Ron van Kemenade is not a man given to platitudes. Sitting down with The Stack in his exclusive first interview since joining Lloyds Banking Group in June 2023 as Group Chief Operating Officer, his prelude to several questions is “to be brutally honest with you” – and he proceeds to be just that. It’s a welcome trait in an industry not always known for such candour when it comes to talking about technology transformation efforts.
Van Kemenade’s role in many senses is also akin to a Group Chief Information Officer (CIO) one. The buck stops with him for technology across the banking group – which includes logos like Bank of Scotland, Halifax, Scottish Widows and many others. The group’s CIOs are among his approximately 13 direct reports and, as a former CIO and CTO (at Dutch bank ING) he’s also an avid technologist himself.
Sitting down to chat with The Stack in Lloyds HQ, he admits that he could hire a Group CIO to sit between him and these technology leaders, but “they would,” he says wryly, “be a poor person, because I have too much of a vision and a view about technology to leave him or her alone.”
His appointment in June 2023 came as Lloyds Banking Group Charlie Nunn (who took office in August 2021) worked to push through sweeping reforms – rethinking internal structures, focusing on improved digitalisation and bringing in a reform-minded cohort of leaders. The reforms needed how technology programmes were delivered to be fundamentally rethought, and it needed, well, a brutally honest view.
Technology reform at Lloyds: thinking big
There’s simply no way to sugar coat it: two years ago, Lloyds Banking Group’s technology was a mess, its talent spread thin and its reliance on outsourcers almost absolute. Was Ron ready for this?
“Charlie and our chairman, Robin Budenberg, did a very good job at making sure that I had no expectations whatsoever in terms of how ‘good’ it would look,” he recalls wryly.
“If anything, they did a very good job in underselling the job, emphasizing how much would need to be done – which was actually why I joined in the first place. I like big things. I'm a bit of a… not a dreamer; but I like to do ambitious things that have a 50% chance of failing massively.”
Putting the horse before the cart again
“The biggest part of my job in the first year was to make sure that [technology] roadmaps were fully staffed, funded, committed, etc,” he adds. “The second part was the way of working. Traditionally, you [had] a bit of a separation between tech people… and they [were] kind of ruled by business people – as in the business people own the budget, they define the priorities and tech people are more in execution mode.”
Ron introduced a “platform model” which has had three iterations. Firstly, he moved to create “dual leadership” so brought together technology platform and “business platform” owners – bringing product owners, project managers, business analysts, software engineers, DevOps engineers, database specialists together in a “two in a box” model.
See also: The Big Interview, with Nomura Group CTO Dinesh Keswani
“I tend to do my transformations a bit like software releases. “So every year we have a release” the Group COO says.
Last April was Platform Model 2.0; April 2025 will see Platform 3.0 released: “Every release has a theme, something big we're going to change, and we do an evaluation of the last implementation, like, bugs in our implementation,” he explains.
Platform 3.0 is focused on “achieving engineering excellence” and delivering “great outcomes” faster for customers.
Version 2.0, meanwhile, was about moving from a “project-driven backlog to product-led” van Kemenade says, something that involved the team identifying and defining “all the products in the company that create value for customers or colleagues; that could be enabling products as well, like a cloud platform or a data platform or an AI platform.
“We’ve now harmonized the way of working across all 40 platforms.”
40 platforms: Say more?
Can you explain what these 40 “platforms” are, The Stack asks?
He lists some of them out. There is the digital platform for web applications, a ‘colleague platform’ (HR et al) a ‘finance platform’, there is core banking, business lending…
Van Kemenade adds: “Then there are enabling platforms like a cloud platform, an engineering platform, a data platform; call it functional domains, if you like. But they are defined by ‘these are the products I'm delivering to customers, directly to colleagues who serve customers, or to engineers that build and maintain all the technology.”
Legacy technology
Like most banks, Lloyds has no shortage of the old rubbing shoulders with the new. It still runs AS400 or ‘I-series’ systems, Solaris, HP Unix, HPE NonStop, Unisys OS, IBM Z/OS, alongside modern containerised applications.
“We have decommissioned over 500 legacy applications and moved almost half our applications to the cloud” he adds.
“This is tremendous progress. We actually have some highly developed, very resilient, and reliable systems in the bank so it’s not about removing every legacy system, it’s about modernising and simplifying those that no longer serve us.”
12 cloud landing zones is… 10 too many?
This legacy of inherited applications and systems, its Group COO emphasises, is “not unique to Lloyds.”
“This is what you typically see in every single organization where the architecture function is not leading but more or less rubber-stamping, and project managers get in the lead. Even on the modern side of things like the cloud environment, we have 12 so-called landing zones” he adds.
Too many?
“I only need one or maybe two…”
It is clear, also, that van Kemenade is not about to start singing a paean to multicloud: “[Do you need] to be able to move applications from one to another cloud provider [for resilience]? I actually think it's nonsense.
“Amazon, Azure or Google have a very resilient underlying infrastructure. So I believe in multi-zone for operational resilience, or multi-region, if you believe that they actually may have outages across multiple zones.”
"You could argue in the end you only need Linux and Windows"
Rationalisation of the data centre estate is well underway, he adds. When van Kemenade joined Lloyds it still had 18 data centres in the UK alone. That’s now down to 12 and will be just seven by the end of 2025.
By the end of 2027, he says “we'll have moved everything to the target two data centers on operating systems…” he tails off gently: “You could argue in the end” he says wistfully, “you only need Linux and Windows…”
We’ll come back to this and the technology, but The Stack wanted to know more about the cultural and organisational sticking points first.
Wrestling back control - and insourcing
Simply trying to brute-force through some technology modernisation without understanding and transforming how the bank got here is critical to delivering real improvements, he emphasises. To make that happen however and make the technology function and his technology leaders real “change-makers”, they need talent. Van Kemenade and a meaningfully supportive executive team have gone after this aggressively.
Arguably Lloyds biggest issue when he joined was a lack of in-house technology talent: “Most engineering work had been outsourced,” he recalls. “I think at that moment in time, 70% of our technology workforce was non-permanent workers…”
How’s that looking now? “It’s flipped to maybe 60% (in-house) 40% (outsourced)” he says; “we’ll probably move it up to 80-20. Over the past 1.5 years, we've hired over 3,500 engineers, both in the UK and in India.”
Building the bench…
That engineering depth has been complemented by leadership hires, including from across the technology world – something he seems particularly proud of, not least that banks (traditionally viewed as compliance-centric, risk-averse, legacy technology-laden) have a reputation for being painful places to be a technologist.
“There is a positive vibe around technology in Lloyds at the moment. Two years ago it was completely unthinkable we would hire senior people from Google and Amazon”, he says, citing the hires of Amazon veteran Dr Rohit Dhawan as Group Head of AI, and Google veteran Ulku Rowe as CIO, commercial banking among the recent examples.
“They come with a very solid background of what good looks like. If you have worked at Lloyds alone for 20 years, how do you know what good could look like? Unless you have a very curious and inquisitive mind, and you go and look for best practices all over the world, but not everybody is tuned like that. So bringing in that view from outside, really helps.”
‘Not multicloud, 'Anycloud'”
Going back to that actual view on technology and what he describes as an ‘any cloud’ approach over a multicloud one, we push for more detail on what he means by this. He’s happy to explain his view.
“We try to choose technologies in our implementations that are not proprietary to a particular cloud provider. So if we want to move, we could do it. Sometimes that is very explicit – like on Kubernetes, we try to use only the functionality that is as close as possible to open source. But it doesn't need to run in two different cloud environments. I think that that's a bit of an unnecessary thing.
Portability, in short, is important; but multicloud failovers for him, are not the default answer to resilience.
How much private cloud will remain?
With that DC footprint being cut back and more applications heading to the cloud, what does the public-private cloud mix look like, we ask?
“Our setup for the next at least five years will be a combination of public clouds and a private cloud on-prem,” he admits.
“Do you need a private cloud?" he asks rhetorically. "That's an often asked question – and my answer is ‘yes’.”
There’s two reasons for that, he says.
“One is there are applications that could run on a virtualised infrastructure, but don't have the right cloud-native patterns to be able to shift them to a container-based environment in public cloud – just moving a VM to public cloud doesn't add a lot of value, and adds a lot of cost.”
See also: The Big Interview with JPMorgan's Group CISO Pat Opet
“The second reason is it would be too expensive to refactor all those non-cloud-native applications. Give you one example, there are still a lot of our applications running on [IBM] WebSphere. Yeah, we could refactor all of them to be working with Tomcat as a web server, or any cloud-native kind of web server. That's simply too much work,” he admits.
He adds: “So the private cloud should add value where the public cloud basically is either too expensive or it's too much effort to even go there. Everything that we believe is highly data-dependent, or everything that needs to scale easily, we will bring to public cloud. Many of our older systems, the systems of record in particular, will remain on-prem.”!
Core banking is being modernised, however, he says: “We recently implemented our first products on Thought Machine, which is a very modern architecture for core banking residing in public cloud, but today that only hosts 1% or 2% of our accounts, but it's a good start…”
“It's too early..."
It’s time to wrap up. Looking back, what is he most proud of so far and looking forward, what are the biggest priorities?
“To be brutally honest with you, it's way too early to call any victory," he says, after a short pause. “Am I proud of all the intermediate results? In how the team is performing, how they have come together to a very strong team that supports each other, that is collaborating? The impact we have on the organization, the delivery of a completely overhauled, new mobile app?” he asks rhetorically. “All of that is true.
“But for me, I signed up for at least four years, and now at 20 months,” he says, with the distant expression of someone seeing a serious job ahead still, “I think I will definitely need my full four years before I call victory.”
