Automated system(s) for order/portfolio/investment management

It’s challenging for investment managers to automate processes and workflows. There are many reasons why, some related to system(s) and some related to how the industry functions. We’ll go into detail in this post about the challenges, which are more easily overcome than others, and concretely how to automate more than today.

Four aspects make automation of processes and workflows a challenge for investment managers:

1. There is a lot to automate. According to a report by Daymi, the number of daily tasks performed by front and middle office teams can be in the thousands. Where to begin? How do we determine what is worthy of automation?
2. Your business is unique. Your exact investment style or approach is, by definition, unique. Likely, your combination of counterparties and systems isn’t used by any other firm on the planet. How would a templated off-the-shelf automation solution solve this?
3. Machines struggle with “sanity checking”. They might get there as AI becomes better, but at the moment, this is a skill that humans outperform software, especially in this industry where data sets to train on are scarce.
4. Workflows typically span multiple systems. This is the biggest obstacle for the majority of automation initiatives. To get data out of several systems, then being unsure if the data is controlled or complete, and continue to operate on it, import it into another system, etc. It’s a very fragile process.

In our experience, the best automated investment platform is a much broader platform covering the workflows to be automated. Automation becomes much more difficult if you rely on best-of-breed order management software or portfolio management software.

With one Front-to-Mid/Back platform, automation just becomes features within the software, without having to control and move data between systems and sources.

The bad news: you can’t “automate on top” of existing systems

AlphaFMC have looked into investment in automation, and found 5 areas of investment (you don’t need to know what all of these mean to continue to read and understand the takeaways in this article):

1. APIs
2. Business Process Management (BPM)
3. Low Code Applications
4. Robotic Process Automation (RPA)
5. Workflow management

Our interpretation of this list, is that these are all attempts to automate on top of existing systems. I.e. an attempt to add a layer between the system and humans using the systems (or “hack” the user interface). But the problem is that the systems aren’t built to support that. It‘s no surprise that AlphaFMC also found that most initiatives in automation fail to provide positive ROI.

The good news: it’s possible to achieve highly automated workflows

Here’s the thing: there is no need to “do” operations anymore. Software can actually “do” operations by now, including tasks such as:

Automation vs Automatic

Below is the interpretation used throughout this article:

• Automatic investment management software - when a process runs entirely autonomously, without any interaction of a person, even when issues happen
• Automation (a greyscale) - when a process runs according to a set instruction, and as soon as something unexpected happens, a user is alerted to investigate and resolve the issue

A real-world story from our CEO and founder, Kristoffer Fürst

When I was on the buy side, we managed a $1.5bn macro portfolio. For this single portfolio, a colleague of mine had a process that started at 4 pm each day. She sat on the desk next to me, so sometimes I’d assist if I had the time.

Here’s how the next 3 hours would unfold:

• She waited for 15min for a system task to complete
• She moved some files from an email into a file area
• She ran an Excel macro
• She then waited another 10min for another system task
• She reviewed a PDF (“sanity check”)
• ... and so on ... until 7 pm

Clearly, this process wasn’t great. But:

• It had nothing to do with our operating model
• It had nothing to do with the quality of our team
• It had nothing to do with our investment style

It had everything to do with subpar software.

We should’ve had more semi-automatic processes:

• Reconciliations
• Data checks: prices, instrument parameters, NAV tolerance, etc
• Importing data: external trades, accruals, subs/reds, cash movements (subs/reds, corporate actions, coupons, etc) and more
• Exporting data: trade files, reporting, etc

... And we shouldn’t have been waiting for a quarter of an hour for a system task.

Don’t get me wrong, I’m not saying everything should (or can) be automated. The entire operations role isn’t threatened by this automation (nor by AI, at least not in the next decade).

What I’m saying is that it’s possible to remove mundane, repetitive tasks from expensive team members (who, by the way, get bored with that type of task) and hand them off to software. Elevate your team to more rewarding tasks that software (and AI) can’t do: break resolution, process development and counterparty coordination.

So what are Exception-based workflows?

In exception-based workflows, the software/AI takes over all “check that...” tasks and anything that resides as a calendar reminder. It automates those tasks as much as possible. As soon as something doesn’t go right (deviates from the “Happy path”), an issue is flagged for a user to resolve.

Ideally, at any point in time, you can see a status dashboard within your automated investment software. Here’s a visual analogy:

Examples of exception-based workflows

• Automated order management workflows can include minimising the number of clicks to route orders and bulk actions
• An automated portfolio management system can include leader-follower (multiple SMAs mimicking one master fund), notifications on portfolio movements, 1-click rebalance, and in general just helping to avoid Excel for anything (look-through, benchmark comparison, portfolio monitoring, etc)
• Within operations, examples are reconciliations, calculations currently performed in Excel, any manual review (e.g. comparing a PDF and a system), etc.

How many firms do we see that have exception-based workflows at scale? Less than 25%, according to our proprietary research.

How does the switch affect your team?

We all know exception-based workflows are the future. Here’s the effect automated investment tools have on the operations role:

1. It becomes much more of an oversight role (control room analogy)
2. Issues aren’t disappearing. BAU (business as usual) in this industry is that a lot goes wrong, every day. With checks and data management delegated to software, your team can focus on resolving issues as opposed to finding them
3. It becomes a design role. Designing processes and setting them up in systems, including controlling notifications and alerts

Advise on how to evaluate and compare vendors

We strongly believe that workflow efficiency should be given a significant emphasis in system selections. Most RFI/RFPs focus on features (yes/no), but forget the “how”, when evaluating automated investment platforms.

A client of Limina saved 50% on their entire operational cost when changing from another system to us. And it’s not a one-off case study, we have more cases that are in the same ballpark. Set Limina aside for a moment and just reflect on that number. That time is worth much more than the cost of the system.

Exception-based workflows really do provide massive efficiency improvements, in the real world, here and now. These benefits aren’t uncovered with a simple yes/no answer checklist.

Tactical advice:

• Use an RFI checklist to shortlist automated investment systems (functional coverage, security questions, etc)
• Then use proof-of-value, demos and business cases to pick the best providers among the shortlisted

The difference between providers might look like <5% on paper. But in reality, it can be more than 2x! As the referenced case study highlights.