When Businesses Go Around IT for Analytics
Business people have been going around IT organizations for about as long as they've been dealing with them. In a recent report, Gartner tackles the emerging problem of going around IT for analytics.
Going behind IT's back is one of the best-known tropes in IT and business management. It's what happens when the business -- frustrated and exasperated with IT -- takes things into its own hands.
This has been a problem for about as long as business people have been dealing with IT organizations. In a recent report, "How to Overcome Business Bypassing IT for Analytical Solutions," Gartner tackles the problem of business people going around IT for analytics.
There's no shortage of analytics products and services to choose from. Thanks both to the complexity of analytics and the traditional slowness of IT, it's more tempting than ever for business users to procure their own analytics technologies. The result is what Jorgen Heizenberg, research director for data and analytics with Gartner, calls a "contextual version of the truth" -- business insights that are suitable for the tasks and responsibilities of individual business users.
This contextual version of the truth is not suitable for enterprisewide decision making. The good news is that organizations are figuring this out on their own.
Learn How to Compromise
By 2019, Gartner projects, half of all traditional (i.e., centrally organized and managed) analytics programs will be a hybrid model in which power and responsibilities are shared between IT and "local" leaders in each business domain.
Most businesses haven't yet worked out the kinks of this hybrid model, however. Too many are still caught up in turf wars between business and IT. In "How to Overcome Business Bypassing IT for Analytical Solutions," Gartner offers an assessment of this problem and, crucially, several recommendations for addressing it.
Gartner's first and best recommendation is an obvious one: compromise. Business people have both the ability and the self-serving incentive to go around IT to procure their own analytics products and services. IT lacks the power to change this; executive fiat can't necessarily change this.
Instead of trying to legislate reality, Heizenberg argues, work with it: "Grant autonomy and support to business leaders when buying analytical solutions, but actively work with them to ensure alignment with central IT, [and possibly] other business units, as well as quality, security, and privacy standards."
Focus on Business Value
Next, work to align self-service and domain-specific analytics efforts with the organization's own priorities and governance model. Make sure, too, that self-service and domain-specific analytics efforts are aligned with overall business value -- not just the short-term self-interest of the business domain.
"Lack of awareness of ongoing data initiatives within IT, or impatience with a lack of progress, could potentially result in business leaders investing in analytic capabilities that are already part of the IT infrastructure, or that are part of the [road map] for future updates," he writes.
Support Self-Service
Heizenberg hazards a third recommendation: IT and the line of business should work together to evaluate external analytics service providers to close gaps -- be they in existing tools or technologies and skills. "Self-service and domain analytics solutions drive the need for different external support. Data and analytics leaders should evaluate service providers including software vendors and new entries to the marketplace," he says.
As Heizenberg sees it, external service providers can help support self-service governance efforts, are often better equipped to support emerging platforms and technologies (such as Apache Spark), provide missing skills (or augment existing skills) such as R, Python, and Scala, and if necessary, provide "turnkey business solutions."
Will these recommendations work? It's difficult to say. The practice of going around IT became popular in the 1980s with the advent of PCs and, especially, spreadsheet programs. Data warehouse architecture was created in part to address the problems -- inconsistent numbers, lack of historical data, a complete absence of governance -- created by the proliferation of spreadsheets -- or spreadmarts, as they came to be called. Growing demand for analytics has exacerbated this problem.
If there's one reason to be optimistic, it's that the business realizes it needs help. "Domain analytics and self-service may require support that is not available within the organization. Data and analytics leaders should work with the business domains and external service providers to ensure new tools, technologies, and use cases are being supported to service the business demand," Heizenberg concludes.