Our panel discusses how to identify legacy systems, the challenge that businesses face when it’s time to replace them and the changing nature of asset managers’ IT departments. Chaired by Nicholas Pratt.
Ian Barringer (Head of IT, F&C Investments), Catherine Doherty (Chief executive, Investit), Martin Engdal (Market strategist, Advent)
Rakesh Vengayil (COO APAC and Emerging Markets, BNP Patibas Investment Partners), Steve Young (Chief executive, Citisoft)
Funds Europe: What is your definition of a legacy system? Is it purely based on age, or that in-house, one-off aspect? Or is it simply an older version of the current system?
Steve Young, Citisoft: The age of the system is clearly an indication of its level of legacy. An important factor is how bespoke it is. But there are other indications that people should look for – something that’s not been sold in the market or doesn’t have a clear release pattern or any kind of user group or community.
Funds Europe: Might some firms not be aware that they have a legacy system?
Young: Absolutely. Vendors are sometimes quite strong at clouding the level of legacy and independence and any influence that it may have, so they will often position and talk about it as if it’s a single product that’s shared across a number of clients, but the reality is that it’s heavily bespoke to an individual client. The user community need to test that and talk to other users of the system and really find out the commonality of the application across the user base.
Martin Engdal, Advent: There’s no one definition, but I guess there’s a checklist. Even standard applications can be legacy. If you don’t really have a roadmap to a product, then I say it should be characterised as a legacy system. If you’re only doing maintenance and you’re not really spending R&D money on that platform, it’s legacy. Is this platform future-proof? Can this support where we want to take our business? If you answer “no” to that, often it is a sign that this is a legacy system. Unfortunately, we’re moving into a world where becoming legacy doesn’t take ages. It can be within a year or two.
Rakesh Vengayil, BNP Paribas Investment Partners: I would not define it by the age of a system. As long as the system is current, relevant, scalable, sustainable and compatible with the new technologies, it cannot be classified as a legacy system. Moreover, it also depends on the context of the operating model, as we operate on both a global and local operating model. From a global context, we may have a piece of software which is currently in frequent use because of its flexibility and agility but, within the front-to-back connectivity of a global system, it is not compatible with the upgrades that are happening. And from a local context, there are systems that are relevant for the local usage but not compatible with our global platforms for interfaces or information exchanges.
In some other cases, a legacy system could be a complex, Excel macro which plays a critical role in the overall flow. This would have evolved over a period of time, which is person-dependant software and not maintainable in an industrial way.
Ian Barringer, F&C Investments: The big issue is a lack of support, whether it’s a vendor system or internally built. Can it be developed or extended? Is it based on a dying language or a platform that is out of support? Plenty of software out there still relies on Windows 2003.
Catherine Doherty, Investit: A legacy system is like a second-hand car. It’s the point at which repairs become uneconomic. It becomes too difficult to change bits of the system and that pushes you into a huge replacement project.
Funds Europe: Has the definition of a legacy system changed or has the length of time by which a system becomes legacy changed?
Doherty: That definition can be triggered by all sorts of things like a major integration, but other systems can hang around for ages because there is not a business case for replacing it, even though it is becoming more and more expensive to run. From a departmental view, things may look OK. They may say that they are using ‘X’ system but when you look more closely, there are Excel programmes on the outside propping things up.
Barringer: It is about extensibility. If a system cannot be extended, you end up with a patchwork of additional solutions to cater for new functionality – another asset class, for example.
Young: If you look at the number of stretches required around a system, that’s a good indication of its point in the legacy curve.
Vengayil: In the fund management industry, operations are divided into three key segments – front, middle and back office – and they are increasingly outsourced or offshored, but at the same time getting highly integrated. Therefore it’s really important that everything works in tandem, even if you are not necessarily responsible for maintaining all these three components. If one of them is moving faster than the others in terms of upgrades, then it raises a compatibility issue and forces you to examine the level of potential legacy in the other components.
Young: My strong suspicion is the age of some systems being legacy is shortening, because we’re moving to the digital age and people want to become more data-centric. So, having legacy systems as core components of the architecture is increasingly problematic and that is putting more pressure on older systems.
Barringer: The rate of change of front-office systems is definitely increasing; in turn, this is dragging other systems with them. However, they are unable to cope, which makes them inherently legacy. If they were standalone they would be fine, but the tight coupling of the many systems in a financial architecture means they have to keep track of the changes.
Vengayil: The ideal scenario would be having a master/slave architecture, but that’s not often the case because there are different external organisations involved in your front-to-back chain. They are changing the components, or responding to other external needs, so it would not fit into that kind of architecture. That’s where sometimes we see the mismatch arising.
Doherty: If you have an open architecture, you can easily pull bits of data out and you can plug in a new risk system in the front office. Whereas if the system insists on keeping its data to itself, it becomes a burden much earlier.
Engdal: A lot of the innovations have been in the front office but not so much in the middle and back office. We see some organisations investing heavily in front-end, client-facing tools but neglecting the back end. This might help them in the short term but in the longer term it tends to be an issue, especially for some of these larger, tier-one organisations. If the core engine isn’t working and future-proofed, you will eventually end up with business constraints and a costly platform.
Vengayil: I absolutely agree. We do see this in local markets as well, because in some of the local markets the middle and back-office technology has to be pretty much aligned with the local market practices, whereas the fund managers tend to use very sophisticated front-end systems because they are globally available, and when you try to integrate that with your back office, it becomes a big challenge. We have experience in countries like Indonesia where the front office is extremely sophisticated, multi-currency, multiple asset class, but the back end has limitations where it cannot respond to the sophistication you’re bringing in the front end... One cannot easily change the local set-up, so you have to be in line with what the market is doing.
Young: Legacy systems are far more prevalent in cost centres. People invest far more where they can generate revenue and they worry far less in areas which are seen as cost centres. But I think this approach is somewhat naïve, because you’re only as strong as your weakest point.
Funds Europe: According to the survey results, 50% of firms said they used legacy systems slightly, 30% said they used them a lot, and 20% not at all. Is that a fair reflection of the industry?
Barringer: It depends on how you define ‘use’. If one of your primary systems is legacy, then saying you only use it lightly because all your other systems are current is a bit disingenuous.
Young: A lot of people don’t want to admit they’re on ageing platforms, so I suspect the results are understated.
Doherty: If you said to people: “How many of your systems are constraining your flexibility because you’re finding it too expensive to extend them?” or “How many of those operations people are there because they’re making up for systems shortfalls?” you’d get a different answer.
Young: It’s also where you find out who truly understands the costs of running a legacy system. They’ll generally say they’re cheaper, which is in my experience complete nonsense. They’re not looking at the total cost of ownership, they’re looking at the vendor spend, and the vendor spend is incredibly low, but the total cost of ownership is usually vastly understated. To create a business case you really have to understand the total cost of ownership, but not many are willing to do that. Building a business case is actually quite easy, if you spend time and effort putting it together, but you’ve got to get out of denial first.
Barringer: This is very much an issue I am tackling now. The problem is primarily the bespoke solutions; the glue that people put between the back and the front, the customisations that change the system to fit the business model. That is where there is ingrained intellectual property. Referring to a previous point, it is the middle piece – the middle office. That is where everything is dumped, because nobody knows how to glue the two parts of business together.
Young: We are on the cusp of more changes in the industry than we’ve seen for a long, long time. The ability to change and acknowledge that you have legacy systems is going to be increasingly critical. It’s probably been less of an issue over the last decade, but it’s going to be increasingly a sharper and sharper issue for businesses to address. The smart businesses are looking at that. People in denial are going to suffer badly,
Doherty: The regulatory reporting thing has been very interesting because it requires you to bring data together from all the way across the enterprise, and that really brings home to you how you’ve patched things together. We’ve had some very good market conditions for the last few years, so people are more inclined to invest in their systems. There’s also a feeling that this is a transformational time in investment management. There are new market opportunities and people want to position their firm as being more customer-friendly and lower cost.
Vengayil: For the last few years the focus has been on rationalising the cost but now there is a clear realisation that what we are doing is sustainable in savings and efficiency. We handle a wide variety of front-office tools in several asset classes and that creates more situations for legacy systems. Sometimes you have a star fund manager who is used to a certain tool and has built in some of his own algorithms. From a technology perspective it may have already become obsolete, but from a functionality and emotional attachment perspective, you cannot take that away. But in the last few years we could drive those kinds of changes because of the focus on cost, people realised that they could not afford the luxury of running multiple platforms and systems. In our case we by and large have moved on to a single, multi-asset class portfolio management system and we have eliminated a lot of tailor-made stuff to bring them all into a common platform.
Funds Europe: How do you manage the economic challenge of replacing legacy systems with new technology?
Vengayil: In a lot of firms, the new innovations and upgrades are self-funded due to budget limitations. That forces the business to find efficiency elsewhere in the chain and use it in a place where one thinks it is relevant. So that automatically drives some of this rationalisation.
Young: The market is quite fragmented. In the UK and Europe, regulation’s still the number-one concern. In the US, they’re much more strategic and the big and middle-sized houses have got a much more modern infrastructure that is built for the future. That is the challenge for Europe, because there have been a lot of tactical, regulatory-led projects but firms haven’t really addressed the issues around legacy systems and they are slipping further behind other parts of the world.
Barringer: The drivers for technology have always been in the front office but more so at the moment, particularly in sales and marketing. However, I think there is an opportunity. With a number of operating systems and platforms facing retirement, there is a focus on the back office. People have realised we should start doing something and big replacement projects are being initiated. Working out the total cost is obviously still a challenge.
Funds Europe: Is there board-level involvement in these projects or are they confined to IT?
Barringer: There is board-level involvement. One of the tools I’ve been using to provide some transparency is roadmaps that highlight pre-emptive milestones – where support begins to fall off, the stage where they are not legacy yet but will be in two or three years’ time. Funding of these projects needs to occur at that point, not once they have hit the legacy threshold.
Doherty: The concept of cost is changing. It used to be solely about the cost of IT but increasingly clients are talking about the cost of the product. They are thinking in much more holistic terms about, can we run this product at 30 basis points that we used to run at 80?
We have also started to build models that look at the impact of operations on fund performance. You get drag effects like running too much cash in the portfolio but you also get the number of fund managers building up – people who are running little spreadsheets on the side – and the proportion of fund managers to AuM is proving not to be scalable. A lot of firms are now thinking about the scalability of the front office as well as the back office and it all comes back to a good platform where managers can run portfolios side by side and replicate them fluently and spend their time having good investment ideas – not trying to run the platform from the front office.
Vengayil: The two big costs for asset managers are staff and technology and a big percentage of the latter is getting earmarked towards regulatory developments and keeping the lights on. That leaves very little for new innovations and development.
That’s where a lot more pressure is coming back to the IT or business to self-fund the new projects and get some sustainable technology which you can redeploy into developing some of these things. Earlier IT projects were seen in isolation but now they are seen as part of a more holistic business process in terms of keeping your technology costs as variable as your revenue.
Young: As we see the rise in passive management, people are looking much more to industrialise their process but there are a lot of legacy systems that are actually quite good at processing. And if it is a rudimentary investment process, it can hide the problem.
The other issue around legacy systems is RoI. The longer you’ve had that system and the more deeply embedded it is, the longer and more expensive is the programme to change it. So in RoI terms it could take years, not months. And your operating costs will rise sharply until that system is replaced. The longer they leave it, the higher it costs, so they have to find a way to generate that cost in the short term. A lot of people just keep putting it off.
Funds Europe: From a vendor perspective, how do you get over the notion that the bigger the project is, the more it’s going to cost and the longer it will take to get that return?
Engdal: The RoI question seems to be popping up all over the place, which means that we are seeing lengthier processes between buyer and seller, and sometimes people going back to the drawing board and redrafting their requirements. We have seen it all over Europe. People are afraid to take the wrong decision because things are moving so fast – this is totally understandable and something we appreciate as a vendor. But in the US, there has been a move to cloud solutions. Adoption is slower in Europe but, if we were to do this roundtable in five years’ time, there will be a larger proportion using cloud-based solutions where you pay as you go, based on the number of funds or portfolios you have on this platform. We are also seeing that the lines between us as a vendor and our client are not as sharp anymore, because we sometimes find ourselves actually ‘competing’ with our clients when it comes to combined technology and outsourcing propositions.
Young: Firms have got a much better, clearer idea of where they compete with each other and in the large parts of their operation that are not competitive, they’re much more prepared to take a standard, off-the-shelf solution or to outsource. Large parts of the middle and back office are relatively commoditised, so why have a bespoke legacy system there, when you can share a platform with the industry?
The problem is that there are very few suppliers. The supplier market’s as weak as I’ve known it in terms of end products, so the legacy issue is not easy for fund managers. Ideally they want to outsource, but there isn’t really a shortlist of outsource providers with a standard model. If you put a shortlist together for let’s say a back/middle office replacement system, if you’ve got four or five you’ve done very well. That’s not a healthy situation, really.
Vengayil: The supply market is weak because the demand is not high enough. If you look at our industry compared with banking, where organisations are addressing the second-generation digital needs of their clients and how they deliver it, we are a bit more ‘old school’ and conventional.
That could also be the reason why this industry lives with a whole lot of legacy systems, because we have lived quite comfortably doing things in a conventional way. But one of the reasons a lot of people are speculating now is that there could be a disruption where some of these new technology providers like Google can capture the asset management market by providing technology and bringing us more management expertise.
Funds Europe: Is there still a place for the in-house, bespoke, one-off design system? Or will we not see their like again? If so, does that mean firms will be less creative in their use of IT?
Young: The only reason to build your own system is if you are so unique, you have to build your own. If you’re not unique, you should avoid building your own systems because of the cost. I also think the industry’s changing, the role and power of IT is diminishing and the number of people who are pure IT heads and managers of IT solely is diminishing. There are some banks that don’t have IT departments, but they have a member of the board with IT responsibility. That’s the way the industry needs to go. We are a long way from that in our industry. Other industries are already there... In the future, vendors won’t just shift code software in boxes, there will be a service software hybrid model.
Barringer: It is about defining technology because nowadays the role of the IT department is more about configuration and design than cutting code and building new systems. Now the focus is on presentation and the facilitation of ideas rather than bespoke development.
Doherty: There needs to be a director of information because that is the company’s value, and what a legacy system does, actually, it separates the firm from its information, and that’s its crime.
Vengayil: We still do some development in-house, but not necessarily the core components. That’s because we clearly understand that’s not our core competency; we outsource. But given the complexity of front-to-back models, when it comes to interfacing and those kinds of stuff, we try to do it ourselves.
Doherty: One of the key developments is that IT lost the battle for the company website. Marketing just went out and bought one from a vendor. As the range of technical competencies gets wider, there’s this outsourcing of technical competence. So there is still bespoke code, but it’s not written in-house.
Engdal: I think IT is going to be more about data. There are a lot of firms getting a chief data officer and I guess one of the challenges is whether firms have enough IT knowledge and data knowledge at board level. You can sometimes realise that as you go higher in the company, people don’t really have enough competence and knowledge to take decisions in some of these more strategic investments.
Barringer: It has been the role of IT to own the data for as long as I have been working in the industry. IT lay down the pipes that data flows through, together with the repository and everything that processes. As the role of IT evolves, people have focused on particular areas of functionality and have spun those off into departments of their own, maybe for survival reasons.
Doherty: Putting the ‘I’ back into information technology.
Young: Yes, and taking the ‘T’ away.
Barringer: Very much so.
Vengayil: IT is no longer perceived as purely back-end function. They are very much part of the strategic decision-making process. There are senior-level forums like the IT Project Review Board where they talk about active IT strategy in conjunction with the business strategy. IT is such a big part of the firm’s investment that it has to be combined with strategy.
Doherty: I think there has been an important shift. IT used to exist in isolation, and then we all hopped up and down and said: “We want to be involved in decisions,” but basically what we wanted to do was be able to say “no” to things. The next phase is that we want to contribute to using IT to transform the firm.
Young: If you replace any piece of IT, you have to consider a number of options, including offshoring and outsourcing, so you’re not taking IT and replacing it with IT. You’ve got to look much broader. So you can’t have IT people looking to replace IT. There is a wider array of skills involved and it is a much more complex world now, which is why the IT voice has got quieter.
The biggest problem with legacy is actually replacing it – working out your RoI, getting the political support and embracing change because you cannot swap like for like. Everyone would say legacy is a big issue in the industry, but let’s not pretend it’s an easy thing to solve. The reason people still have legacy systems is because it is so hard to get off them.
Engdal: My prediction would be that you would see fewer systems in production that are categorised as legacy in the future.
More firms will realise that if they are running a legacy platform, the chances are that you can improve profitability and business flexibility by moving to a standard application. They will likely be better off buying the best off-the-shelf products they can find and reap the benefits of shared R&D investment on maintaining their own systems.
We are seeing that to a larger extent than we did just three, four years ago.
©2015 funds europe