Big Data Analytics Strategy Tips for IT Leaders
If you ask company leaders about data, they’ll probably say they have too much of it.
In the age of big data and the Internet of Things, the value of data has become common knowledge.
For many CIOs, though, it’s a challenge to store all the data coming in, let alone analyze it.
With experts predicting a 4,300% increase in annual data production by 2020, it’s imperative to get on top of data storage.
Every aspect of business is affected by big data, from operations to finance to sales and marketing. Without a plan to handle big data, you’ll soon be drowning in it, unable to come up for air.
If you’re facing this type of situation, you need to create a big data analytics strategy. Pronto. Here are 5 tips to get you started.
1. Define Your Big Data Goals
It may seem obvious, but you want to start by really thinking about the primary concern of your business and what goal you want to achieve through big data analytics.
For instance, are operational costs a big concern?
If so, your main data collection solution may be adding sensors to your machinery to help lower maintenance costs. Tesco, the European supermarket chain, cut its annual refrigeration cooling costs by 20% using analytics to ensure its chillers work at the right temperature.
Churn, or the number of customers who leave for competitors, is a major issue for cable and internet providers.
By analyzing everything from a customer’s personal information, the neighborhood they live in, the frequency of the services they use, and even how they prefer to pay their bill, big data analytics uncovers ways to boost retention.
To start creating your analytics strategy, talk to a small group of peers to develop your initial direction.
Hash out the major challenges big data can help solve. Narrow your choices down to the most important and vital ones. Ensure that your analytical goals align to your company’s overall strategy.
2. Promote a Data-Driven Mindset and Culture
Remember the book and movie, Moneyball?
This true story tells the tale of Billy Bean, manager of the Oakland Athletics baseball team. The A’s didn’t have the budget to compete with high-spending teams like the Yankees, so they created an entirely different approach.
By focusing on evidence-based sabermetrics – a statistic that measures in-game activity – the A’s constructed a roster full of bargain players whom the analytics showed to have game-winning potential.
The end result: The A’s made it into the playoffs and the conventional wisdom of baseball was turned upside down in favor of this new analytical approach.
It’s hard to believe, but the story of Moneyball happened in 2002, well before the big data boom.
Some may say that the A’s even hold responsibility in bringing analytics into the mainstream.
The important thing to take away from their story, though, is that they built an entirely new business culture.
They brought in players who saw the benefit of analysis to fundamentally change their business. For your analytic strategy to succeed, you need to inspire the same leadership to build your team and culture around a data-driven mindset.
3. Source Data Creatively
It’s been cited so many times it’s now a truism: 80% of relevant business data is mined in an unstructured form. Social media generates terabytes of unstructured data, mostly as text.
When you add unstructured data to structured data streams from sources like sensors, monitored processes, and local demographics information, you are broadening the scope of your analytical capabilities.
You open yourself to new ways of thinking beyond the borders of your structured data, and you gain the knowledge to get creative and innovative with your solutions.
The best ideas are rarely found in a bubble, and that’s where your structured data lies. Reach outside and start sourcing data creatively.
4. Embrace the Cloud
The traditional data warehouse is great for structured data, where information falls neatly into rows and columns. However, unstructured data is more suited to the cloud.
Unstructured data, such as streaming data, doesn’t fit the framework of traditional relational database management systems (RDBMS) very well.
RDBMS can store and retrieve unstructured data. However, the issue comes from the need to manipulate and extract the essence of this data, which is under the purview of the cloud.
As such, the cloud is emerging as an increasingly popular option for developing and testing big data analytics applications.
The adoption of open source Hadoop is growing fast, and the ability of non-proprietary and affordable hardware to perform analytics is making more and more sense for business.
This isn’t to say your physical data warehouse is going to be replaced. It will instead be augmented by affordable and scalable cloud services.
5. Work Hand-in-Hand with the Business
Coming back full circle to a point made in the first tip, Defining Your Business Goals, if your big data analytics strategy doesn’t align with your company’s strategy, it will never get off the ground.
The longstanding divide between the IT department and business leadership is closing.
The role of IT continues to grow and become intertwined with the success of the business.
For your big data analytics strategy to succeed, you need to assert your leadership and work with other business decision-makers as a team dedicated to the growth and health of the company.
Keep in mind that many traditional business leaders don’t understand the complexity of big data or the possibilities it holds.
You need to be prepped for some resistance and combat it with a good, sound argument.
In short, you need to educate your business peers on the value of big data and show other business leaders the potential ROI.