Learning from London’s leading Big Data practitioners

Perhaps one of the most recognisable symbols for data is the floppy disk. Interestingly, the data an individual creates in an average day would fill up more than 500 floppy disks and it’s only set to grow as we interact with more connected devices. In one sense, big data is a big problem – how can businesses store and use it easily? On the other hand, it represents big opportunities. The more you know about your customers, competitors and internal operations, the more your decisions can be informed and strategic.

Making data intelligible can be a daunting task. However, innovative technology is readily available that can sift through the data and make it palatable. And if this technology is easy to use and integrate into a business’ day-to-day operations, then it can significantly reduce friction and inefficiencies in the supply chain.

This is what we are passionate about. Through Tungsten Network Analytics, we aim to give firms greater insight into their supplier invoices which can lead to smarter spending decisions and reduce costs. Our analytics technology helps companies have a unified view of their spending and provides actionable insights. While we believe it can transform our customers’ businesses and impact bottom lines, we never rest on our laurels.

For this reason, we attended BigDataLDN recently, a two-day conference in London which focuses on how to build dynamic, data-driven enterprises. More than 3,000 delegates gathered to learn from leading Big Data practitioners, see some of the most ground-breaking technologies in action and find out how to deliver business value from data projects.

It was great to see that even in a highly technical audience, the end user wasn’t forgotten. In the sheer volume of data and complex tooling, it can be easy to forget the main goal – to make data more accessible to everyone.

Perhaps the most interesting talk from this perspective was on dashboard design. This is a topic that my colleagues and I often discuss – how to make it quicker for users to get from the headlines of their data to real actionable insights. The talk was fascinating because it looked at the use of eye tracking to test designs and considered two factors: how long does it take users to find important information and how long do they look at it for?

Most research on eye tracking has focused on web pages (such as the classic “Don’t Make Me Think” by Steve Krug) and has always suggested an f-shape to the way users skim a page. People tend to pay a lot of attention at the top, scan through the middle and then attention trails off as they come down the page. The same is true of how people skim most text-based information.

But how does this apply to more visual formats like dashboards? It seems we find it easiest to see big numbers at the top of the page. Certain placements took users three times as long to find the same information, and when they did find it, they didn’t look at it very long. We might only be talking about seconds but it adds up and tires the user out.

With Tungsten Network’s Analytics applications, we work on a simple methodology called DAR (Dashboard-Analysis-Reporting) that emphasises these insights. We start from having a Dashboard with a high level overview of the numbers in the dataset. This allows executives to see at a high level what their numbers are. We then move to Analysis which shows breakdowns of the information on the basis of different metrics, in quick and visual ways. Finally, we have the Reporting which shows the underlying granular level data. We want to build tools that allow people to move from high level insights to individual actions as quickly as possible.

Data is big and getting bigger but so is the way we think about viewing it. In our Analytics team, we are obsessed with data and finding ways to design dashboards that help our users find and understand it as quickly as possible. That way, they can take the insight and apply it to their businesses, reducing inefficiencies and friction throughout the business.