It has an air of hype or, worse still, even mysticism about it: businesses and even individuals are afloat on seas of data, with everything they need to know somehow hidden in there.
If it sounds like another tech industry ploy to prize wallets open, that may well be because it is. But when you strip away the jargon, data analytics has a lot to offer.
For most of us it might not be the revolution we have heard about. However, a level-headed focus on transforming raw and often confusing data into information that can be used to make decisions can help even small businesses to work more efficiently, drive sales and, perhaps most importantly, stop wasting money on things that don’t work.
This is all fine for the giants of enterprise, of course, but what about the everyday business?
Michele Neylon, chief executive of Blacknight Solutions, said that there is a cost of entry to the world of analytics and, as a result, larger companies are more likely to be in a position to be able to build it into their business processes.
“It depends on the size of the company: overall, making data-derived decisions is something larger companies look at a lot more. They will have analysts and data scientists and so on. For smaller companies, that will seem more out of reach,” he said.
This need not mean that data is a closed ledger for smaller businesses.
First of all, Neylon said, data can be a misleading term: not all of analytics is about analysing vast quantities of unstructured information collected from user behaviour online. In fact, a lot of it is traditional business information, such as sales contacts for example.
Secondly, the wide array of tools means there are options for businesses of all sizes.
“The reality is that there are a lot more tools out there that smaller companies can use. There’s Octoboard, and Microsoft 365 has a range of tools like Business Intelligence,” he said.
Neylon said tools like these make the key information that businesses want to make use of more accessible. “I am not a data scientist and I do not get excited by spreadsheets – but I do need to know things, so if you can make that data easy to understand that’s great,” he said.
Ciara Dempsey, senior manager in data centre, compute and solutions at Dell Technologies Ireland, agreed. “It’s not necessarily only larger companies. We are seeing it across all industries and verticals,” she said.
There is nonetheless a disparity, not only in technology but also in skills. “It’s the large enterprises that have access to the resources: data scientists and engineers. They have the budget and can snap them up.”
This is exacerbated by a pipeline problem: there aren’t enough data specialists and their training has not always been straightforward.
Dempsey said that this is further complicated by the fact that data scientists may not even be the people you want analysing the data. Instead, the ideal would be to make the data accessible to other specialists in key business areas.
“When you have someone within that company analysing data, it can’t just be mathematical. You need to have the domain knowledge,” she said.
In other words, sales staff understand sales and human resources staff understand HR in a way that disinterested outsiders do not, no matter how sophisticated their mathematical models.
“It's not enough to approach it as simply data. The role of a data scientist has changed. There are now more data engineers, and the change is being addressed in Ireland with the likes of CIT’s [Cork Institute of Technology] AI master’s degrees – but that takes time,” she said.
The initial questions being answered may seem banal – whether you need to hire more staff, whether you need to change the roster, how many phone calls you get a day, what time of day they come at – but they help drive decision-making and, therefore, cut to the heart of business.
“That’s just basic business, but the amount of data we’re dealing with has made it more complicated,” said Neylon.
Integrating the systems
If data really is to become central to every small and medium enterprise (SME) in Ireland, there will need to be more than a glut of data scientists or even new software that makes analytics accessible. The data itself will also need to be accessible. The fly in the ointment is that, right now, it simply isn’t.
“The tech nirvana is having everything in one place,” said Neylon.
“The reality is we have legacy systems and we’re trying to get them to cooperate with one another. Any business that’s transacting these days, be that business-to-business or business-to-consumer, is going to have some sort of social media presence. You’re probably advertising, too. That’s a lot of data coming in and, frankly, that’s only a little sliver off the top.”
Dr Scott Fischaber, co-founder and chief operating officer of Analytics Engines, said that his goal is to integrate and augment that data so that users can pull out insights from it.
“What we’re really trying to do as a company is around the fact that lots of organisations have data that’s sitting around in different places. If you have a nice relational database, it makes it easy – but the real world isn’t like that; it’s messy and it’s difficult to connect CRM data to publicly available data, and so on,” he said.
Information across industry
Naturally, tech and social media companies are those we associate with making use of data – in some cases, for the worse. The scandals around social media scraping being used to target political advertising being just one example.
Somewhat less controversially, companies like Google and Amazon have built their entire business models around understanding their customers – perhaps even understanding them better than they understand themselves.
Loughnan Hooper, founder of Dotser, said his recent work with the agricultural sector showed that data can be used in any sector.
“Everybody has to move forward with the online side of things. We’ve been developing our platform for the last 15 years. At the most basic you are asking, ‘Did that person get as much work done as we wanted or did we just spend two or three grand on nothing?’” he said.
Hooper said that entry-level options, right down to sharing spreadsheets among staff, should not be mocked. “You can do a lot with a Google Sheet: it’s online, you can share a live figure, you can have 20 people in there. Discussions can be recorded with time stamps,” he said.
When the time comes to move beyond the spreadsheet, though, the software is out there regardless of the sector.
“I’m reviewing Australian agriculture shows; there are 580 of them, and they’re broken into regions and sub-regions. How do I focus on only 20? Do I go by size? By date? That’s all data. It can be huge amounts, or it may be internal information.”
The coronavirus pandemic, specifically the restrictions placed on workplaces, has also driven Dotser’s use of data. The company dramatically cut back on the amount of people working from the office, so data became part of daily operations.
“With lockdown we track every task, every job. It’s not to catch people out; we’re lucky to have a team who are dedicated. We shut down the office, people went home with their computer and they were straight up-and-running from home,” said Hooper.
“We’ve done bigger deals in the last six months than ever before. We did them online. People say, ‘If I have to do another Zoom meeting...,’ but learn to manage it better. Show a bit of passion.”
Beyond transactional data and public information, one key source of information can be found on the factory floor: machinery.
Internet-of-things (IoT) technology has come into its own in the manufacturing sector in particular, with sensors measuring environmental factors such as heat and vibration to perform predictive maintenance. This cuts costs in the sector by minimising unscheduled downtime.
“Manufacturing has always been a leader in data gathering, particularly from an IoT perspective. Obviously 5G is essential for them, moving forward,” said Dell’s Ciara Dempsey.
What is instructive about this is the location of the computation: increasingly it is performed not ‘in the cloud’, but right there on the factory floor.
“It’s all very well having it all in the cloud, but that model’s not going to help you with the volumes of data we’re seeing. The risk is that the data explosion will result in unactionable noise,” she said.
Where data should be processed is a technical question, but it too can be driven by focusing on business outcomes: moving everything to the cloud and back has costs associated with it, both financially and in terms of time.
Dempsey said that a less haphazard approach will start to make itself clear once business decisions are put at the centre.
“It’s very ‘we need to gather data’, and then people don’t know what to do with it or where they are storing it. A recent Gartner study said [only] 20 per cent of data is processed at the edge, but that that will rise.”
Dempsey said the question was not about pitting on-premise versus the cloud so much as sorting data to ensure that the right information was processed and stored in the most effective places.
“It’s getting people to understand that infrastructure is important,” she said.
“While part of the approach we take is developing multi-cloud solutions, I think what people need to look at is the capability of the devices – and to not underestimate the compute and network power necessary to process all of that structured and unstructured data.”
Businesses are beginning to recognise that the ability to make use of data allows them to move up a gear, said Scott Fischaber. His experience has shown that interest in analytics is growing, and has long since escaped the lab.
“I would say data is essential for all businesses, even down to one-man plumbers: there are trends in their workflow that they could better understand,” he said.
“We’ve run a big data conference in Belfast yearly and when we started it it was very technology focused, but over the last seven years we’ve seen it become much more business focused.”
This change indicates how entwined data has become with daily life, and Fischaber said that businesses that are aware of this are seizing the advantage. “In my opinion it’s the people who are using that data that are moving ahead,” he said.
The sector is in its infancy when it comes to moving away from ‘pure’ data specialists to domain experts, however. Fischaber said that this inevitable move will be key to the integration of data into business sectors not typically considered leaders in IT and computing.
“We provide a lot of expertise in data science. We’re not museum experts, but we work with professionals at the National Gallery [in London] who are. It’s about taking their knowledge about that and overlaying it with complex analytics,” he said. “It bridges the gap in terms of what they want from the data and what they have the ability to do.”
Leaving aside whether the data being subjected to analysis is structured or unstructured – and typically it is easier to begin with structured data – one aspect that must be structured is the approach a business takes to its data.
“I think there’s a lot of pressure from the top in organisations to say ‘we need to be doing something in data, we need to be in this space’,” said Fischaber.
By focusing on clear goals, businesses can start to think about what data they need to work with and begin to expand beyond the starting point of internal operational sales data or customer data
“The next buzzword will be ‘small data’. You don’t need massive datasets, you can work with the data that you already have and combine it with public data. That combination is powerful,” he said.
This, along with automating processes, has an obvious appeal as it gives staff back time that can then be used to concentrate on their actual jobs.
“Some are already using publicly available data within their manual processes, but it’s generally a manual trawl so they’re not doing it in an automated way or in a way that the data can be captured for more than that one-off task,” he said.
Fischaber said that lingering confusion about technologies like artificial intelligence (AI) was finally clearing up as people experienced them in the real world rather than mediated through fiction or in breathless reports of pure research.
“AI definitely still has some connotations around it, but people are coming to the realisation as to what AI can do. When people start seeing things like self-driving cars or technologies that assist radiologists in finding tumours, they understand it,” he said.
“It’s that collaboration of the human and the machine that is much more powerful than what each can do individually.”
Loughnan Hooper from Dotser said that while there was still work to be done, data was best understood as a by-product of doing business, not a magical commodity. “Data is important if you can pull it together, [but] it’s time consuming to get it collated,” he said.
For Hooper, any obscurantism in how technologists and analysts – and journalists, for that matter – have presented the use of data in business is simply a roadblock that can be bypassed.
“Some of it is the industry’s fault. They want to claim this [almost] ‘magic’ or science behind it because they want to charge a fortune for it. What is AI? At the end of the day it’s a database. It might be smart, but it’s a database: rows of information.”