There’s a long-standing myth that we only use 10% of our brains. Luckily it’s not true – otherwise who knows how far back in the dark ages we’d still be languishing.
But here’s what’s surprising. Even though we could almost compare data analytics to the ‘think power’ of an organisation, on average most companies analyse a very small percentage of their data. In fact, the IDC predicts that even by 2025, just 15% of useful data created in the global datasphere will actually be tagged and analysed.
There are two key reasons why this statistic is so low.
Swimming in data but struggling for insights
The first is the sheer volume of data out there. Many companies are already managing exabytes of data, with the IDC estimating that this will continue to grow each year by around 30%. The result? People are swimming in data but struggling for insights.
Equally problematic are siloes of which there are three different kinds.
It begins with data and architectural siloes – this means that when an application is created, it’s typically placed within a siloed architecture. The application creates data, but that data is effectively stranded on the application. How can you then correlate it with data from other sources?
This challenge is exacerbated by organisational siloes. Most companies have changed the way they do IT, shifting from operations to more functional groups. But the problem then becomes actually achieving objectives which are greater than those individual groups.
And even worse are mind-set siloes and behaviours, which are the number one barrier to digitisation. In fact, as difficult as it might be to swallow, achieving your digital transformation objectives might well require you to have a change of leadership within your organisation.
From edge to outcome
So the real question is – how does your business start leveraging more of its data and delivering outcomes that matter? We believe it’s about looking at your core data sets and analytics, and correlating those with what’s happening at the edge.
While IoT is often only associated with the edge, it’s important that the data produced here isn’t looked at in isolation. In fact a major focus for Hitachi is to help businesses deliver edge to outcomes – in other words, regardless of where your data is created, you’ll be able to deliver the outcome you need to benefit your business.
Metadata is the new data
When it comes to adding new data streams that you want to analyse, having an architecture that focuses on and leverages metadata is critical. Basically metadata is the new data, and as architecture changes, metadata will become the layer on which we do all our analytics.
If you look at a call centre, for example, the audio data of an interaction between a customer service employee and customer is not important. But you could potentially translate that data into text, converting it to metadata.
This is the layer which becomes really important. Let’s say you have a particular customer service rep and there is a complaint about them. You would be able to search your metadata to determine which calls that rep has been on, and then pull that data to analyse it.
Achieving the outcomes you want might mean you need to look at your business differently.
Are the right people in place to embrace new technology? For example, where will blockchain be in five years and what should you be doing to ensure it won’t be disruptive to your business?
At the end of the day, people and data will prove key to everything that you do. You need to make sure that your business is leveraging its data to differentiate its service to customers.
Deliver outcomes that matter
Let’s look at the example of a fleet manager who is tasked with identifying problems across their fleet before they happen.
If that manager has sensors across each vehicle, which are all collecting data, then using several days’ worth of that data, they will be able to spot that the rate of change in a particular tyre’s pressure has been very abrupt.
By creating a dashboard which contains key information about the health of the fleet, not only can the fleet manager pinpoint the location of every vehicle in their fleet, but they can also see the particulars of each vehicle, enabling them to attend quickly to a vehicle which needs their attention.
Running a predictive maintenance model will also tell the manager whether there is a high probability of failure over the next few days. This allows them to make an immediate appointment, and with a model like Pentaho’s, it’s easy to find the nearest service centre with immediate availability, as well as to send on the necessary parts information.
At the end of the day it’s about accessing the right insight at the right time to deliver outcomes that matter.
By JP Smith, Pre Sales Director, Sub-Saharan Africa at Hitachi Vantara