Skip to content
Home>All Insights>Podcast: Harnessing the Storm: Data-Driven Decision Making with Glen McCracken

Podcast: Harnessing the Storm: Data-Driven Decision Making with Glen McCracken

Harnessing the Storm: Data-Driven Decision Making with Glen McCracken

In this episode of The Digital Lighthouse, Zoe Cunningham is joined by Glen McCracken, the Head of Data and Analytics at ION, a leading firm in automation solutions for the financial sector.

With a rich background in corporate finance and applied statistics, Glen delves into the pivotal role of data analytics in shaping business strategies and enhancing operational efficiency. His unique perspective, born from years of experience and a passion for innovation, offers invaluable insights into harnessing technology for better decision-making.

Listen now and discover why adopting a data-driven culture is crucial for success in today’s dynamic business landscape, making this a must-listen for professionals seeking to leverage data for growth and innovation.

Listen to the podcast on this page, or wherever you get your podcasts, and read the transcript below.

A collage of promotional images for the Softwire Techtalks podcast series. The graphic features a bright color scheme with pink, blue, and orange elements. Each image showcases different speakers, with their names and titles provided. The speakers are framed within abstract geometric shapes that match the colors of the Tech Talks theme. Logos of Tech Talks and their affiliates are visible, alongside snippets of topics covered such as leadership, healthcare technology, and software development. The overall layout is modern and dynamic, designed to attract viewers to the podcast series.

Digital Lighthouse is our industry expert mini-series on Softwire Techtalks; bringing you industry insights, opinions and news impacting the tech industry, from the people working within it.

Listen to every Digital Lighthouse episode on SoundCloud


Zoe Cunningham: Hello and welcome to the Digital Lighthouse. I’m Zoe Cunningham. On the Digital Lighthouse, we get inspiration from tech leaders to help us shine a light through turbulent times. We believe that if you have a lighthouse, you can harness the power of the storm.

Today, I’m excited to welcome Glen McCracken, who is the Head of Data and Analytics at ION, a technology firm specialising in automation solutions for the financial sector. So, hello and welcome, Glen.

Glen McCracken: Thank you, Zoe.

Zoe Cunningham: Can I ask you to give us a brief overview of your role at ION? And what ION does and how data analytics plays in your operation and strategy?

Glen McCracken: Yeah, sure. So, as you might take a little step back to talk a bit about me so people can understand why I have a funny accent. And it helps inform also the role that I do.

So, I was born in New Zealand, hence the slightly funny accent. I studied corporate finance and applied statistics. And I was lucky enough to work with Ross Ihaka and Robert Gentleman, who were the creators of R, the Data Science software. And we used R way back in 1993.

So, it was released in 1995. So, from a relatively early age, I’ve had a fascination with data and a fascination with data analytics. And that brings me nicely to I suppose ION. So, ION Analytics is the division that I work with then.

So, ION, as a parent company, is very similar to Virgin in that it has many different interests and aspects, but they all form somewhat of a common theme. So, we have ION Markets, ION Corporates, Treasury, Commodities, and we have ION Analytics. So, the role that I have is within ION Analytics.

And our mission or our vision is to transform the capital markets by providing our clients with data insights, editorial content, and merging those insights with human intelligence so that they can make more informed decisions. And my role as Head of Data Analytics and Automation within ION Analytics is the internal element of that.

So, if we say to our clients we want to help you be more effective and data-driven, and here are the products and services and data, in order to do that, my role is looking inward and saying, how, as an organisation within ION Analytics, can we leverage data to help us grow, to help us be more operationally efficient, to understand more about our clients so we can serve them better? That’s kind of the summary of the funny accent and also the role that I play with ION and ION Analytics as well.

Zoe Cunningham: Well, I think it’s really important if you are offering these services to your clients and you’re essentially setting yourselves up as an organisation with expertise, if you’re not, you know, walking that walk yourself, it’s actually kind of an important part of who you are.

Glen McCracken: Yeah, totally. And there’s a really nice quote I like, and I always forget who said it, but it says, “If we have data, let’s go with the data. If all we have are opinions, then let’s go with mine.”

And I suppose the danger we have in these modern times is typically the opinions that are sometimes followed are the more senior people or the loudest people that happened to be in the room.

 And that used to work, I think, because often those senior managers were in a position of authority because they understood intuitively the data; they understood intuitively what was going on within the business. But these days, you know, there’s so much data being produced, there’s so much complexity to all the things we do as an organisation, that you really do need to be quite open-minded around the use of data and the role that it can play as well.

And one of the things we talk about within ION Analytics, and we try to live by it, is to have strong convictions but loosely held. It’s good to believe in something; it’s good to have a set series of beliefs and principles and things that guide you. But at the same time, you need to be open-minded to the fact that there might be data or insights that come your way that cause you to challenge some of those beliefs and some of those principles.

And just like you described, in order for us to be effective with our clients, internally we have to do the same thing. We have to say, we’ll be led by our principles, but we’ll be willing to reassess those principles if data comes to mind that is compelling and we believe in and is credible.

 And then we can adopt a new set of principles and beliefs and harness that. So we’re on this kind of cycle where we try and have firm beliefs and what we’re doing, but always informed by the data and willing to take a step back and say, something has happened, the market has changed, conditions have changed, or we just now have access to new data which suggests that some of those assumptions we’ve had in the past may not have been as well-informed as they could be.

Zoe Cunningham: I think that’s really important. And I really like what you said before about the idea that maybe it used to be the case that one C-level manager at the top of the organisation essentially was acting like a piece of software and had all the information going in, and their brain was like processing and coming up with the answers.

But nowadays, it’s very hard for any human being to have all of the information, you know, for any one person not to be missing critical information. And also, if you outsource it to a system, you’re making a conscious choice about what information goes in. As human beings, we can seek out information, but we don’t have control over exactly the data set we’re operating on in quite the same way.

Glen McCracken: Yeah, it’s a really good point you’re making there as well, because there’s an emerging trend within the industry to talk about artificial intelligence and the replacement of people leveraging these LLMs and OpenAI type tools.

And I suppose our focus has always been the augmentation of humans with intelligence. So we still see the critical element that people play in the decision-making process. And that’s not to downplay the huge benefit that we’re seeing from OpenAI and whatever it’s called now. And these things are amazing as an augmentation to the intelligence of a person.

And in many cases, it may well automate a large degree of someone’s role if they’re in copy editing or so forth. But you still need that human; you still need the person to look at it and say, “This is great, that fits well, the nuances are there, the tone of voice is there, it’s consistent with all the things that I want.”

And so it’ll be great one day to be at a place where we can replace certain lower-skilled, more repetitive administrative roles. But we’re in the early days, and we’re still in that belief cycle right now, where we see the augmentation that comes from all these fantastic tools.

But still, the central point being the person using those tools, so great to have the tools, great to augment our own intelligence with these fantastic tools and do things in a far more productive and effective manner. But we still need that gatekeeper, that human gatekeeper to say, “Yep, all of this makes sense.” And hopefully, pick up on some of the hallucinations that can sometimes slip through.

Zoe Cunningham: Exactly. And I’m sure we will come back and talk about how humans fit in more later on in the episode. But I would like to dive a bit deeper into your kind of tech and infrastructure and how do you prioritise what to invest in because there’s always, you know, an unlimited amount you can invest?

Glen McCracken: I suppose for us, our starting point, and it’s a little bit cliché to say it, but we pride ourselves on being very client-centric.

So for us, as new tools and technologies and infrastructure and providers become available, our first kind of default position is how will this help our clients.

So we’d like to think we have developed a very deep and rich understanding of the problems that our clients have that caused them to use our services and our products and our data in the first place.

And by having that strong understanding of their pain points, the needs that they have, the jobs they’re trying to fulfil through the use of technology, that allows us to then, as we see other tech stacks and emerging technologies coming online, it allows us to say, we think this could be a good fit. So that starting point is the client.

And we do that in a couple of ways. We have what’s called an internal tech radar; we’re lucky enough to have around 1500 people within our division, we have very, very smart people. And we have these cross-functional teams that are looking at the different technologies coming into the market and assessing it, maybe in terms of an academic assessment, and maybe in terms of a proof of concept to kind of try things out.

But we’re always kind of looking at those technologies and saying, how does it fit? How does it make our clients’ lives easier so that when we do approach our clients, they continue to have that confidence that we’re looking out for them, that’s wearing our unselfish hat.

 The Selfish bit is where we do the exact same exercise for ourselves. So we say, the clients are taken care of, the priority is there, now from our perspective, are there ways in which we can utilise this to be more effective, to drive the business outcomes that we want, to increase our productivity or efficiency, to gain more insight into what we’re doing or what we’re not doing?

So the combination of those two things, looking at it externally with respect to how we can benefit our clients, and then internally as to how it can benefit ourselves. And there’s somewhat of a feedback loop as well.

So in some circumstances, we develop and leverage some of the technologies for our betterment, and then we decide, or, “Well, now that we’ve learned more about it, this could actually be beneficial to some of our clients.” And conversely, there might be some things that we build into our products or services to our clients that we, in turn, look at and say, “You know what, this has been fantastic for us. We could repurpose the same benefit and realise that internally.” So it really is a bit of mixture for us.

Zoe Cunningham: Yeah. And that idea of a feedback loop. You’ve answered that next question that I was going to ask, which is that sometimes you can’t know how something is actually going to be applied until you apply it.

So actually, there’s always this sense of iteration, right? That you, it’s not just a thought experiment of which one can we use, there needs to be some practical experimentation as well.

Glen McCracken: Yeah. And there’s a great example of that. So when OpenAI first started releasing some of the API functionality, the documentation was, I think I’m being generous when I say poor.

And so really, you have to play around with things to see what worked and what didn’t. And we kind of fell for the 101 mistake of thinking we now have a solution; we just have to find the problem that it can address. So we knew the solution was OpenAI and using the API.

And we thought, “Oh fantastic, we can leverage it to help our internal salespeople use Salesforce CRM software more effectively.” And so we spent some time on that, we looked at creating a type of language model associated with the training materials we had, we pulled some of the transcripts from the training videos, we pulled the frequently asked questions, we stood something out reasonably quickly, and then went to the clients, in this case, which was the sales team, and said, “Hey, go, we’ve solved a problem you didn’t even know you had.” And they looked at it. And it was on average, to say the least.

And it was average because the data was poor. So whilst we do training sessions, and we have FAQs, we don’t always keep them up to date. And we don’t always refresh them. Because we’re constantly looking at ways of improving the CRM system and making it more of a productivity tool for our salespeople.

So what we didn’t fully realise was even transcripts and training materials from six months ago may well be out of date. And when you’re feeding in conflicting and potentially out-of-date information, it’s very hard to the inside of the model put more emphasis on the newest stuff, where there’s a conflict, where there isn’t, trust the old stuff unless we haven’t told you that it’s changed because we haven’t done any training on it. It became this kind of quagmire of, have we done it wrong, have we built it wrong? Do we not understand the API well enough? Have we not implemented in a clean enough way?

When in actual fact, the real issue was the data that we had to begin with? And the fact that we started with the solution and not the problem? It’s an interesting insight into I suppose companies like us, that, I mean, we have analytics in our name, you would think that a relatively mature company like us wouldn’t make those types of mistakes. And yet, we do?

Well, I think the lesson there is we all need to be a little bit careful about getting the basics right, getting the clean data, starting with the problem, and then trying to fit a solution to the problem rather than being excited by the shiny things that come along and then trying to surprise our stakeholders with solutions that they may not have thought they actually needed.

Zoe Cunningham: And also, nice to know we’re all human, right? Everyone falls for this when there’s new technology.

It’s so exciting. And that brings me really nicely on to my next question because I wanted to ask, not just about how you’re leveraging the kind of rapidly evolving AI and machine learning, how you’re leveraging it to realise value, but also looking at it from a risk and governance perspective.

Glen McCracken: Yeah, so two really good points. And I suppose there’s that balancing act there as well. So for us, again, part of the tech radar that we have looks at the scalability, the security, the integration capabilities, we were lucky enough to have a relatively long-term vision that we’re aiming towards.

And we’re also, to a degree, governed by ION as an overarching organisation moving in a very similar direction to what they’re doing as well. So rather than us having to guess as to what our standards are, we have a relatively strong North Star that we know that we need to hit towards.

And I suppose the great thing about that is, if you’re considering risk and governance, the balanced approach of implementing robust data governance processes, ensuring that you’re using AI in an ethical manner, being concerned about data privacy, about compliance risks, about how you’re utilising AI and machine learning, and ensuring that they are aligning with those values and aligning with that North Star. It does make those decisions a lot easier because you’re not having to reassess all the time.

So even with new technology, if you have reasonably good standards in place and a reasonably strong governance and security protocols in place, then making those assessments become a lot easier as opposed to potentially if you’re not as mature, if you are not as potentially lean, disciplined in identifying what those standards should be, then every new thing that comes along potentially is a much longer and wider discussion than it needs to be because you’re figuring out new things on the fly.

So I suppose one of the great things about being part of a wider organisation, so the wider ION, but also having the autonomy to do our own things within ION Analytics, is that we have a strong sense of direction, we have that strong kind of North Star, we have a larger governance team to draw upon.

And we have kind of the intellectual rigour to go through in advance and figure out, what are we going to stand for? What, how valuable is it for us to our clients and to ourselves to have some of these disciplines in place? And what we find is that subsequent decisions become a lot easier by having that documented, by having it well-known and circulated, and not having this creeping, moving of standards when we potentially don’t have them well enough defined.

So that is how we deal with it, with kind of leveraging the fact that we’ve gone through a lot of this exercise before. Of course, they’re gonna be updated on a regular basis. But we have that kind of that strong north guiding star which allows us to make some of these decisions in a much more timely manner than potentially some of the people can.

Zoe Cunningham: I think it’s a good reminder that actually what encompasses human technology is a lot more than machines; we nowadays think of machines as technology.

And that’s what we’re building on and utilising, but actually, this kind of human technology of our processes and understandings and ways of doing things is as important, or I think no, thing maybe even more important to be able to make use of the new kind of digital innovations.

Glen McCracken: Yeah, so I spent about 10 years in consulting, and one of the areas I worked in was process reengineering. And one of the premises there was the first step within the consulting creed, the first rule is to seek first to understand. And in a process reengineering context, what that means is to even devise what the future state should be, you have to start with a solid understanding of the current state. And that current state again doesn’t constrain you to say, well, we’ve always done it this way.

And that’s why we’re doing the current state. In fact, it’s the opposite. It’s let’s understand what the current state is so that we can challenge that current state, so we can understand what are fundamentally the things that are coming into the system, the value that is created by the people or the systems, and then the outputs that come from that system.

And that system might be multiple different mini systems or subsystems within a larger end-to-end value chain, in which case, understanding what each party is doing and why they’re doing it and the value they’re creating, and what they’re relying on in order to produce something for the downstream thing. In many cases, it allows you to cut either some of those steps out or fundamentally overhaul the entire system. And the great thing about leveraging technology is you can use the exact same approach.

 So by starting with the current state, you can understand in some cases the historical reason by which people have made choices to design a system or a process in a certain way. And then the great thing about some of the emerging technology is it allows you to fundamentally redesign that system; it may well be you don’t need all the same people within it.

And maybe you need more in some cases, it may be in many cases that the people shift from doing the work to actually being the exception handlers for the work that can’t be done by the automation and the systems that are implemented.

And so what I love about the field of FinTech is that whilst it’s ever-changing, it’s ever-changing in the context of the kind of core principles that don’t really change.

So, the same principles that were used many years ago in doing process reengineering are the exact same principles you use today in understanding the end-to-end system, understanding how you can leverage and augment humans, and use technology to our advantage to speed things up, or to reduce the variability, or to allow us to identify anomalies in a far more time-effective way, or to automate some of those repetitive activities that people, by choice, don’t really want to be doing.

They want to be doing the more creative activities, they want to be more exception handlers and advisors, as opposed to, in many cases, doing an activity that really could be automated in a relatively simple way.

So, it’s an exciting time, the new tools and everything coming online are causing, I think, everyone to question a lot of the things they’ve done and we love questioning things, we love breaking things and rebuilding them, and hopefully making them better, and in turn, creating a more stimulating environment and in the type of work that we have people doing within our organisation as well.

Zoe Cunningham:

Well, just quickly, we’ve heard you talk there about bringing the same principles from the past to the present. And obviously, those same principles will run on to the future. Are there any specific skills, like trends or technologies that you’re focusing on, to apply those principles to in the future?

Glen McCracken:

What I’ve found most effective when talking about this is probably not the skills in the sense of the technical skills, but the culture of how you can establish or help establish an organisation where you are truly data-driven. And from what I’ve found, you really need two things: you need the pull, and you need the push.

So most organisations do the push. That is, they seek to have a golden source of truth with respect to the data, they seek to leverage that data and produce Insight and Analytics, they seek to have good data management, they seek to democratise access to the data and empower people to utilise that. And those are all push things, how we can help the organisation and kind of push the technology and some of the solutions on them.

What’s sometimes missing is the pull, it’s the pull from the organisation to say, we want to be data-driven, we want to rely on insights, we want people to be providing analytics and leveraging the data.

And sometimes if you have one without the other, so if you have the pull, like people saying I really want this, but you don’t have the investment in data governance, clean data, the data warehouses, the tooling, and the skills you need, then you’re left disappointed, because the organisation is saying, I’d love to rely on data more, but I just can’t.

If you have the push without the pull, then you’re putting all the right things in place, but often the insights that you’re deriving are not actually being leveraged, so you’ve got well-meaning people producing insights that may or may not be circulated and consumed in a way that is actually effective and helps the organisation.

So the Shannon information theory talks about it, and I’m torturing the quote, but it basically says that the greatest hallmark of data and information is for it to be used in business decisions to affect a change. And that’s true of an organisation and truly endless.

So the greatest satisfaction I, as an analyst, will get will be to produce analysis that has an effect to change my organisation. So your question is a good one as to what I think is important.

And in addition to Python, Snowflake, and Databricks, and those technical things, I think often the starting point is having the push mentality of knowing that data is super important. And it starts with having good clean data that’s well-managed, secure, that there is democratised access where all the right people can gain access to it and use it in a way where they can provide insights and analysis.

And the second element is the appetite. Is there appetite within the organisation that is pulling or demanding those things to ensure the investment in the tools and the training and the right people, and even creating the audience and the opportunities for those analysts to present so that they have a chance to kind of fulfil their purpose, which is to help affect change in the organisation through the use of data?

Zoe Cunningham: Yeah, absolutely fantastic. And that reminds me of a much simpler quote:  :The proof of the pudding is in the eating”, which I don’t think about very often with technology, but actually, “whether it’s being used: right? Whether it’s being used effectively is the most important part of the chain.

Well, thank you so much, Glen, for coming on the show and for sharing your insights and helping us to shine a light for others.

Glen McCracken: It was my pleasure and lovely to spend some time with you, Zoe.

Digital Engineering

Get expert help with your digital challenges and unlock modern digital engineering solutions.