Delivering value using data is top of mind for business leaders today, yet many are unsatisfied with the results from their data initiatives. It’s no wonder: research1 at the Massachusetts Institute of Technology Center for Information Systems Research (MIT CISR) identified nearly 100 obstacles that impede progress as companies attempt to move from data to insight to action and subsequent business value. Obstacles are problematic, particularly because business leaders are held to tight timelines these days; in other words, speed to market matters.
Suzanne Hoffman of Tableau Software briefed Dr. Wixom on March 18 in The Briefing Room. Click here to hear the conversation.
The good news is that MIT CISR research also identified practices that help companies move either through or around obstacles to value. For now, let’s focus on one: agile methods and approaches. This refers to development practices that actively engage multi-disciplinary teams (which include business users), scope delivery products and services to include only the most meaningful functionality, deliver desired business value in frequent intervals and readjust to meet new or evolving needs.
Surprisingly, the world of data was late to the game regarding agile. For years, many organizations had development shops that leveraged agile methods and practices for development projects other than data projects like business intelligence. That has changed. Today, agile is increasingly used to create and evolve data-based products and services. In fact, in research2 conducted for the Society for Information Management’s Advance Practices Council, agile emerged as a frequent practice among best practice business analytics initiatives. Leading companies have figured out that agile is critical to embrace for a number of reasons, some described below.
Agile helps companies identify decision-making requirements more quickly.
Identifying the right business requirements for decision-making products and services can be tough, particularly when business users are new to business intelligence and analytics and simply don’t know what to ask for. With agile, business users roll up their sleeves and have “skin in the game” from the outset of the development process. In this way, business users learn and grow in their “data savvy” – and shape deliverables in real time as development evolves. One net result is a positive impact on delivery schedules. For example, an insurance company moved to an agile development process for all business analytics projects, adopting techniques such as paired programming, story walls and test-driven development. The company realized a four-fold increase in analytics usage over two years, attributed to increased development productivity.
Agile facilitates good communication, which results in the right decision-making products and services.
In 2012, the Business Intelligence Congress3 (BIC) surveyed 446 business analytics recruiters, and they discovered that the top five skills4 desired by employers in descending order include: 1) communication, 2) SQL and query, 3) basic analytics, such as descriptive statistics, regression and ANOVA, 4) data management, and 5) business knowledge. Interestingly, communication skills ranked #1 in the BIC survey from 2010 as well, although other skills gained and lessened in importance. This makes sense as communication is incredibly important in all facets of decision making; it is required for tool development as well as effective tool usage. Agile supports communication in a number of ways, such as prototyping, collaboration across multi-disciplinary teams and techniques like story boarding.
Agile facilitates user adoption.
Countless academic studies have found that business users embrace decision-making tools that are easy to use, useful and engaging – relative to their other choices. This means that when a user is offered a new decision-making tool, they will subjectively compare it with their current tools and processes (even if the process is manual). If the old process is easier, more useful and/or more fun for the user, then the chance of that user migrating to the new decision-making tool is quite low. Agile engages business users in ways that ensure that the tool characteristics that matter to the users are ultimately delivered. For example, if “easier” translates into a tool that is Web-based, visual and flexible, then users will surface these as important requirements during the development process.
Agile kick-starts the change management process.
At the end of the day, data products and services generate value when users take some action based on the insights that they uncover, which typically means doing things in new ways. This may require changes to business process, organization roles, incentives, skills, organizational structures, etc. Too often the leadership and commitment for change is missing and value is lost because action never happens. With agile, users learn about the possible implications of data products and services as they are being developed; buy in and understanding builds along the way. Thus, users are much more willing and able to embrace change and use the tools as intended once development ends.
Some companies may feel that they are not set up for agile – and have a steep learning curve to climb. One option is to begin instituting agile practices gradually, beginning with select practices, such as prototyping and heavy user involvement. Another option is to build or expand agile capabilities by buying them. Fortunately, outsourcing5 is now an option for building/deploying/using data products and services so that companies can leverage offerings like cloud-based software quite effectively to support agile efforts.
Given the great number of benefits that result from agile combined with the accessibility of agile-friendly resources, the use of agile methods and approaches will only continue to soar in the data space.
1 Wixom, B.H. and Beath C.M. “Winning the Data Race” MIT Center for IS Research, Research Briefing, XIV, 3 (March 2014), available publicly in July 2014 at http://cisr.mit.edu/
2 Wixom, B.H., B. Yen, and M. Relich. “Maximizing Value from Business Analytics,” MIS Quarterly Executive, 12, 2 (2013) 37-49.
3 See http://www2.commerce.virginia.edu/bic3/ for more information on the BI Congress 3
4 Wixom, B. H., T. Ariyachandra, D. Douglas, M. Goul, B. Gupta, L. Iyer, U. Kulkarni, J. G. Mooney, G. Phillips-Wren, and O. Turetken. “The Current State of Business Intelligence in Academia: The Arrival of Big Data,” Communications of the AIS, 34, 1 (2014). http://aisel.aisnet.org/cais/vol34/iss1/1
5 Watson, H.J., B.H. Wixom and T. Pagano. “Analytics Outsourcing: The Hertz Experience,” Journal of Business Intelligence, 18, 4 (2013) 4-7.