A FINOV8 Signals series examining the assumptions, technology architectures and operating models shaping institutional trading.
And if you're thinking of building one, I'd seriously question why.
Having spent more than 25 years designing, selecting, implementing and building institutional trading platforms, including a next-generation EMS myself, I've learnt one thing.
Start with how the desk wants to trade, not which system it wants to buy.
Start with the workflow
Consider a typical fixed-income workflow.
The order sits in the OMS. It's staged to a venue. The trader executes through the venue's desktop, and the fills return to the OMS.
If that's all the connectivity allows, what exactly does an additional EMS contribute?
You may simply need reliable order routing and execution reporting.
Connectivity isn't the same as execution capability
But here's where things get interesting.
Venues may offer sophisticated RFQ protocols, streaming prices and automated execution capabilities.
That doesn't mean every client or EMS can access them through an API.
There's a significant difference between staging an order, instructing a venue to execute it automatically and managing an interactive RFQ yourself.
And access to one doesn't guarantee access to the others.
So before investing in an EMS, establish what your dealers and venues will actually permit you to do.
Not what the EMS brochure promises.
What if you need more connectivity?
Consider a connectivity hub or managed FIX service.
Want to aggregate liquidity?
Consider dedicated data aggregation and analytics capabilities.
Need sophisticated multi-market execution workflows?
Then evaluate whether an EMS genuinely delivers them through the connections available to you.
Beyond the execution record
There's another question.
If a trader sees an axe, indicative price or streaming quote and decides to act, does the technology preserve that information and its relationship to the subsequent trading decision?
Or does it simply display the market data, route an order and record the fill?
There's a difference between showing traders information and helping them make, explain and evaluate decisions.
As AI reshapes trading, shouldn't we be questioning whether today's technology architecture can support tomorrow's trading operating model?
My point isn't that nobody needs an EMS.
It's that an EMS shouldn't be the default answer to every trading technology requirement.
Build or buy the capabilities that support how your desk actually trades.
Sometimes that's an EMS.
Sometimes it's an OMS, a connectivity hub, a data hub, and specialist tools.
And sometimes the smartest technology decision is not buying another platform at all.