Democratising analytics: A case of IT vs the business

'Keep your enemies close,' says Toys 'R Us strategic analytics director Pat Murphy

Two personalities at April's Teradata Universe conference in Hamburg stood out in terms of their crossover appeal: a presentation by strategic analytics director Pat Murphy, from Toys ‘R Us, and the flipside argument from IT man and oddly-titled "chief subject matter expert for information management" at Lloyds Bank, Simon Howarth.

Murphy counts himself as ‘from the business' while Howarth is very much ‘IT'. But when it comes to democratising analytics, everybody needs to be asking the same question as funk band War's 1975 hit: "Why can't we be friends"?

Data scientists like MacGyver

"I'm ultimately in control of everything - I need to work with my business partners and IT partners to get things done," says straight-talking New Yorker Murphy.

"So with business and IT, I've taken the stance, ‘Keep your enemies close'."

Murphy talks of the business side of a company needing intimate access to processing software in supply chain, and the need to 'operationalise' analytics.

"But now, I'd argue we're in a world where you need to be flexible and play things quickly, but how can that possibly work?" he asks.

"So we now have weekly meetings with our IT partners, and we are almost attached at the hip when it comes to the complexity [of analytics needs] coming through, and when technology means we need to do something."

At this point, explains Murphy, he has both "the business case and IT" on his side.

"And then we go to the powerful person - the CFO - asking for what we need. So technology through relationships has become easier to acquire and to fit with the overall vision."

It all sounds very equitable, but Murphy admits it's the business side that has the biggest say on things, with IT often left to improvise.

"Statisticians and data scientists should be like [TV action hero] MacGyver," says Murphy.

"With a shoe, some duct tape and a little bit of wire, he could get out of jail and build a bomb, and save the day. So that's kind of the way I like to think about our analysts - they may not know the best technology, but they should be able to take a shoe, and build [what they need] to get the job done.

"I think analysts should be able to break stuff, but they shouldn't necessarily know the right tool, or if you haven't given them the right tool, they should be able to build the right tool. So we've taken down Aster [Teradata's big data "discovery" platform] multiple times with bad theories, and we've hacked together solutions with tape and glue and all that. You should be innovative and creative enough that you're solving problems no matter what," he says.

Enter Lloyds' information management IT man Howarth, who has led the democratisation process from a technology viewpoint while combating his sector's own particular problems.

Democratising analytics: A case of IT vs the business

'Keep your enemies close,' says Toys 'R Us strategic analytics director Pat Murphy

"Getting the banking world to realise the value it's got in its data and that that data is the future has not been a short journey, and it's a difficult message to get across," he says.

"But it's an important message to get across, and Lloyds has to learn that message, so we generate all our customer leads, all our campaigns, through this system, and we feed this out into fraud detection, risk, regulatory reporting, financial services, compensation schemes, and all to UK regulation [standards]; the warehouse is core to doing that."

Howarth describes a "very active user community" who've required a little hand-holding when it comes to accessing the bank's data.

"We've worked to make it easier for users to get it to the point where a person who's not a data scientist can work with the data, ask questions, and work out where the business has to go," he says.

"We have a capacity expert, and it's his job - and he's superb at it - at managing the unruly business users who want to come in and use our data, and mess up our database and trash the box.

Murphy says that his analysts thought they weren't doing their job "unless they trashed the box and brought it down when experimenting, and we took this big intake of breath and said 'they're not bloody doing that with my system'".

Power to the people

But this unruly attitude can be vital to unlocking the value of data, and spreading the power of that data across the company, Murphy adds.

"I should be able to run through a brick wall to solve an issue - I don't care if technology is an issue. So if your analyst isn't breaking things and doing really stupid stuff, they're probably not that good, because you need that creativity and that proactive striving to solve at all costs, even if it's dumb when you try and 'operationalise'," continues Murphy.

"You need to say ‘You guys can go with the technology - I'm going to build the solution'. And the one thing we've got really good at is - and this may be true especially of smaller organisations - is that we're democratising analytics.

"That's a buzzword," admits Murphy. "But if people aren't using our intelligence, it's useless. So we've done monthly newsletters, we're out there doing campaigns with marketing, we have SharePoint sites, with their own in-site surveys. We're just presenting everything. We say 'Hey, we know a bunch of stuff about our customers you probably just want to know'. We're basically trying to make our brains downloadable - that's our mission."

Lloyds' Howarth may be wary of the Flash Harrys from the rest of the business wrecking his carefully-constructed servers, but he definitely agrees about the need to enlighten business users about the power of data.

Democratising analytics: A case of IT vs the business

'Keep your enemies close,' says Toys 'R Us strategic analytics director Pat Murphy

"The key is the people, and unlocking the potential of the people, the data and of the organisation that you've got, and the interactions you have around that organisation," he says.

"So core to that is culture, and we probably don't spend enough time talking about culture. So are people committed to data - do they understand data and get that information daily? Do they interact with that information? Is that information right - have they got the right numbers?

"Because [it's useless] if all that time and effort is spent, but different people have their own pet data sources, and they spend time arguing about the bits they've got as to who's right and who's wrong. Basically people - not to mention the regulators - want to have certainty that the decisions being taken are based on understood, agreed, reconciled, great data."

Howarth describes an "interesting dynamic between business and IT".

"Murphy put the case very well that from a business point of view, they want to be quick, they want to be agile, they've got things they need to do, they know what's going to happen and they know what the vision is, but they need IT to help them.

"'IT are [seen by the business to be] slow and cumbersome... too process-bound and non-responsive,'" he says.

"But from an IT point of view, we're looking at the business and saying, 'Look at that cowboy over there, trashing my system. He doesn't understand the security aspect, or the governance required, or policy updates, or how to write the codes to run the system with'."

Howarth insists that this friction needs to be tackled - and he believes this is where the use of multiple business routines comes in.

"That's where I'm passionate really, about people changing everything, and people coming together under that reorganisation. It's not easy, but it's done by building up relationships between people.

"To meet their demands and service needs, you need to provide them with easy ways to work with that data, to interact with it and have good experiences."

CDO David Grant, who joined Lloyds in 2015, also "understands the value of data", and the board has "seen the use of digital explode" in the time since.

"We have senior management with two big mantras - service is our number one priority, as well as challenging the norm and unnecessary process. Which could be seen as opposite, clashing priorities, but the key is around agility and control," says Howarth.

Democratising analytics: A case of IT vs the business

'Keep your enemies close,' says Toys 'R Us strategic analytics director Pat Murphy

The bleeding edge

It's not sounding wildly different on either side of the business vs IT fence at this point.

For Murphy, who obviously values his analysts and data scientists, his main priority is keeping them engaged.

"To keep them fresh and to keep them engaged I want exploratory problems, short-term, hidden, and long-term or behemoth problems," he says.

One area that is keeping them busy and challenged is the supply chain, he says.

"We need to understand our [supply chain] stakeholder groups - it could be store visits, or process paths, we just need to understand the problem. I'm not necessarily going to know your pitch points or your business, but we get out there and we get smart, so we can build whatever it is into an actual problem.

"We ask where we're going to land data, and then we need to solve it."

But the hard part, he explains, is making sure the processes are in place that ensure all his analysts' hard work actually translates into business benefits.

"The relationship between analytics, IT and the business is critical," says Murphy.

"So if I come up with all these solutions, but we can't come up with an efficient way to deploy them, it's all been essentially useless. If you go to your business stakeholder with everything you've done and they go ‘well, what do I do with this?' you've failed."

To explore the power of analytics and then help the business harness its data, Murphy set up what could be described as a centre of analytics excellence.

"We did some analytics work over the time as a natural part of our business, but we just happened to get good at it, and it got more complex, as did the problems that came through."

Over time, the unit was bolstered with the addition of "formally-trained data scientists".

"You start small, you prove things - it's fun to do. It's enjoyable," agrees Howarth.

"Get multi-disciplinary teams together - IT and people from the business who know what they're doing, and work with people from the [partner] companies who also know what they're doing. We know our business and we can make it happen."