How to leverage analytics to become a truly data-driven organization– three things to consider
By Dorman Bazzell and Scott Schlesinger, Capgemini
There is a new role within many leading organizations that’s focused on helping transform data into information, and information into intelligence – the data scientist. It's a bit of an ambiguous role, as not many people understand what they do and how they do it. Yet, it is a role revered as significantly important. Data scientists help the organization chart the path forward by leveraging data and creating sophisticated statistical algorithms in an effort to solve problems in the journey to actionable intelligence. This job is so in demand that McKinsey estimates that there will be a need for up to 190,000 data scientists in the US by 2020.
Today, however, there are very few data scientists and talent available that has in-depth knowledge of how to make the most of data. Among the rest of us, many are eager to quickly take advantage of this trend to tap into the vast amounts of new data coming into our organization by amassing and assessing more data points than ever before to support decision making.
What many have come to realize is that their interest lies less in the trend itself but more so with the analytics. That eagerness, fueled by a desire to immediately derive valuable, actionable insights from data, is unfortunately leading to low impact outcomes. Fundamental key elements needed to truly realize an effective analytics vision are lost in haste.
Through our client initiatives, we believe there are three essential elements that leaders in any business function must adopt in order to ensure ideal outcomes thorough data analytics. These include creating a classification making a distinction between big data and analytics; defining a clear business use case; and optimizing existing infrastructure and tools to support the effort.
Separate Big Data from Analytics
The general idea behind big data is not new. Companies have been looking to take advantage of available data to improve business operations and customer initiatives for years. What’s different today is the vast amount of data types and volume, their entry points, and the velocity at which organizations seek to acquire and analyze this data. What often gets lost through eagerness is the understanding that big data is not just about collecting and managing data sets. Rather, big data must incorporate analytics – the ability to assess and identify outcomes and trends to make more educated, impactful business decisions. To reap the benefits of analytics, companies need organizational rigor around the efficient acquisition and leveraging of big data. At the outset, everyone involved in an initiative must have the same understanding and expectations.
Define the Business Use Case
The big data and analytics journey begins with a relevant, clearly defined and well-understood business use case. Yet, one of the primary causes of frustration is jumping into an analytics initiative too soon without clearly defining what it is the company wants to accomplish.
As companies seek to leverage data – be it internal, external, structured, or unstructured – to increase operational efficiency, improve profitability, increase market share, analytics offers the ability to glean insights on areas of the business that previously were a distant reality. What are the insights that you hope to obtain? What are you looking to measure? What do you hope to achieve through analytics for your department or enterprise? These questions must be answered before you start the journey, otherwise you will not arrive where anticipated, and your ROI expectations will be dashed.
Organizations must understand why and where it makes sense to leverage big data. Establishing a set of prioritized use cases at the functional-area level will ensure clarity and strategic guidance for each initiative and provide success criteria for measuring accomplishments. Use cases help create business cases that provide both IT and the business a common framework to address the ROI questions expected from organizational leadership to investors.
For example, it’s fairly straightforward for a clothing retailer to track the number of men’s ties sold. But, it is less obvious for them – unless they make use of the social data conversation – why particular ties where such hot sellers. Oil companies might use machine sensor data to determine when a blow-out preventer might fail and what is the likely cause.
This use of data is improving worker safety; preventing environmental damage, and eliminating potential costs.
Optimize Existing Technology Investments
Many companies looking to quickly take advantage of analytics rush into purchasing decisions. They seek out the latest and greatest analytics software and tools. But, by not having first created a clear business use case or put the proper data governance frameworks in place, they’re likely to attain a low return on investment from the start.
Companies should first look at their current IT infrastructure and enterprise planning systems and seek ways to pair those with external sources such as social media and sensors. This will help them to gain initial data insights that will likely ensure greater flexibility to scale analyses of subsequent data sets that will have a broader impact on the business, its partners and customers. More importantly, this practice will provides a path to more efficient data optimization efforts to take advantage of open source data tools, such as Hadoop, to store and assess larger data sets.
Attaining real time answers from data is the ideal outcome of analytics. Acting on these three critical elements requires an appreciation for patience – they require time. Eagerness to employ analytics and attain quick results runs the risk of creating more hardships than benefits for the business in the long run. In order for organizations to net their ideal outcomes with analytics, these blocking and tackling steps are essential. They no doubt take time, but their return on investment will be invaluable to any analytics initiative.