Big data does not make business smarter

data science

Photo: iStockphoto

Michael O’Connell argues that analytics and data science, rather than just big data, is what will provide businesses with the leverage they need for competitive advantage.

For the past few years, big data has been driving the digital revolution toward the ‘smart everything’. The basic idea behind ‘big data’ is that everything consumers and businesses do is now leaving a digital trace – which can be turned into smart insight. While this is not wrong, this reasoning leaves out the most important part of the process i.e. making actionable information out of the raw data.

‘Big data’ has become so prevalent in the business lexicon that it now seems to refer to any data-related project. The confusion had its benefits in the early days at it helped businesses understand the value of their data. But the downside is that data-related projects are associated with barriers such as complexity, price or unknown ROI.

Big data does not convert data into actionable information. Big data does not create value. But data science does, and it does not have to be complex or expensive, or even big.

Data science is not about data scientists, it’s about business acumen and passion

Data science is a three-legged-stool that combines business acumen, data wrangling and analytics to create extreme value. Focusing on the hard science skills such as statistical methods is a common mistake when actually, developing the knowledge about a particular business and wrangling the relevant data are usually the most important skills to bring to the table.

It’s one thing to know how to play with numbers, but it’s more important to understand what insights these numbers reveal on the business, and what actions to take based on these insights. Experience and business knowledge plays a role, as well as curiosity and passion. Sometimes the best results come from unlikely people just because of their desire and persistence.

Driving insights to actions with analytics

Turning data in to information, and insights into action is a multi-step process:

  1. Ask the right question by deciding which specific business problem to tackle. Make sure you are focused and spend every minute on high value business problems.
  2. Sourcing and accessing the appropriate data sources. Organic big data are cheap and non-intrusive, but often not representative to the business problem. Seeking out the best, most representative data is a key initial task.
  3. Clean and transform these data sources. Massaging available data to address the business problem is time consuming but crucial. The best analytics methods can’t make up for low quality data.
  4. Clearly define the response variables and explanatory features that inform the business problem. Crisp definition of variables and features is central to deriving actionable insights.
  5. Prepare visualisations and dashboards highlighting the key response variables and features so that everyone can quickly derive insights on the business issues.
  6. Create predictive models and rules that encapsulate the insights for ongoing analysis and action.
  7. Publish the dashboards, models and rules for ongoing use across the business.
  8. Refresh with source data as often as practical for the business problem at hand. Some data don’t change very fast and/or can’t be actioned right away. In other cases, we can drive action within a day or even in the moment.
  9. Define alerts and notifications so that appropriate stakeholders are informed when there are opportunities or threats on the business.

Each of these steps is rooted in common sense, and keeping intensely focused on the business problem at hand. With this kind of focus, data science drives significant value across industries and functional areas. For business problems where data are in motion, we can get all the way through these nine steps and drive extreme value to the business. Some examples include:

  • Omnichannel customer engagement and appropriate offers while customers are transacting on a website
  • Order fulfilment and inventory restocking based on demand, inventory and purchases
  • Proactive machine management and maintenance to prevent/address failures and enhance machine productivity
  • Oversight of transactions for fraud and compliance
  • Logistics and transportation optimisation and routing
  • Price optimisation based on updating demand and conditions.

Michael O’Connell is chief analytics officer, TIBCO.