Putting a value to big data

Most businesses have woken up to the fact that big data is an area in which they need to invest in to realise its rich potential. Finding the investment for this can be hard though, as most organisations need to make a compelling ROI case before spending on new tools for the business.

This is exacerbated by only 10 per cent of firms having assigned a monetary value to the data they hold, according to recent research by The Economist Intelligence Unit (EIU). So what can organisations do to monetise their big data, what do they need to make big data deliver big ROI? 

Big data and ROI
Big data is now an established business tool, potentially providing organisations with game-changing insight into customers, competitors and the wider industry they operate in. It helps executives make smarter and better-informed decisions.

But it has been said that the benefits big data offers are less tangible than they need to be. Organisations are investing significant money in big data and they want to see proper return on that investment. And this means getting hold of the right skills, tools and mindset to do so.

First thing to consider is exactly how organisations should be assessing ROI – what metrics and over how long? A majority of people would readily acknowledge the insight that can be derived from big data, but one would hope they are also aware that there is a time and cost associated with this. It requires some form of big data analytics engine, such as Hadoop. This in turn requires an infrastructure capable of processing large volumes of data at sufficient speed and it will also require some resource – a data scientist or data analytics consultant – to extract the required insight.

Minimising costs to improve ROI
This is where big data can run into issues. Data scientists are a sought-after breed and good ones do not come cheap. And how many organisations can afford a costly infrastructure investment, especially when such computing power won’t be required all the time, only when running specific big data queries?

Hadoop was designed to run in large physical data centres, on dedicated physical machines. When you run Hadoop in a virtual environment, performance is compromised greatly. Rather than purchasing their own dedicated Hadoop environment, businesses can now buy affordable bare-metal services – using only what they pay for and need. By removing the most significant investments, the time scale for ROI will be much shorter.

Business units not technology
Part of the sea change around big data is that it is increasingly a business unit investment, not technology. This is particularly true of marketing, where targeted offers based directly on a customer’s preferences are hugely successful. According to McKinsey, around 35 per cent of what people buy on Amazon come from product recommendations, which are generated from big data from other transactions.

So marketing departments are deploying big data in ever-growing numbers. The digital nature of much of modern marketing means that results are easy to track and a big data deployment for a specific project or campaign is both less costly and easier to monetise than a grand over-arching big data project across an entire organisation.

So as well as the core requirements – powerful analytics capability, a flexible and scalable infrastructure, strong data – organisations need to start thinking about big data for specific uses, rather than as an overall solution. Such deployments will be key to achieving ROI from big data.