5 hurdles slowing down your uptake of data analytics

In speaking with senior information technology and data science professionals across the Financial Services, FMCG and Health Care sectors, I’ve come to understand there are a number of common hurdles that keep appearing on the proverbial track to adopting and making the most of big-data analytics for actionable insights.

These hurdles haven’t completely blocked the track or stopped the uptake of data analytics, but they have certainly slowed and or delayed the uptake, particularly amongst mid-cap companies. Several professionals that I spoke to had also come a cropper after clipping a hurdle and falling on their wrist, as Australian Olympic champion Sally Pearson found in Rome a few years ago.

Below, I outline the 5 main hurdles I’ve come across to date. Do these insights ring true to you?

Please note that this insights paper is also meant to be a discussion starter about big-data analytics within the industry. I am interested to know whether you have encountered these same hurdles and how are you addressing them. I am equally open to learning from your experience if your insights and approaches have been different.

#1. PLATFORM. How to manage all this data for easy access.

With the proliferation of new data and new data types arriving in unseen-before quantities from multiple sources, both internal and external, a new type of data platform is needed. This is not an easy feat. As if the data warehouse for structured data was not challenging enough, the concept of a ‘data lake’ is now in vogue. This raises new types of question for the CIO, “Do I really need a data lake when I already have a data warehouse?” and “If I’m going to move to a data lake, what should I do with my existing data warehouse? Decommission it or integrate it?” That’s just one aspect.

There are other questions on governing data, acquiring it, organising it and delivering it for discovery, modelling and analytics. The challenge is building a new style of platform underpinned by technologies, from the plethora on offer, that is effective for your needs and budget.

Then there’s the cloud – easy to access but complex to architect. Big-data analytics is difficult unless the platform is underpinned by cloud computing technologies (private, public or hybrid). The reason cloud computing is fundamental for any real analytics at scale is that it provides cost-effective abundant compute power to process big-data, in addition to the services for advanced analytics, machine learning and AI.

#2 TRUST. Low confidence in the data.

Relatively speaking, data science is the easy part but getting the data, internal and external, in the right place, at the right time, and then getting it ready for analysis, is the hard part. The challenge of getting secure access to trusted and quality-assured data exacerbates the stresses of making timely decisions on actionable insights.

Many factors contribute to this situation, reducing the level of confidence in data-driven decision-making. For example, there’s often no common understanding of what data exists, where it can be found or how it can be provisioned. Others have come to realise that their company’s data stewardship, data governance and quality management is ineffective.

Data is often deeply embedded within systems, needs to be enriched from external sources, and getting acess to it requires effort and layers of bureaucracy. By the time it makes it to the users, it is often late, stale, not as needed, or not well understood. This is also because data lineage or provenance is sometimes not clear, and the users are unable to trace it to its source or how business rules transformed the data set along the way.

What this means is that the platform architecture for data analytics needs to account for data sharing, trust, secure access, and delivery at different speeds.

#3. TOOLS. Explosion of new analytics tools makes choosing the right ones difficult.

The explosion of data analytics tools is making choosing the right one difficult. There is no dearth of tools, but it is amply clear that the use of older and best-known tools such as MS Excel quickly reach their limits when it comes to complex queries and analysis on large data sets. Take data visualisation tools for example. There’s a suite of them to choose from that ostensibly offer similar features and capabilities. Look closer and you’ll see overlapping capabilities with other aspects of the data platform. Which one’s right for you? If not properly assessed there’s the likelihood of regret spend on duplicate capabilities or choosing one that falls short of our needs.

#4. INERTIA. Organisational factors get in the way.

Often executive leadership is quietly sceptical about the benefits of analytics. I’ve heard it said by executives that “We have other burning issues to tackle right now and we’ll get to data analytics later.” They don’t see it as integral to business strategy and therefore, sponsorship is lacking resulting in suffocated analytics in preference to other enterprise initiatives.

Another belief is that what worked in the past will continue to work in the future. After all, the leadership team has steered the company through challenging times, and they are well prepared for new ones. But the competitive landscape has changed. Industry boundaries are amorphous making it easier for new digitally-savvy entrants to pull customers away. The belief that past experiences is equally transferable to the new digital economy may not hold. An analytics mindset must permeate through the enterprise to allow assumptions to be tested against facts, evidence and data.

A holistic strategy for data and analytics is missing that provides a clear vision, proactive business engagement, road maps, platform architecture, outcomes and performance metrics. Analytics is fragmented with islands of data smattered across the enterprise. Or, it is centralised to a single function or group, causing it to become a bottleneck because they are unable to scale beyond the initial stages. Instead, think of analytics as a fabric woven with collaboration. IT builds the analytics platform to deliver data, the business owns the data, business users consume the tools and data, and a data governance group ensures that data quality is assured – all working in concert.

Enterprise initiatives that cast too wide a net limit their ability to show benefits. Outcomes and business opportunities are not aligned. There is not the linear relationship between more data and action and so, thinking that more data the better is not necessarily correct. The value of more data may not be worth the effort and cost of processing. It is more effective to start analytics with a specific business question, easily accessible data, and learn from the experience while improving it iteratively.

#5. TALENT. The war for talented data scientists, analysts and engineers is on!

All companies are becoming increasingly reliant on data scientists, but good data scientists are in short supply! Companies and organisations with traditional workforce cultures need to emerge to foster work practices and cultures that attract and retain the new generation of data scientists for the long-term. A flexible workforce with an ecosystem of partners works well but gets overshadowed as a buyer-vendor relationship. Similarly, companies are finding their good data engineering resources are hard to find and retain, and need a similar approach, as with data scientists.

How Irada’s dialectic approach can help in your digital business transformation

Analytics-driven decision making is a critical component of the digital transformation puzzle. But making your track clear and overcoming the hurdles on the way through is never easy.

This is why Irada’s unique ‘dialectic’ approach to transformation will help you win. We understand that what works in one situation may not work in another. That is why any game plan we work on with you will not be a cookie-cutter approach. We take into account all the factors influencing the uptake of big-data and analytics, such as support and understanding from C-level executive leadership, alignment with business aspirations, organisational culture, and buy-in from stakeholders in finance, marketing, sales, customer service, and supply chain.

We do ask a lot of questions but we know which questions to ask to whom.

As discussions around big-data and analytics become mainstream and move from conversations among technicians in the server room to strategic conversations among Senior Executives in the C-Suite, it helps to have someone on your side, with close to a decade’s transformation experience, who has seen the pitfalls and mistakes made in other companies like yours.