5 ways to maximize the value of data analytics

It’s one thing to gather large amounts of data and apply analytics to it; lots of organizations are doing that. It’s another thing entirely to gain optimum business value from that data and analysis.

Enterprises that have invested heavily in analytics tools might be doing so without finding ways to ensure business value actually results from their efforts. This could be happening for any number of reasons.

Here are some tips on how to ensure that investments in analytics pay off by delivering insights that make a difference, rather than just generating attractive reports that don’t say much.

Align analytics with business goals

More closely aligning IT efforts with the organization’s business goals is a key IT leader mandate in generating more value from data.

“Data analytics need to solve real business problems,” says Dan Simion, vice president of artificial intelligence and analytics at technology consulting firm Capgemini, who advises clients on the best ways to maximize insights and value from their data and analytics. “Starting with use cases that are business-specific can be a great approach to earn buy-in from more stakeholders who may be outside of IT.”

This strategy helps the rest of the organization to see the value across various functional areas and business units, because the data is driving clear results that are understandable in business terms. “Starting with the business problem, developing a business case, and following up on it is pivotal to uncover the value from the data,” Simion says. As part of this process, the alignment between the business and IT is critical, he says.

IT and data analysts need to work with the business units to find real-world wins and internal use cases, says Gary Kern, CIO at financial services provider Middlefield Banking Co. “This helps everybody understand the real value and the teamwork benefits of getting there,” he says.

Middlefield has struggled to get some business units more active in using data-driven decisions and new processes based on more detailed information, so Kern has worked “to find early adopters within those departments that we can partner with to get wins and sell the value to others in that area,” he says.

Healthcare provider UnityPoint Health began investing in analytics tools years ago and they continue to pay off for its regional and community network hospitals, clinics, and homecare division, helping manage population health and addressing future problems, says Laura Smith, CIO.

Strong alignment with business leaders is a big reason for the success. “It’s imperative to partner with the business to understand the problem to solve or opportunity to be achieved,” Smith says. “What, specifically, are we trying to accomplish? What important data are we missing right now?”

A great way to start building a relationship with the business is to

research the problem or opportunity by meeting with stakeholders and performing observations in the field.

“For example, our analytics team built a model to try to reduce the number of patients who had to be readmitted to our hospitals,” Smith says. “We started by working with the business to understand what questions the model needed to answer. Those included: Who are the right patients to target for intervention? What action should we take, and when should we take it?”

As a result of this model, one hospital decreased its 30-day readmission rate by 44% within two years, surpassing internal performance targets.

Get key executive sponsors on board

Having executive sponsors or stakeholders who can push for outcomes and insights from data analytics can help generate greater value.

“This executive champion drives adoption across the organization, and can help to mold the operating model to allow people to take action with the insights that have been drawn from the analytics,” Simion says.

With a high level of buy-in and support, “the business can start to activate the findings and gain value from them,” Simion says. “If you simply build a report from the findings and nobody takes any action, the business won’t realize any value. The data will generate the insights, which will be used for business decisions at all levels within organizations, from tactical actions to strategic decisions.”

Middlefield Banking formed a Data Governance Council (DGC), which includes the CIO, CFO, CMO, and two or three other senior managers in data-intensive areas, Kern says.

“This group meets monthly to discuss ‘one version of the truth’ issues, data clean-up, data quality, evolving analytic efforts, and other high-level concerns involving the analysis of information and ownership of data,” Kern says. “The DGC allows us a way to escalate concerns and assure there is a decision-making body in place to direct efforts at the senior level.”

Update your operating model — and measure its success

Although the drive to being data-centric is common among enterprises in the age of the digital business, many companies still fail to grasp the true value of information.

“Companies need to move away from operating based on a ‘gut feeling’ and move toward becoming insights- and data-driven organizations,” Simion  says.

Having a data-driven operating model creates a far greater probability of success, Simion says, and enables organizations to see value from their data analytics faster — with a clearer path and vision of how to achieve their goals.

“The data, through the insights, will power the decision-making process,” Simion says. “Through the new operating model, the people within the organization will be motivated to change behaviors, [and] the value of the data will be achieved at a faster pace.”

But to recognize the value resulting from a particular insight or piece of data, companies need a framework to measure success. “This helps organizations to assess their current progress, make adjustments, and optimize the way they are tracking toward their data analytics goals,” Simion says. “By demonstrating the amount of value and results driven by data analytics through a clear measurement capability, it will help data leaders to showcase the return on any analytics investment.”

Establish data pipelines with business value in mind

Getting value out of data doesn’t happen overnight or by some magical formula; it takes time and effort.

About 20 years ago, Lonnie Johnson, CIO at healthcare organization KVC Health Systems, decided to start developing a long-term analytics strategy at the company that has paid off over the years. The first step was to organize data in a relational database, which enable the analytics team to categorize existing data points.

“We normalized the information by cataloging lines of business, offices, programs, chronological identifiers, transaction types, and a host of characteristics about our patients,” Johnson says. “We gathered and connected information from a number of standalone databases and spreadsheets.”

The team then created digital forms, applications, and user interfaces to transform the company’s paper documents. It also created interfaces for those documents as a way of entering information into databases going forward.

“In our user interfaces we enforced data integrity and learned to autofill fields as much as possible,” Johnson says. “We included the user community heavily in the development of these digital interfaces, to ensure we were capturing actual business value. We still practice this today.”

The team created custom query builders in applications, which enables users to pull information from select fields based on descriptions of data points. “This freed up the data team to focus on more advanced analytics,” Johnson says. “We also encouraged users to give feedback about the query builder to help us better organize information.”

The team began capturing large amounts of text and form data in NoSQL databases for both rapid development and future natural language processing. “If you use digital forms for surveys, legal documents, customer information, or any other document that may change at any time, using NoSQL can expedite your data capture and free up developers for other more innovative tasks,” Johnson says.

The company invested in data science acumen and tools, with the goal of developing these needed skills inhouse. “We also found a partner we could work with on a regular basis to help us craft solutions using machine learning for predictive analytics,” Johnson says. “This range of onsite [skills] and deep outside experience has developed into a new service, which generates an ongoing supply of actionable insights.”

Leverage cross-functional partners or teams to improve data accuracy

This cuts across all the areas previously covered, especially aligning analytics with the business and updating the operating model. The analytics team should be in regular collaboration with business users to help ensure value through higher-quality data — or include business users as part of their cross-functional teams.

“Partnering closely with business teams in your organization creates an added layer of protection for data accuracy, which improves how both the data and business teams leverage the data they see,” says Jessica Lachs, vice president of analytics and data science at online food ordering and delivery platform DoorDash.

“When more teams are looking at the same data, you have more eyes to spot anomalies that automated alerting might miss,” Lachs says. “Close partnership also ensures that the data team builds business intuition to better understand the practical applications of the data they manage.”

This enables the team to be autonomous and make better decisions about accessibility, accuracy, and scalability based on business needs, says Lachs, who oversees an analytics team of 85 people.

Another key is to treat data as the currency to evaluate business decisions and trade-offs.

“We believe that by quantifying as many things as possible, we can best evaluate trade-offs, determine what’s working and what we need to improve, thereby maximizing our impact and building a better product,” Lachs says. “To do this, we must have current and accurate quantifications of our key business levers, which is a critical part of my team’s roadmap.”

From there, “we can use data to create a common internal currency that allows us to evaluate and compare trade-offs in like terms — for example, whether it would be better [to] lower delivery fees by $1 or improve delivery times by five minutes,” Lachs says. “If you can frame the question in like terms, such as in terms of incremental orders, then the tradeoff becomes more clear, and so does the value to the business.”