The world’s depen­dence on data has nev­er been greater. Enter­pris­es rely on ana­lyt­ics to take what are often huge vol­umes of raw data cir­cu­lat­ing through the busi­ness and trans­form it into rich, action­able infor­ma­tion. How­ev­er, with many organ­i­sa­tions, this isn’t the case.

“Imag­ine you want to make a tool or a car or a ship or a bridge out of steel, and you’ve got a huge pile of iron ore. You need to refine the ore into iron or steel before you can make some­thing out of it. That metaphor is incred­i­bly accu­rate for the world of data,” says Matthew Scul­lion, founder and CEO of Matil­lion, a cloud data inte­gra­tion and trans­for­ma­tion plat­form.

“We have a gigan­tic pile of data that’s get­ting big­ger and big­ger at log­a­rith­mic speed, but you can’t make any­thing out of it and turn it into ana­lyt­ics-ready data. There’s this huge moun­tain of pos­si­bil­i­ty that you can’t tap into until you’ve refined it.”

Like those iron work­ers before the Indus­tri­al Rev­o­lu­tion, the qual­i­ty of the out­put was ques­tion­able, the pro­duc­tion rate was low and, impor­tant­ly, only a few peo­ple had the skills to do the work. This, says Scul­lion, is where we find our­selves with data today. This scarci­ty of skills and process­es has cre­at­ed the infor­ma­tion gap with­in organ­i­sa­tions.

This infor­ma­tion gap is par­tic­u­lar­ly preva­lent in large enter­pris­es. The larg­er the busi­ness, the big­ger the pile of data that’s valu­able but not yet use­ful. Addi­tion­al­ly, there’s more tech­no­log­i­cal com­plex­i­ty across the enter­prise, and if you’re a bank or a man­u­fac­tur­er or a huge retail­er, only a tiny frac­tion of the peo­ple who work for you will be data ana­lyt­ics experts. This means the prob­lem is more pro­nounced — and the stakes much high­er — in the world’s biggest organ­i­sa­tions.

Data, data everywhere…

So what is caus­ing this infor­ma­tion gap? Why is there such a dis­con­nect between own­ing valu­able data and turn­ing it into some­thing use­ful and usable?

There are sev­er­al rea­sons. Data may be dis­persed across the organ­i­sa­tion in silos, out of sight – and mind – of busi­ness deci­sion-mak­ers.

There are also tech­ni­cal bar­ri­ers to com­pa­nies fail­ing in their efforts to achieve data-dri­ven deci­sion-mak­ing. Datasets that live in lega­cy tech­nolo­gies that can’t keep up with the demand for rapid insights and the diver­si­ty, wide dis­tri­b­u­tion, and dynam­ic nature of mod­ern data. The result: ana­lyt­ics that yield inac­cu­rate and incom­plete infor­ma­tion, or infor­ma­tion that is siloed from the users who need it.

For exam­ple, any high street bank will have a com­put­er sys­tem where accounts are stored. They will know cus­tomer infor­ma­tion such as bal­ance, with­drawals and deposits, and any charges. They will also have a cus­tomer rela­tion­ship man­age­ment sys­tem with the customer’s con­tact details, mar­ket­ing pref­er­ences and any­thing else they’ve learned through their inter­ac­tions. Per­haps there is anoth­er sys­tem that tracks how and when a cus­tomer request­ed help from a ser­vice desk, or how many times and in what ways they’ve used online bank­ing or mobile bank­ing. And then final­ly, there will be infor­ma­tion from cred­it ref­er­ence agen­cies about cred­it his­to­ries.

“The prob­lem is that if a bank wants to cre­ate a machine learn­ing algo­rithm to tell them if it’s a good idea to extend cred­it to a cus­tomer, or whether they will appre­ci­ate you offer­ing them a cer­tain prod­uct, then they need many of those parts of the sto­ry in one place,” says Scul­lion.

The process of mak­ing data use­ful relies on join­ing it togeth­er and organ­is­ing it so it’s neat and tidy, and aggre­gat­ed to the right lev­el to make it use­ful for ana­lyt­ics. But in real­i­ty, there can be hun­dreds or thou­sands of steps to go from raw mate­r­i­al to ana­lyt­ics-ready data. And of course, these steps need to be done care­ful­ly. They need to be main­tained over time. They need to be secure, to make sure the right data doesn’t get into the wrong hands. And things like per­mis­sion rules and geo­graph­i­cal reg­u­la­tions must be care­ful­ly fol­lowed.

“On a very small scale, it’s pos­si­ble to do that with­out much extra tech­nol­o­gy to help – you could just write some code and you’ve got the ana­lyt­ics-ready data,” explains Scul­lion. “But when you scale that up to a large com­pa­ny, then doing it in that man­u­al way isn’t quick, scal­able, secure or main­tain­able enough.

“We can use data to make the world a much bet­ter place. But if the pipe sup­ply­ing us is too small, then the rate at which we can do that is mate­ri­al­ly reduced. What the world needs and wants is tech­nol­o­gy that widens that pipe and allows organ­i­sa­tions to turn raw data into use­ful data more quick­ly.”

Bridging the information gap

What, then, is the answer? “The world needs a tech­nol­o­gy plat­form, a high­way of data, where all these dif­fer­ent bits of raw mate­r­i­al can come togeth­er and be refined into ana­lyt­ics-ready, trust­ed, use­ful, clean data that can then be put to work,” says Scul­lion.
“We need a tech­nol­o­gy lay­er to solve that prob­lem, and that’s what Matil­lion has built – a data oper­at­ing sys­tem on which enter­pris­es can at scale make that data use­ful. And in doing so, they can bridge that infor­ma­tion gap, fix that con­straint in the sup­ply chain of data and make the world a bet­ter place by point­ing to this new com­mod­i­ty that is data.”

---

There are now more than 1,000 enter­pris­es around the world using Matil­lion to bridge their infor­ma­tion gap. After launch­ing the plat­form in late 2015, Matil­lion has become one of the fastest grow­ing and most valu­able late-stage star­tups in both the UK and West­ern Europe, dri­ving excit­ing lev­els of growth.

Matil­lion makes use of the lim­it­less, cost-effec­tive pow­er of the cloud. “The cloud has unlocked this gen­er­a­tional shift in the use of data,” says Scul­lion. “Before the cloud, we couldn’t afford, and there­fore we didn’t have, the ambi­tion to cap­ture the data and to put it to work. But the cloud pro­vides this incred­i­bly cost effec­tive, pow­er­ful and easy-to-make- use-of tech­nol­o­gy to cap­ture and store data, and then put it to work.

“That’s the back­drop against which the world is piv­ot­ing to data and against which data is becom­ing the new cur­ren­cy.”

Addi­tion­al­ly, Scul­lion says Matil­lion can be stood up in min­utes, and can be learned in a few hours or days. The time to val­ue is a few days or weeks. “That’s because it’s sim­ple to launch, man­age and use, but not at the expense of capa­bil­i­ty,” he says.

“With a tech­nol­o­gy like ours, which is visu­al, we call it ‘low code, no code, code option­al’. So you can make data use­ful with­out need­ing to write code. That makes it much eas­i­er, and more intu­itive to use. It also makes it much faster and more pro­duc­tive, but still pow­er­ful.”

Matil­lion is secure, scal­able and per­for­mant, and can coex­ist with the enterprise’s rules, poli­cies and pro­ce­dures, and oth­er tech­nolo­gies. Addi­tion­al­ly, the tech­nol­o­gy can be used by a much broad­er spec­trum of busi­ness users, putting the val­ue of the data into the hands of more users.

---

Q&A: Modern analytics and the business of misinformation

Ciaran Dynes, chief product officer at Matillion on how businesses can make better use of data analytics and the skills they will need to do so

What are some of the chal­lenges fac­ing mod­ern organ­i­sa­tions when it comes to data ana­lyt­ics?

A lot of indus­tries evolved with exec­u­tives that were nev­er required to work with each oth­er; that cre­at­ed arti­fi­cial bar­ri­ers. But if every­body is gen­uine­ly putting the cus­tomer at the heart of their busi­ness, there will be no prob­lem with shar­ing the data. They need to fig­ure out ‘what would a bet­ter cus­tomer expe­ri­ence be for us?’ And then work out ‘what ques­tions do we want to ask?’ Then it should be the case that the team is required to work togeth­er.

That is what hap­pened large­ly in the big data era. Peo­ple have realised it’s bet­ter to share inter­nal data, and then to pick their bat­tles — the one or two things that they could do to basi­cal­ly improve cus­tomer expe­ri­ence, reduce risk, cut costs. It’s pos­si­ble today because the tech­nol­o­gy is avail­able to do those things.

Mod­ern ana­lyt­ics is much more about pre­dic­tive­ness; there’s no absolute cer­tain­ty. Com­pa­nies must ask ques­tions to build an answer

What is the dif­fer­ence between tra­di­tion­al and mod­ern data ana­lyt­ics?

With tra­di­tion­al ana­lyt­ics you might add up a bunch of num­bers to get an answer to a ques­tion such as, ‘who are our top 10 cus­tomers?’ But with mod­ern ana­lyt­ics, there’s not nec­es­sar­i­ly one answer. There might be a scale of answers, with a min­i­mum and a max­i­mum, and some­where between lies the answer.

Mod­ern ana­lyt­ics is much more about pre­dic­tive­ness; there’s no absolute cer­tain­ty. Com­pa­nies must ask ques­tions to build an answer. Some­times the answer will be defin­i­tive and some­times the data comes back and it’s incon­clu­sive. The skill of the busi­ness is chal­leng­ing the answer that is returned, say­ing ‘I think there’s not enough evi­dence to sug­gest that’s the answer.’

What about ‘fake news’ in the con­text of busi­ness data?

You can’t escape the mis­in­for­ma­tion in soci­ety. There’s actu­al­ly been no rise in mis­in­for­ma­tion, it’s just digi­tised nowa­days so it gets into the pub­lic domain instant­ly.
Organ­i­sa­tions also have mis­in­for­ma­tion in their busi­ness. But mis­in­for­ma­tion is not always inac­cu­rate infor­ma­tion, it is inter­pre­ta­tion, it’s con­text, it’s bias. That’s the real chal­lenge for busi­ness­es going for­ward – they need more and more and more of this dif­fer­ent data that we can put togeth­er to either prove or dis­prove that argu­ment to be true or false.

How do we get infor­ma­tion? How do we look at mis­in­for­ma­tion? How do we come up with a ques­tion or a hypoth­e­sis that may or may not lead to an answer? Mod­ern busi­ness is based on those things.

What’s an exam­ple of mod­ern ana­lyt­ics?

There are bril­liant things that we do in the world with data; there are some amaz­ing sto­ries in gov­ern­ment, in health­care and in insur­ance. I’m immense­ly sur­prised all the time by these inter­est­ing sto­ries.

For exam­ple, an insur­ance cus­tomer in Wash­ing­ton in the US is look­ing at the data of for­est fires. They pro­vide car, house and prop­er­ty insur­ance, and with for­est fires so preva­lent they have an enor­mous impact on their busi­ness. They are pulling data from things like geot­her­mal sen­sors in the ground and using weath­er pat­tern infor­ma­tion to per­form pro­jec­tions and analy­sis based on tem­per­a­ture change.

They have no choice but to have an opin­ion on things like the envi­ron­ment and cli­mate change. That is the busi­ness of insur­ance – hedg­ing bets against a set of out­comes, and some of those out­comes are pro­found for soci­ety.

Learn more about Matil­lion: matillion.com/demo

---

Spon­sored by

Matil­lion