In our ‘age of acceleration’ as described in Thomas Friedman’s latest book, speed is the number one business currency.
To compete in today’s unpredictable and fast-changing world, businesses are looking to artificial intelligence and advanced analytics to make trusted data-driven decisions. In fact, Forrester has coined insight-driven businesses as a new breed of companies that “use data, analytics, and software in closed, continuously optimized loops to differentiate and compete” and they are expected to steal $1.2 trillion in 2020 from their competition.
While enterprises are increasing investments in AI and analytics, there’s still a long way to go before we’ll see organizations reap the true benefits of becoming a data-driven business. That’s because analytics, like any other major technology, moves through cycles.
We’re still in early phase of the analytics evolution
The order of evolution or progression in technology goes like this: You have early adopters, followed by a crest of adoption and then growth drops off. It takes a very long time to move through the entire cycle. Case in point, according to the SAP Digital Transformation Executive Study, only 3 percent of study respondents have completed digital transformation projects spanning their organizations.
So where do most organizations stack up in the analytics evolution today? Using Gartner’s definition, we move through a progression from descriptive to diagnostic to predictive – with prescriptive analytics being the most enlightened stage. Achievement of prescriptive analytics is characterized by industry leaders using advanced analytics to uncover insights that enable them to determine future actions.
Most companies operate within the descriptive and diagnostic stages, using basic data warehousing and BI approaches to get quick views on what HAS happened. Predictive analytics is when organizations project what WILL happen … graduating from rearview mirror to human intervention combined with the automation of repetitive patterns through the application of predictive machine learning (ML) models.
Data complexity is hard to solve
So why are most companies not further along the analytics progression? Frankly, most enterprises are drowning in an abundance of data types and sources – many of which contradict each other as data size and ingestion rates are also on different levels. Moreover, many organizations are not taking advantage of new technologies that can unlock and manipulate data. If you look at the market, most companies operate within a rigid data environment, one that’s slow and methodical, characterized by basic data warehousing, descriptive analytics and use of dashboards.
In an automated environment, your data is integrated into a system without human intervention. For this to happen, algorithms need to rely on good data. With prescriptive analytics, algorithms are smart enough to digest the data and put it to use by applying self-service ingestion and rationalization. The data you feed these algorithms will have a significant impact on the outcome. As such, data quality is more important than ever and this is often where businesses struggle.
Skill-set constrained IT organizations
Furthermore, many IT organizations lack the adequate resources and skills to fully understand the capabilities of advanced analytics and how to implement it in a practical manner across their businesses. We are seeing some digitally advanced companies move toward sophisticated analytics with ML tools for categorizing data to help make predictions. For example, MasterCard is using AI to improve the accuracy of its fraud detection. However, the vast majority of business-deployed AI systems are yet to make business-critical decisions on their own.
To realize the full potential of advanced analytics, your organization will need to design a roadmap for becoming a data-driven business. Consider the following to optimize and manage your data, and find the necessary and sought-after data science talent:
Start with building a solid data foundation
Advanced analytics will only work if you have clean and reliable data. Begin with building a solid foundation for collecting, managing and analyzing data so you can get to trusted insights. By operationalizing and governing your data, you’ll be able to automate it with little to no human interaction.
The first step is to assemble your data into your ‘information supply chain.’ Just like a manufacturing supply chain, it involves taking the raw materials (data) from the source and adding value at each step in the journey. Once assembled, you can leverage AI tools to automate the critical data processes to assure the governance, quality and reliability of your data, as well as the clean-up of your ever-increasing data sources.
Embrace new technologies
Accessing disparate types of data and bringing them together is a fundamental challenge requiring reliable data connectivity as a foundational technology. Not only do you need the right technology to provide the physical and logical connection to data that’s coming from your on-premises and cloud-based applications, but today, you also have to consider non-traditional big data and IoT sources bringing in ‘new’ data.
These sources are even more open-ended in terms of fragmentation, governance, quality and reliability. Since there are very few standards defined or consistently implemented across all systems, your data drawn will likely be ‘apples and oranges,’ making it difficult to aggregate without significant effort to standardize and harmonize the data.
To bring advanced analytics to bear on the consolidated data, you must validate and rationalize the data across the different systems. Therefore, leveraging pre-built connectors that already know the protocols and structures for extracting the right data will be invaluable. Once you have the right connectors to access and collect the data, the fun begins … because it becomes easy to leverage new big data technologies to manipulate large amounts of data.
Get the right resources
As we’ve consistently seen in the media, data scientists are a highly sought after resource and the best data analysts bring conceptual thinking, deductive reasoning and decision-making expertise. This blend of technical programming, analytics know-how and solid business savvy will remain in short supply. My advice? If you find good ones, hang onto them.
With the deployment of more ML tools and the automation of more business analysis and reporting tasks, businesses will have the opportunity to do more with technology and free up resources for more critical-path and strategic roles.
The future of advanced analytics promises to be transformational and it’s paving the way for new business models that will impact how we live and work. For now, we remain early in the analytics evolution. That said, businesses will need to quickly transform themselves to be truly data-driven – seeking out the best data science talent and managing smart machines to realize the full potential of AI.