Alexei Miller - DataArt - EVP - Custom Development is Alive and Well - wAT - Dec 18 2013

Because the market doesn't stand still. Because there's more and more pressure from investors, from regulators, from clients, from internal users, and there's this pressure on IT to do more, pay less, and figure this all out. It's not business problems that you have. All these competing pressures, and oftentimes the answer is managed service. Oftentimes the answer is pushing the responsibility for running a certain piece of operation outside of the shop. But in many situations, there's no choice but to develop a custom piece of software. For example, if the firm is engaged in sophisticated credit trading strategies, you have to develop, and that strategy is supported by proprietary research process. There's no choice but to develop a custom piece of software if only for your workflow. You have to do that, and then you expand on that idea. If you are in the derivative space, if you are trading with multiple counterparties, in order to set all of that up in a manner that will work uniquely for your firm, there's no question: custom is the answer.

I think trends are different on the buy side and the sell side. Let me comment briefly on each. On the sell side, it is true folks have developed massive custom software systems internally for many, many years. Then about seven, eight, nine years ago, this massive trend to outsource started. But they started it by outsourcing work that was non-mission critical: development of their HR system, their CRM systems, and whatnot. They kept the juicy stuff—trading, risk, compliance—in-house. So what happened? What's happening on the sell side? They're bringing portions of the corporate technology back in-house for a number of reasons, but at the same time looking for additional expertise to support this dynamic, to support the requirement to do more with less and faster. So a lot of more sophisticated stuff is being outsourced while the simpler stuff is occasionally brought in-house. That's on the sell side. On the buy side, it's different because buy side never outsourced much. So the buy side, in the last two, three, five years maybe, is just now experimenting with this. Because buy side firms never have massive internal operations. For them, outsourcing corporate technology is not that big of a deal; it doesn't really drive any results. So buy side in many respects is jumping straight to the meaty, juicy stuff, sort of skipping that phase about outsourcing that sell side firms went through. So they definitely retain a lot of capability to do things internally still, but I'm seeing a lot more openness to engage with an outside party.

In my experience, not too much is being done on the extreme front end of the house, things like trading, trade execution, trade allocation systems. People still prefer to buy. But as we move down to the middle office and the back office, everything from reference data, custom analytics, custom investor reporting, multi-broker functionality, especially if you have a multi-strategy operation, will require some bits and pieces of custom code. And there's really no limit today with the sophistication of third-party providers. There's really no limit as to what you can or cannot outsource. It's about capability and about time to market.

They're not mutually exclusive at all. Managed service is a new way of thinking of owning versus paying someone else to run it for you. In both scenarios, you can go with custom or with a package software. So those are parallel arguments. Custom software is all about feeling that your shop is unique enough or you know how to operate it in an efficient enough manner that if you invest a little bit of money into supporting it with custom technology, you can run your ship more efficiently. That's the whole conversation. You can then take that piece of software that you've developed in a custom fashion and run it as if it was a managed service. So there's no one or the other; you can combine the best of both approaches.

Well, outsourcing used to be about cost, no secret there. Then it became about capability: can you do what I cannot do? That kind of conversation. Increasingly in the last few years, it's about time to market. Can you do it cheap? Can you do it better than I can? But can you also do it faster than everybody else? Right. So it's a lot more about doing it faster. It's about time to market. Therefore, having prior knowledge is very important. Because if you're a smart group of people, you'll learn everything. I mean, you can take a good developer and teach him algorithmic trading; that's not that big of a deal. The trouble is it's going to take a year. If it takes a year, your time to market argument goes out the window. So if outsourcing decisions or custom development decisions are being made on the premise of time to market, you better partner with someone who understands the market. That is likely to cost a little bit more. So you are engaged in the classic risk-reward calculation. Good news is Wall Street is used to that kind of conversation.

Thank you.