts to round out their risk assessments: Does the organization have suppliers in regions that are vulnerable to disruption due to weather? Or they might include data from social media to help give them clues about shifting patterns of demand. If social media indicators show demand for a certain product is spiking, should the organization increase purchases of raw materials for those products?
"That's the next level of excellence that companies are aspiring to," he says.