Although big data gets bigger by the day, the availability of increasingly-affordable storage solutions and user-friendly applications is putting the effective management and analysis of all that data within reach for every part of the organisation.

Finance is analysing data to leverage advantage and manage risk; marketers are deploying models to better understand the consumer and improve customer lifecycle value; planners are accessing public data sources to anticipate demographic trends and production needs in the years ahead. And in sales, HR, manufacturing, logistics and the boardroom, we see analytics increasingly being adopted for faster access to business insights.

But enthusiasm for access must be tempered by sound management of the whole analytics environment. Centralised standards and processes must be established to ensure against ill-defined projects, overlapping effort and non-compliance. Without sensible disciplines in place, it’s all too easy to waste valuable resources building models that never get used.

As widespread demand for models increases, SAS recommends setting up an enterprise level ‘analytics council’. Oversight should be with a C-Level executive who is already analytics savvy, and members of this council – call it a committee, if you like – should be made up of representatives drawn from every part of the organisation, including those who are new to analytics.

Representatives must be committed to the organisation’s high-level objectives, and clearly understand how their own operations complement the whole. Further, and very importantly, they must also be able to articulate the business needs of their own sections and visualise what could be improved.

The council’s first action is to draw up policies and procedures for all analytics activity. These must include prioritising workloads and ensuring that no individual project is started without clearly agreed goals and measures for success. Wherever possible, development should specify models being embedded in systems that can trigger automatic decisions. Audit trails must identify what libraries and data are being used for which models, and histories must be kept of who did what, when and why. These records will not only facilitate model modification as business conditions change, but will also be important in helping resolve any issues that may arise. In addition, models must obviously comply with industry regulations and be secured against unauthorised interference from within and espionage from outside.

Records of activity should be communicated appropriately. And the council should invite input from across the organisation and formally respond to user suggestions for the ongoing improvement of models. Openness on the part of the council will help everyone in the organisation see analytics as a crucial part of the business and a contributor to success.

ABOUT GEOFF Beynon//

Geoff Beynon is the general manager of SAS Institute New Zealand. SAS is a global leader in business analytics software and services.