Trumid: Design as differentiator

How Trumid is building the interface for the next generation

Jason Quinn, chief product officer and global head of sales at Trumid, speaks to Dan Barnes about why user interface design matters more than ever in electronic fixed income trading, how intelligence is replacing clicks, and what traders can expect from the platform over the next twelve months.

The DESK (TD): As trading protocols have multiplied, how has that complexity changed the demands on interface design?

Jason Quinn (JQ): That hits nail on the head. When you put every possible way to trade on one platform, we think of it as a single-platform EMS, it sounds great in theory. But the cognitive load is significant, and as liquidity grows across each of those pools, it becomes more and more important for clients to understand what’s happening and not miss opportunities.

The answer to that is intelligence. If you’re faced with cognitive load, I can either ask you to figure it out yourself, or I can try to figure it out for you and serve it up on a platter. There are two ways to do that: automation – where you just tell the system what you’re trying to do and it’s designed to ensure that you don’t miss opportunities – and proactive delivery of what we think your optimal opportunities are, so you don’t have to sift through everything yourself.

That really means using machine learning and AI to improve user experience, because there’s a limit to how much you can reduce clicks and improve design on its own. At some point, you have to do it with intelligence instead.

TD: You’ve used the word intelligence deliberately without saying artificial in front of it. What’s the distinction you’re drawing?

JQ: When people say ‘AI’ right now, they tend to mean powered by a large language model. But there’s also machine learning, pricing analytics, and other tools that deliver insights to clients without relying on an LLM. And there’s a very good reason for that: LLMs are non-deterministic. I don’t really want to be delivering non-deterministic intelligence when clients are making trading decisions on millions and millions of risk. I can still deliver intelligence, but in a deterministic way, without leveraging an LLM. That’s intentional.

TD: How do design and user experience shape Trumid’s thinking on function?

JQ: The current version of Trumid’s platform is about to turn ten years old. From the very beginning, we thought there was an opportunity to differentiate with design and user experience (UX). Some of that came from observing competitive platforms and where they were relative to platforms we were using in our everyday lives.

If you rewind the clock to early 2016, everyone had a smartphone in their hand. People were using modern apps. Things were gravitating from web pages to really intuitive apps with great design, and it always boggled our mind that you would go from that experience in your personal life to using the things you were using in your working life. And that’s not just a shot at the established platforms. When I was on the sell side, the pace of change and evolution in the systems we were using to do our jobs was also way behind the standard you would expect.

So as Trumid we said: we’re the newcomer, we’re trying to be scrappy here, we think we can differentiate with design. And I have a background in it — I actually went to school for industrial design before ending up in trading, so this was always part of my DNA.

TD: So is design a differentiator for Trumid?

JQ: Yes, and I’d like to think we were responsible for some of that being a competitive thing amongst platforms now. Everything we use in our personal lives continues to improve in design and usability, so the bar keeps rising. More importantly, clients legitimately care now. When electronic trading was fifteen percent of what they were doing, they didn’t really care that a system was a bit painful to use. When it’s half of everything they do, they care a lot more.

So do we. I was in a twenty-minute debate yesterday about the specific icon being used for a new enhancement going out. Because those little details matter enormously. There is no broad stroke you can brush across to get design and UX right. It’s every small detail that adds up to the overall experience, and if you don’t care about every micro detail, the overall mission falls flat.

TD: How do you elicit feedback from clients about what they actually want from an interface?

JQ: We have to engage directly. This is such a niche market, with thousands of users, not millions, so we can’t just A/B test our way to the answer the way a consumer app can.

Some clients are very opinionated, and others don’t care very much at all. It’s a bit like being a car person or not. I can see a headlight in the middle of the night and tell you what car it is, and what year it came out. Other people just want to know if it has Apple CarPlay. It’s true in our world too. When you find the clients who really care, the ones who notice you changed a button overnight and tell you exactly what they think about it, you talk to them a lot, because they give you really good feedback.

What they care about is clicks, intuitiveness, the presentation of data, whether interactions work together logically. On our platform, we have multiple workflows all in one system. That sounds great, but it’s really hard to pull off without confusing everyone. There are years of iteration there, both internally and in collaboration with clients.

TD: How do you capture the positive signal, and not just complaints, which tend to be more forthcoming?

JQ: We have discipline around this. We look for patterns. If one person says something and five others say something similar, maybe we missed something. And for the things nobody mentions, we observe how people are actually using the system. If a feature we thought was valuable isn’t being used the way we expected, that’s a signal. If usage of something is going viral and nobody’s said a word about it, we can see that in the metrics too.

Product development is about taking in the feedback, collaborating, studying the data and measuring things. The version of Trumid you use next month needs to be better than the version you’re using today. That’s the bottom line.

TD: What can traders expect from Trumid’s interface over the next six to twelve months?

JQ: There is a real drive toward making it very clear on our platform when you are using an intelligence feature. We have a suite of them now, and we’ll continue to invest there. The UI components and design patterns will be familiar and obvious, so it’s immediately clear when you’re working with an intelligence feature rather than a standard function.

And it’s not just for traders. We consider sell-side salespeople an important and relatively underserved cohort of our client base, and we invest as much in our sell-side sales application – Trumid Link – as we do in the trading app. The intelligence features arriving over the next twelve months will be available for both trading and sales. In some cases, we’ll actually be rolling things out to sales before they make their way over to the trading side. We’ll be looking for a lot of feedback from both cohorts, because the fundamental question we’re solving for is: how do you manage cognitive load? The answer, increasingly, is intelligence features.