Top challenges facing data-driven enterprises – and how to overcome them

As technology developments reveal new data analytics frontiers, CEOs should look towards current best practices to unlock valuable, actionable insights.

Five years ago, it would have been almost impossible for a telco to overlay heat maps with channel touch points to identify its optimal operating hours. Digital agencies could not analyse footfall to focus advertising spend in high-traffic areas, and health officials were unable to use geolocation data to track disease outbreaks.

It’s a different story today. As companies collect and store more data, and small sparks of innovation add momentum to growing capabilities, the data analytics landscape is changing quickly. Recent advancements include real-time data analytics, combining datasets to generate new use cases, and an improved ability to understand customer behaviour at a granular level.

The big data and business analytics market is expected to grow from US$130 billion by the end of 2016 to US$203 billion by 2020. In this environment, CEOs are challenged to extract maximum value from their data, while also using analytics to better understand – and engage with – customers across physical and digital touch points. At IGNITE, a client engagement forum organised by DataSpark, speakers shared best practices for how data-driven enterprises should navigate the new data analytics frontier.

Understand big data collection, analysis and storage options

Every organisation will have a different approach to collecting, analysing and storing data. Some choose to store data in server farms and analyse findings in-house. Others prefer cloud storage and outsourcing analytics work to an external provider.

Cloud storage is an increasingly popular choice. Big data puts pressure on networks, storage and servers, so outsourcing to the cloud is a convenient way to reduce IT headaches. Cloud storage also makes real-time analytics more accessible by eliminating the need for additional hardware. (Open-source real-time data processing tools are now widely available.)

In addition, cloud storage can be scaled as needed. Organisations only pay for what they use rather than investing in hardware that’s rarely exploited to its full potential.  However, some cloud storage providers can struggle to guarantee performance as data storage increases.

Evaluate data scientist teams with rigour

Data science requires expertise across disciplines such as statistics, machine learning and mathematics. As a result, it’s almost always a team sport. To extract meaningful insights, CEOs should be able to recognise the hallmarks of an effective data scientist team. Here’s how:

Seek data scientist teams with a broad range of skills

It’s rare for a single data scientist to be an expert in everything from structured data analysis to human behaviour and data visualisation. For in-depth, specific analyses, ensure you have a team with skills that are comprehensive and complementary.

Emphasise real-world experience

While qualifications and theoretical knowledge are essential, leading data scientists also have experience outside of academia. Unlike their academic counterparts, these scientists know how to create business value while working with tight budgets and time constraints.

Be realistic about budgets

In a 2014 Accenture survey, around 41 per cent of respondents cited lack of data science talent as a chief obstacle in their big-data strategies. Similar shortfalls are anticipated in South-East Asia. As a result, skilled data scientists are in huge demand – and often have starting salaries as high as US$200,000 and above.

Understand that data analytics is a continuous process

Compared to traditional research-based analysis, data analytics is a continuous process. Every discovery opens up opportunities for further innovation. Recent advancements, for example, have enabled DataSpark to go beyond analysing Singapore’s MRT usage to using predictive modelling to understand how new roads or rail networks may affect congestion.

CEOs, therefore, must adapt and improve data analytics practices to maintain a competitive edge. Building on current knowledge and tools is critical for ongoing success.

Make priorities clear

Data scientists are always working to enrich the granularity, accuracy and speed at which insights are available. However, prioritising one data analytics performance metric (i.e. speed) might affect another (i.e. granularity). Some types of data anonymisation can affect data accuracy, for example, while real-time insights may not be granular enough for their intended purpose. CEOs should identify their most urgent data analytics requirements, and work with trained data scientists to determine the best way forward.

Follow industry recommendations for keeping customer data private

While customers might reap the benefits of sharing location data, such as improved urban transport, or alternate route suggestions based on real-time traffic flows, most are not willing to exchange identifying information to do so. A 2015 DataSpark survey of subscribers in Singapore and Australia found that real time location data was something consumers were particularly concerned about – only 9 percent respondents in Singapore and 13 percent in Australia were willing to share this type of data.

data-driven enterprises

When it comes to big data, security and privacy are paramount. Companies should ensure data privacy measures comply with local laws, such as Singapore’s Personal Data Protection Act, and proactively de-identify data.

At a minimum, data should be:

Encrypted – to ensure that only intended users can read and understand information.Anonymised– so that individuals and businesses cannot be identified.Aggregated – grouped into small aggregates (also known as microaggregation) to prevent disclosure of individual information.

Noise addition, which is the process of adding random noise to data, can also decrease privacy loss and stop individual records from being identified.

Data analytics provides CEOs with countless opportunities to improve business outcomes. By understanding how to store and collect data, evaluate data scientist teams and adhere to privacy recommendations, CEOs are positioned to unlock valuable insights that deliver maximum returns in a fast-changing environment.

For more on how you can become data-driven in your organisation, get in touch with DataSpark today.