If there is one thing we can be sure of when we look at the future of the IT industry, it’s that data is going to get bigger.
Already, the amount of information the average person generates is growing exponentially-a recent estimate from Cisco put the amount of data generated by internet traffic per hour to be enough to fill around seven million DVDs. And estimates of the size of the net by 2015 put the total amount of data online at 966 exabytes, that is 966,000,000,000 gigabytes.
As more and more of our personal and professional lives are lived online, facilitated by net-connected devices, and more and more companies embrace cloud-type services and technologies, the amount of data that the average company is sitting on has grown rapidly, and this growth is showing no signs of slowing down.
This explosion of data, however, is only one aspect. The vast amounts of rapidly generated data can contain incidental material that is of no great significance, but it may also contain vital information from which patterns can be determined that can help shape everything from daily procedures to long term business strategies. This phenomenon is known as Big Data-and its advent is presenting Irish companies with a challenge. How should it be dealt with? How do you separate the data that should be stored from that which can be ditched? Should any be ditched at all? Is there business value in hanging on to large amounts of data, and if so, what is it and how can it be realised?
Two roads
"There are two camps on how to deal with this. One camp wants to know how we can get that data into our data warehouse to understand it and use it, and the other camp thinks that realistically, there’s too much of it and that’s never going to happen," said Paul Pierotti, client lead with Accenture Analytics.
"Although they will totally acknowledge that data processing costs will reduce, a lot of companies have come to the conclusion that there’s a lot of data they should just use and lose. I think we need to be pragmatic-there are a variety of sectors where things are changing very rapidly with the digital world."
"The idea of putting data into a warehouse and waiting four months to see the results, most companies just aren’t in that kind of world anymore. My take is that there is a pile of structured and unstructured data inside your organisation that you can use. The more you can use that quickly to improve customer outcomes the better."
According to Tim Arkless, system architect with IBM, big data is beginning to reach a tipping point in the ICT world because of the degree to which data is being pulled together.
"Obviously companies are now generating more data in more ways, but it’s when that data is cross-referenced and related to each other that it becomes big. Everything is connected and integrated-there’s a lot more federation, so it’s an evolving thing," he said.
"How we handle big data is important today and it’s going to become even more important tomorrow as these developments show no signs of slowing up. How business handles this is very important moving forward."
The Knock
IBM suggests that this isn’t an issue that will only concern huge companies sitting on massive data warehouses-in fact, Arkless suggests that the kind of business an organisation is involved in is more or less irrelevant. Big data is knocking on everyone’s front door.
"The kind of business you’re in isn’t that important-this is something companies of any size can use to their advantage. The question is ‘how big is big data?’ Well, the definition is different depending on what an organisation does. It’s relative to what you’re already working with," he said.
The thinking is that if you’re a small company, then the growth that you’re seeing in the amount of data you work with is, for all intents and purposes, Big Data for you. How you handle that growth, and what value you derive from the avalanche of proportionate data growth coming your way is your Big Data challenge. A global corporation is going to have its own Big Data challenge, and while they will be related to each other, the solutions may look quite different.
Don’t have, don’t trust
"We did some research recently that said that one in three business leaders make decisions based on information they either don’t trust or don’t have," said Arkless.
"We also found that one in two business leaders say they don’t have access to the information they feel they need to do their job properly. So the information we have is that people don’t feel empowered by the information they have, and think there is more data out there that would allow them to make better decisions."
So why is that we’re collectively sitting on more data than ever before, yet it seems that many business leaders perceive themselves to be data-poor? The answer lies in the way companies store and structure their data, and the way they lay the foundations for good data analytics.
"Analytics is all about the better use of information to drive business performance. That happens in three categories," said Pierotti of Accenture.
"The first is investment decisions, where do you put your money and resources? The second is operational performance, if you have a chain of shops for example, how can you know which are the better performing outlets? What’s driving that performance and how can you take that information and roll it out across the wider organisation? The last one is about customer outcomes," he said.
Pierotti believes that too many organisations in Ireland deliver vanilla services to their customers, rolling out the same services to each customer, regardless of their individual needs.
"Really, the key here is embedding analytics in the decision making process and in the actual customer experience, so that you and I-two different customers-have fundamentally different experiences based on the information you know about each of us," he said.
Fraud analysis
A good example of specifically how this can happen can be seen when analysis is applied to data security and fraud prevention.
"We’re doing a number of pilots around fraud reduction and fraud management at the moment, where we are using a strong predictive approach to deciding who is a fraud risk and who isn’t. When you do this correctly, it means you can change the experience for the customer," said Pierotti.
"For example, take something like online banking. There’s analysis that shows that if your mobile phone is in the same location as you when you make a purchase, the chances of that being a fraudulent transaction are massively reduced."
"This applies just as much with online purchases and also to people who travel and shop abroad. People who have used credit cards in unusual or out of the way places will be familiar with the phone call from the credit card company asking you to confirm your identity."
But if location-based data is used, phone calls such as that could be a thing of the past. People deemed to present a lower risk or potential fraud could be presented with fewer questions when they log into an online service or present themselves at a cash register.
"It’s important that people opt in to that kind of service to respect their privacy. But for fraud services, I think most people would happily opt in as they will see the value to them. Consumers need to be educated and consulted before that kind of service is rolled out," said Pierotti.
Ways and means
There are also ways for smaller companies to benefit from this kind of approach to working with Big Data, according to Sean Moynihan of Information Technology Services.
"Working effectively with big data stores requires the power of techie buzz-words like Sql Server Analysis Services (SSAS), Online Analysis Processing (OLAP), OLAP Cubes and so on. But at the smaller end of the scale, businesses just want to query their data," he said.
"For example, they want to query their accounts database to analyse margins, market performance and so on, in ways that their standard package can’t do. In the past, the main way of doing that was to use products like Crystal Reports, but the most recent release of SQL Server Reporting Services has got a lot of very impressive new features."
"We’re currently using SQL Server reports with a fairly typical mid-size business in Clonshaugh to analyse margins, market performance, year-on-year comparisons and so on from their Sage 200 database. It does take some expertise that the small company will naturally not have in-house , but in effect SQL Server reports comes for free. So you don’t need to spend huge amounts to benefit from data mining."
Data value
Just what is the value to be derived from spending the money that will be necessary to get enlarged quantities of data working for the average company? Why is big data better than small? According to Damian Moloney, business intelligence manager with Avnet Client Solution, the answer depends on what you use it for.
"From a statistical perspective it enables you to work on a larger sample, so you could potentially end up with a more accurate trend analysis. If your data becomes big because you increase the frequency, for example, instead of taking a meter reading each month, you take it from a smart meter every ten minutes, then potentially that allows you to make decisions more quickly," he said.
"So instead of waiting till tomorrow to analyse today’s information, you can do it at ten minute intervals. That could have real business value, but it would generate a lot more data that would need to be worked with."
"Some companies will have the additional capacity already in-house to cater for the increase in data that something like this would create. Others might have to make significant investment to allow them to handle it. We recommend that a company should have a mature business intelligence capability before they strive to analyse big data. In other words, make sure you can walk before you run."
Toolset
There’s no doubt that the ability to work with enlarged quantities of data, significantly more than at the moment, will be a hallmark of successful companies in the near to medium future. Recent developments such as HP’s decision to sell its PC business, with its ultra-thin profit margins, and buy into UK software company Autonomy for a reputed EUR*11 billion, are symptomatic of how some of the big players are positioning themselves.
Autonomy is an analytics company that will allow HP to focus on the more profitable business of helping other companies deal with Big Data.
"Deriving value is the key issue. Any observer of the industry will be able to see a pattern emerging as some of the biggest players in the market position themselves in order to be at the forefront of the race to deliver big data style tools," said Chris Coughlan of HP.
"The future lies in helping organisations derive the most value possible from the data explosion that’s happening all around us. In my view, the future lies in cloud and in that area."
According to Coughlan, advances in cloud technology are creating opportunities for companies of all sizes to benefit from collaboration, generating more and more data in the process.
Cloud-ifying
"Smaller companies that want to position themselves in this area could do a lot worse than fully embrace the cloud, cloud-ifying their business. There’s a whole pile of opportunities there for new services, systems and products to be developed to help organisations of all sizes grapple with big data, and it would be a great area to be positioning yourself for now. It’s not going away, and it’s only going to get bigger," he said.
"These down-stream opportunities for services will be needed to service the ecosystem that will be created as an effect of big data, in areas like security, building services, in supporting the many different financial models that will undoubtedly appear. We should incentivise SMEs here and elsewhere to come here, to corner this area, and there are opportunities there for Irish companies."
Subscribers 0
Fans 0
Followers 0
Followers