Betting the business on big data

Betting the business on big data
Nik Healy, solutions director, Convergent

Big data: it’s an intimidating subject, doubly so when it comes with artificial intelligence in tow, but does it have to be so confusing? If big data means anything it means adding to the sum of knowledge in an organisation, and, Jason Walsh asks, isn’t that something every business would want to do?

Data is the future of business; we’ve all heard it a million times by now, but the hype cycle keeps on spinning.

With each of us now leaving an unending trail of data in our wake, it was only a matter of time before someone worked out how to make money from it. Throw unlimited processing power and storage into the mix — thank you very much, cloud computing — add a sprinkle of artificial intelligence (AI) and, hey presto, we’re all living in the future.

If only anyone other than largest corporations on earth knew how to make any sense of the ocean of data in which we now swim.

In fact, any business can. More about KPIs than rampaging AIs, step back from the daydreams about the Star Trek transporter or the cybernetic singularity and big data is really about improving mundane, but important, business processes.

“It’s a question of operational intelligence, really,” said Phil Ryan, customer relationship management solutions manager at Egro Group.

Ryan said what while big data will prove transformative for business it will do so only if it is linked to actual business processes.

In other words, while data might be kept in the cloud, feet should remain planted on terra firma.

“You need to think about it mostly in terms of strategic aims,” he said.

Different industries have so far taken different approaches to data, he said, often reflecting long-standing cultural issues, not to mention compliance and regulation.

“There really are two ends of a spectrum,” he said.

“On the one hand, retail is [already] using data for revenue and profit generation, whereas, for example, the banking industry — outside of the fintechs — would be quite conservative.

“They work mostly within their own private spaces, innovation centres, and so on.”

Picture: Maura Hickey. Noel Brilly, Cloud & Consumption Services Manager, Ireland.

Noel Brilly, GreenLake consumption services manager and business developer at HPE Ireland, said that big data can be understood simply enough: it’s about doing more with more.

“The term ‘big data’ comes from the relatively recent ability for organisations to store and process colossal amounts of data, which previously would have to have been discarded because of the cost and effort of keeping and processing it,” he said.

The business value does not itself come from big data per se, he added, but from using artificial intelligence methods on the enormous datasets that big data enables.

“The larger the data set you can process theoretically gives you more ability to identify more patterns and derive more insight. Artificial intelligence is essentially the simulation of human intelligence processes by a machine: systems perform tasks such as searching or identifying patterns in data to give a business insight into their data where previously they would not have been able to do so.”

Simple enough in theory: big data enables machines to do some of the ‘thinking’ previously vouchsafed for humans, preferably repetitive data that humans struggle with, and on a much larger scale.

The end result, though, should not just be the streamlining of business processes, but using the insights gathered for the development of new business, whether repeat or from ongoing customers.

“The change to business comes where it is now possible to analyse a huge data set for patterns or insight where as previously it was not, and therefore identify insight which was previously unavailable. Insight normally hopefully translates to competitive advantage,” said Brilly.

Mark McArdle, group sales director at Intact Software, said that the goal was to make useable information available where and when it matters.

“The overall objective of analytics is to enhance decision making at every level, with instant access to accurate, real time information,” he said.

Speeding up the process of obtaining useful information — in effect, extracting it from masses of data — means that business can be driven in a rational manner, focused on how customers are behaving.

“We put the power to extract and interrogate data into the hands of our customers and their users which brings significant benefits to their business. Our customers find they can respond quicker to day-to-day business requirements; becoming more agile and responsive as a result. They can spot trends quicker. They can identify gaps quicker. And they can see who hasn’t been buying from them lately.

“With centralised information accessible to all via a dashboard or report, customers are reducing the amount of time they spend on monthly and management reporting. They’re identifying wasted resources and reducing costs. But [it] goes beyond financial data, with customers looking at the time of day, source of enquiry, pick rates, slippage, missed opportunities and more,” he said.

Edge-to-edge information

Of course, in an age dominated by the likes of Google and social media firms like Facebook, big data has become something of a shibboleth. Nonetheless, behind the usual games of buzzword bingo, there is something real at work —and the involvement of major technology companies like HPE is a testament to that.

And HPE is not alone. Dell EMC is working with clients in fields as diverse as healthcare and finance to put AI to work on mountains of data.

“We’re extremely interested in it and extremely active in it. You probably expect me to say that but we really are,” said Marc O’Regan, chief technology officer at Dell EMC Ireland.

Dell EMC isn’t playing around, said O’Regan, and the company does, as the expression goes, ‘eat its own dogfood’.

“We play heavily, we practice AI on ourselves and write it into our delivery models,” he said.

For external customers, though, the company uses its expertise in data to develop AI platforms.

“We’ve done work around genome sequencing in life sciences, for example,” he said.

Hugh Nolan, Lead Solution Architect, Singlepoint

“It’s about developing an engineered system that brings in all the compute, network and storage, but also the binary components; that is libraries and compilers. What you’re doing is freeing up researchers to drop data into the analytic engines and get up and running.”

What Dell EMC’s customers want are powerful, flexible systems; often ones that work in conjunction with public cloud.

“What we might do [for a customer] is develop an analytics engine and integrate it into the likes of Azure. What we’re interested in doing is understanding how we solve complex problems and [in doing so] create business opportunities. Building an analytic engine is one way of doing that and bringing it into the Microsoft or Google cloud brings a lot of value,” said O’Regan.

At the other end of the spectrum from the vast public cloud providers is on-site processing — or even off-site; something that goes beyond traditional industries and right into daily life.

“We’re also deep into the IoT [internet of things] and bring AI, neural networks, to the edge devices. That could be in a production environment or an EMR [electronic medical record] or in a hospital, pulling it across a low bandwidth network and doing analytics right at the edge: at the patient’s bed, in the operating theatre, or on the production line,” he added.

Getting real

The chatter around big data brings challenges itself, not least of which is the dynamite combination of the hype train and jargon. Half-digested ideas are not going to help businesses make serious use of the information they record.

The answer is a strict focus on the business case.

“Our ideal client is one who has a good idea of what they’re getting into. The worst is when someone has read up on big data and thinks ‘we need to get into that’,” said Hugh Nolan, lead solution architect at Singlepoint.

Singlepoint approaches big data by taking a consultative approach with its clients, he added.

“The most important thing to understand about any customer coming in the door is to understand their business: to look at the existing KPIs [key performance indicators] and look at how they are reflecting the real world — or not — and how they might be improved.”

For Nolan, it is essential to make sense of what the goal is and, from there, how data can get you there. One aspect of this is looking again at KPIs to see if they are really performing their tasks.

“Then you can look at what kind of data you have or can gather,” he said.

“The KPI having aged would be a major one [issue]. Another would be [when] a business puts in an IT system to deal with a specific issue but, over time, operators have come up with workarounds. Humans adapt. The big thing about big data and machine learning is that it’s about getting machines to adapt, including to how humans have adapted.”

Phil Ryan, of Ergo group, makes a similar point, but also said that newer, better KPIs develop as a result of using big data.

“When you get involved in analytics, there is no confusion about what KPIs mean. The traditional way is often too slow and people want the information in real time or near real time.

“What tends to happens, organically, after a couple of months is that the KPIs can take shape following the data that is being collected.”

The other side of the story, however, is that radical changes to expectations can leave staff feeling abused.

“There is [now] a paradox in management,” said Ryan.

Mark McArdle group sales director at Intact Software

“[It is essential that] education goes along with the technology. You could be using to the KPIs that could be wrong, something that might become clearer when you get the real data in. That’s an educational process that they have to go through,” he added.

For Marc O’Regan, of Dell EMC Ireland, the key thing to remember is that the goal is, ultimately, to create a new opportunity to solve a complex business problem.

“First we define big data and big data is all data in my view: big data, small data, unstructured data, structured data, dark data… What’s interesting is how we actually monetise the data to give somebody a competitive edge or save a life.”

This will, of course, mean different things in different industries, but analytics is at the core every time.

“If you look at manufacturing, we have a number of things we can bring to it: looking to asset management, deficiencies on the line, and so on, build data into the line and use it analytically.

“We can do the same in hospitals, we can do it in finance. It’s not about the ‘wow’ factor, it’s about tapping into [the] every day. What Mastercard doesn’t want is someone commenting in how good the AI is. They just want the customer to transact easily and move on in their life,” he said.

Small business, big data

Big data has, like most new technologies, been subject to rather breathless reporting: it will cure cancer but obliterate our privacy in the process. Both ends of this propaganda spectrum have something in common, though: scale.

Our love and hatred of the large are not so much juxtaposed as intertwined, so while we cheer the curing of diseases by researchers at giant institutes we also fear the internet behemoth will eat the private sphere.

Both are points worth considering, but where does it leave big data when it comes to the small guy?

Nik Healy, solutions director at Convergent, said his goal is to bring these benefits to Irish indigenous businesses, which are often smaller.

“Our customer base would be small- to medium-sized forms, of all sorts from retail, food producers, to a range of people,” he said.

This is precisely where Irish business is: 2014 numbers from the Central Statistics Office indicate that small and medium enterprises accounted for 99.8 per cent of total number of enterprises, and nearly 69 per cent of all employment in the country.

And just because they are smaller businesses doesn’t mean that they are small beans. These SMEs account for over 56 per cent of Irish business turnover.

“They’re not enterprise, but they’re progressive and successful. A lot of them would be well-known brands,” said Healy.

Convergent works with them on strategy, he added, and the market is a large one, not only because of preponderance of SMEs in the Irish economy, but also because these businesses, unsurprisingly, need help to do the things that multinationals can get done with the wave of a virtual chequebook.

“Most of these organisations, from what I see, are still years behind what big data is promising,” said Healy.

“For an SME it’s not big data, but it is data, and they want to be able to make decisions based on it. The focus has all been on the enterprise level — and it has left a gap.”

So where does all of this data that has the power to magically transform business come from in the first place? The information is there; it just needs to be made understandable.

“If the data is already being collected somewhere you can evaluate it and see if it can be brought together in a meaningful way. If not, it’s a case of ‘how do we change our systems to collect it’,” said Hugh Nolan of Singlepoint.

Here, automation is the key.

“One problem is getting the data in the first place: if you’re running things manually, on paper, there’s no data being generated, except maybe for some data that is manually entered into an accounts package or ERP [enterprise resource planning system] at the end of the day,” said Convergent’s Healy.

He also said that smaller businesses often struggle with entrenched processes.

“This is an issue. There’s a lot of preparatory work that needs to be done so that a particular business process or function is able to produce data that can be gathered and analysed so that it can be turned into a report or a future analysis.”

After all, they are getting on with the business of business, and while disruption might have positive connotations in Silicon Valley, no one can afford to disrupt their actual business with an ill-conceived plan to implement a new technology.

“There’s a lot of unstructured data out there: documents all over the place.

“Most company information, to this day, is held within documents: HR contracts, client contracts, decisions. Most information that is designed to help with processes is trapped somewhere in a spreadsheet, a folder, or, god forbid, a paper document on some shelf somewhere,” he said.

The first selling point, though, is that information lag can be eliminated: something is working, or perhaps it isn’t? With digitised data the truth can be discovered immediately — and this is done with already-familiar tools.

“The numbers shouldn’t come in at the end of the month. If you digitise what you do and automate your processes, then it becomes live. After that, your BI [business intelligence] tools, say Qlik or PowerBI, can start to tell you things.”

Healy said that a decision has to be made, however, on where to draw the line, otherwise what should be a process of digital transformation will become a Sisyphean task of converting old information into new formats.

“We were recently approached by a company that has, in many ways been progressive, but they’re only now talking about digitising things like procurement, invoice approval. A line has to be drawn otherwise it becomes a legacy conversion project that can take months and years, or maybe even get nowhere,” he said.

Can technology itself be put to work here, though, making the decision on what is or isn’t relevant? Healy said yes.

“With AI becoming, I wouldn’t say a norm but becoming available, it is bringing with it a new level of possibilities to the table. It can make structured data out of unstructured data. It’s a level of automation, and the AI is provided by the big public cloud platforms. We have a tool called M-Files that uses them. It goes through the legacy repositories, meaning you can get data out without doing a complete migration.”

HPE’s Noel Brilly agreed.

“This [big data and AI] is now more feasible and cost-effective for smaller organisations through HPE and our partners’ technological advancements, as well as HPE GreenLake’s pay-as-you-go service,” he said.

Whether SME or enterprise level, however, Brilly said that the model for IT in enterprise has changed, and big data is the latest development resulting from increased connectivity.

“What we hear from our customers, and from industry analysts, is that the cloud experience has shaped what business leaders expect from IT. The pay-as-you-go, rapid scalability, operated-for-you model will become common. By 2020, consumption-based procurement in datacentres will account for as much as 40 per cent of enterprises’ IT infrastructure spending.

“It is much more flexible and simple to work with, and it’s expected that the industry will adopt a consumption model across the board, including for on-premises IT,” he said.

In the end this will mean businesses themselves will change, said Marc O’Regan of Dell EMC Ireland. If data really is virtual ocean, it’s sink or swim time. “Doing things using the big data is what the paradigm is about. Everybody has to look at their organisations. The impact will be [felt] across industry and if you’re not looking, your competitors certainly are.”

Big data, simple idea

Nik Healy, of Convergent, said that some of the discourse around big data is distinctly unhelpful, as it clouds what is, in theory at least, a fairly straightforward concept.

“I feel anyway, that there is this massive step up to big data. People look at it and say: ‘This is Star Trek stuff!’ It’s just not realistic for them,” he added.

Thought about in more grounded terms, however, and big data need not be daunting.

“On the other hand, businesses produce massive amounts of data and they still have the challenges of gathering data and being able to analyse it.”

Hugh Nolan, of Singlepoint, also said that problems with terminology can create confusion, specifically the buzz around artificial intelligence.

What is AI anyway? Extracting new KPIs from sales records or contracts is important, but does it qualify as intelligence?

“I think about AI as getting a machine to interpret the real world. On the other side you have ML, or machine learning, which is getting a machine to iterate over datasets.”

This, he said, is where the strategic value can be produced.

“A huge amount of time, effort and investment is being put into algorithms that can pull out the salient legal points in a contract. The actual interpretation of contract law is going to become a lot simpler as there will be fewer grey areas. That already is a huge change, the work that is going on in that field.”

The data dawn

Venerable tech writers Mary Jo Foley and Paul Thurrott think that analytics will transform business — and society — in hitherto unknown ways.

Visiting Dublin for the Cloud Camp conference on October 17, they told Connected that its effect will be transformative.

“I think it changes everything, everywhere,” said Thurrott.

“A big part of it is that these capabilities are no longer only available to a few companies. Public cloud makes it something that any business can use.”

Foley said that she agreed, and that real-world examples were beginning to make themselves obvious.

“I met an insurance company guy who was raving about AI,” said Foley. “It was finally affordable and now something that had been pie in the sky is a real prospect.”

Foley and Thurrott, who co-host the Windows Weekly podcast, have a combined experience of 55 years in reporting on technology, and both think that the shift has to be understood in material terms.

“Don’t just think about dripless cars; Microsoft thinks about the ‘connected car’: you can get data from your office into the car, you can get predictive maintenance. There are a lot of real-world applications for this,” said Foley.

“Of course, there’s the whole ethics of AI: how much data do you want the AI to have? There are [also] questions about jobs. There are downsides, but more and more tech companies are being proactive about that.”

On the question of the job destroying potential of AI, Thurrott is sanguine, and said that our ability to predict the future is haphazard at best.

“I imagined, when I was a child, [that] we’d have flying cars, but we don’t, and settlements on the moon,” he said.

He has a point: visions of the future, even those supposedly grounded in reality such as the gritty cyberpunk novels of the 1980s, may have imagined something sort of like the internet, but they missed out on the mobile phone.

As for tech jobs being de-professionalised, Thurrott said that we go through the same hype and fear responses every time a new technology appears on the scene.

“The shift to cloud computing is, in a different space, the same. The reluctance to see cloud in IT departments is gone. People maybe got complacent in some places and felt they were losing something to the cloud.

“As this has evolved, this [sentiment] has changed.”

Related Articles

More from The Business Post