5 big data best practices

Analytics projects that use big data or advanced analytics grow increasingly popular but present a heightened risk of failure, says Gartner, Inc. Analytics leaders can improve the likelihood of success by following five best practices recommended by the consulting firm.

"Although big data and advanced analytics projects risk many of the same pitfalls as traditional projects, in most cases, these risks are accentuated due to the volume and variety of data, or the sophistication of advanced analytics capabilities," says Alexander Linden, research director at Gartner. "Most pitfalls will not result in an obvious technical or analytic failure. Rather they will result in a failure to deliver business value."

Gartner also predicts that, by 2018, 50% of business ethics violations will occur through improper use of big data analytics.

Failure to properly understand and mitigate the risks of poor big data execution can have a number of unintended and highly impactful consequences. These include loss of reputation; limitations in business operations; losing out to competitors; inefficient or wasted use of resources; and even legal sanctions.

5 essential big data strategies

Key best practices include:

1. Linking analytics to business outcomes through benefits mapping—Analytics must enable a business decision maker to take action, and that action should have a measurable effect—whether the effect is directly or indirectly achieved.

Linking analytic outputs to traceable outcomes using a formal benefits-management and mapping process can help the analytics team navigate the complexities of the business environment, and keep analytic efforts both relevant and justifiable.

2. Investing in advanced analytics with caution—Many organizations believe that big data automatically requires advanced analytics.

However, the data-crunching power required to manage the big data characteristics of volume, velocity, and variety does not inherently require any more sophisticated algorithmic processing. The complexity of the analytical question to be addressed should drive the need for advanced analytic tools. Gartner says that in many cases desired outcomes can be achieved without resorting to more sophisticated analysis.

3. Balancing analytic insight with the ability of the organization to make use of the analysis—Because analytics can only be beneficial in organizations that are willing to embrace change, it makes sense to limit investment in analytics to a level that matches the organization's ability to use the resulting insights.

Analytics may not be the most suitable approach:

• If pertinent data is absent.

• When high levels of ambiguity are present.

• Where there are entrenched opposing points of view.

• In highly innovative or novel scenarios.

In these cases, scenario planning, options-based strategies, and critical thinking should also be incorporated into analytical approaches to better support the organization's ability to take action.

4. Prioritizing incremental improvements over business transformation—Using big data and advanced analytics to improve existing analyses, or to incrementally update and extend an existing business process, is easier than using them to deliver business transformation. That’s because there are fewer dependencies to overcome to ensure success.

Care should be taken to validate the level of overall change required. In some cases, deep reform of the business strategy may still be necessary—for instance, when a new disruptive vendor enters a market, when technology innovation changes the business model, or when an organization has become dysfunctional.

5. Considering alternative approaches to reaching the same goal—Few objectives can only be achieved in one way.

Statistical modeling, data mining, and machine-learning algorithms all provide means of testing ideas and refining solution propositions. Big data and advanced analytics help validate proposed hypotheses and open an even wider range of potential approaches to addressing corporate priorities. Not all problems even require a fully engineered analytical solution. Investment may be better targeted on human factors, re-education, or reframing the problem.