Big data best practices, and where to get started

Big data analytics is on the ‘to do’ list of every large enterprise, and a lot of smaller businesses too. But perceived high costs, complexity and the lack of a big data game plan have hampered adoption in many South African businesses.

By Mervyn Mooi, Director, The Knowledge Integration Dynamics Group

Big data as a buzzword gets thrown around a great deal these days. Experts talk about zettabytes of data and the potential goldmines of information residing in the wave of unstructured data circulating in social media, multimedia, electronic communications and more.

As a result, every business is aware of big data, but not all of them are using it yet. In South Africa, big data analytics adoption is lagging for a number of reasons: not least of them, the cost of big data solutions. In addition, enterprises are concerned about the complexity of implementing and managing big data solutions, and the potential disruptions these programmes could cause to daily operations.

It is important to note that all business decision makers have been using a form of big data analytics for years, whether they knew it or not. Traditional business decision making has always been based on a combination of structured, tabular reports and a certain amount of unstructured data – be that a phone call to consult a colleague or a number of documents or graphs – and the analytics took place at the discretion of the decision maker. What has changed is that the data has become digital; it has grown exponentially in volume and variety, and now analytics is performed within an automated system. To benefit from the new generation of advanced big data analytics, there are a number of key points enterprises should keep in mind:

  • Start with a standards-based approach. To benefit from the almost unlimited potential of big data analytics, enterprises must adopt an architected and standards-based approach for data / information management implementation which includes business requirements-driven integration, data and process modeling, quality and reporting, to name a few competencies.


(Image not owned by KID)

In context of an organized approach, an enterprise first needs to determine where to begin on its big data journey. The Knowledge Integration Dynamics Group is assisting a number of large enterprises to implement their big data programmes, and we have formulated a number of preferred practices and recommendations that deliver almost instant benefits and result in sustainable and effective big data programmes.

  • Proof of Concept unlocks big value. Key to success is to start with a proof of concept (or pilot project) in a department or business subject area that has the most business “punch” or is of the most importance to the organisation. In a medical aid company, for example, the claims department or business might be the biggest cost centre and with the most focus. The proof of concept or pilot for this first subject area should not be a throwaway effort, but rather a solution that can later be quickly productionised, with relevant adjustments, and reused as a template (or “foot-print”) for programmes across the enterprise.
  • Get the data, questions and outputs right. Enterprises should also ensure that they focus on only the most relevant data and know what outputs they want from it. They would have to carefully select the data/information for analytics that would give the organisation the most value for the effort. Furthermore, the metrics and reports that the organisation generates and measures itself by, must also be carefully selected and adapted to specific business purposes. And of course, the quality and trust-worthiness of sourced data/ information must be ensured before analytical models and reports are applied to it.
  • Get the right tools. In many cases, enterprises do not know how to apply the right tools and methodologies to achieve this. Vendors are moving to help them by bringing to market templated solutions that are becoming more flexible in what they offer, so allowing organisations to cherry pick the functionality, metrics and features they need. Alternatively, organisations can have custom solutions developed.
  • It’s a programme, not a project. While proof of concepts typically show immediate benefits, it is important for organisations to realise that the proof of concept is not the end of the journey – it is just the beginning. Implementing the solution across the enterprise requires strategic planning, adoption of a common architected approach (e.g. to eliminate data siloes and wasted / overlapping resources), and effective change management and collaboration initiatives to overcome internal politics and potential resistance and ensure the programme delivers enterprise-wide benefits.




Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Google+ photo

You are commenting using your Google+ account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )


Connecting to %s