CMU Liquidnet - Research

Overview

In this phase, we collected information about our users and "dark-pool institutional trading", which Liquidnet specializes in. Our work focused on understanding the trader (our user) and their technology (the Liquidnet 5 product).

Problem

We needed to quickly onboard ourselves in the financial world to study traders and Liquidnet's product.

Challenge

Our team had to understand traders well enough to design for them according to our project brief.

Research Journey

The team began learning about the average trader's lifestyle and technology usage. Once we had a better understanding of traders, we studied Liquidnet's software, Liquidnet 5. Finally, we examined the trader's current usage of Liquidnet 5 features.

Users

Product

Users + Product

Users

Many trader types emerged as we studied trader's priorities and metrics for success, with varying Liquidnet usage.

Product

We examined Liquidnet 5 to understand its current use. Different styles of usage emerged across user types.

User + Product

The user types found earlier matched different interactions with Liquidnet's front-end and back-end software.

Users

Traders are core to Liquidnet's success - they execute trades, they analyze research, and they develop algorithms, all benefitting Liquidnet. Understanding trader’s goals and needs tells us where Liquidnet should target traders. We'll walk through the research we did to understand the different stakeholders in dark pool trading.

Jeevan the Equity Trader

GOALS

As an equity trader Jeevan's goals are to improve his performance in his trades, and to create innovative trading strategies.

NEEDS

Jeevan needs to be confident in his trades even if they fail, and needs to grasp high-quality data and algorithms fit his trading strategy.

Jed the Research Analyst

GOALS

Jed wants to provide useful insights on 'good' trading decisions, thus becoming a trusted resource for traders.

NEEDS

Jed needs to be aware of market direction, and must establish a rapport with traders so that his research is taken seriously.

Jane the Portfolio Manager

GOALS

Jane wants to achieve 'good' portfolio performance with low risk overall, as well as avoid information leakage at all costs.

NEEDS

Jane needs to be confident in her opinions, even if her team disagrees. She must also understand clients’ need to build portfolios.

How did we study the user?

RESEARCH METHODS

Only a few niche traders use Liquidnet or take part in dark-pool trading. As such, our team used generative research to understand the elusive dark-pool traders and proxy users that experience trading-esque work. Through these activities, we learned how traders think and act on the job.

Observation 1

Dark pools stand out

Our team played poker and a stock management game
Both games emulate the stress and timing of real trading
Poker is a commonly used game to train young traders

Observation 2

A trader's day and peers

Preliminary research gave us an idea of a hypothetical trader
Interviewees discussed unique trader types and teams
We learned that trading teams don't have standard sizes or roles

Observation 1

Trades are an 'action stream'

Institutional trades 'stream in' and change throughout the day
Change in news, pricing, and market may influence the stream
Traders are prepared to change actions mid-stream

Observation 2

The three phases of a trader's day

Open: Traders input their day's trades, hectic time
Mid: Most calm, ideal for smaller mid-stream actions
Close: Pressure rises to finish trades and report work

Observation 1

Poker is analogous to trading

Poker is often training for incoming traders
The pace and mental load of poker mimics trading
In-game planning mimics the portfolio management

Observation 2

Poker players struggle like traders

Players need to recover and gain emotional balance
Most time is spent disengaged from the game itself
More engagement can build positive feelings on wins

Observation 1

Gaming can emulate trading

Our team played poker and a stock management game
Both games emulate the stress and timing of real trading
We played competitively to mimic trader competition

Observation 2

Some lessons learned

In poker, strategy management is mentally challenging
Both games require strategy to succeed long-term
Even strategic choices can 'go bad' due to luck

SYNTHESIS

After gathering a considerable amount of generative data on traders, our team synthesized findings by using the following techniques. These maps and graphs helped us collect insights and communicate our research findings with ease to our clients.

Insight 1

Collecting insights

Traders use in-house advisors to develop trading strategy
External services and brokers provide 'unique' information
This information allows traders to improve their strategy

Insight 2

Liquidnet serves overlapping roles

Liquidnet serves as a broker and external analytics platform
Relationship managers sustain loyalty and trust with the trader
Proprietary data brings traders back to the Liquidnet product

Insight 1

Actions repeat

Pre-trade is determined by the client and/or portfolio
Trades are untouched unless an anomaly occurs
Mid-day anomalies are handled as a team

Insight 2

Emotions dip mid-day

The highlights of the day are the open and close
Mid-day is mundane, and is mostly updating prior trades
Open and close are both exciting and stressful

Insight 1

Screen real estate

Traders have 2-6 monitor screens, all used in the day
Typically monitors are dedicated to a type of application
Liquidnet front-end is actually infrequently used

Insight 2

Front-end distractions

Liquidnet usually appears on-screen as pop-ups
These pop-ups are for immediate action or alerts
The front-end is often seen as a distraction to traders

Insight 1

What does Liquidnet do for traders?

Shows unique data through Discovery analytics
Builds trust with traders via relationship managers
Provides the most liquidity compared to the competiton

Insight 2

A trader needs...

To trust the information provided from software and peers
Unique or proprietary data to inform trading strategy
A robust risk management strategy for any situation

Product

With a better understanding of traders' workflow, our team studied the Liquidnet software. We learned about the target customers and the features marketed towards them. Then, we compared the target to the actual users of Liquidnet. Below are some of the generalized findings about the product.

Customers

An entire company or a single trader may purchase Liquidnet for their trading teams. When interacting wtih the customer, Liquidnet must keep the customer’s portfolio in mind.

Features

Liquidnet has acquired analytics technology to build upon its order management and execution software. They want their customers to do more than execute trades on Liquidnet.

Users

Traders often avoid Liquidnet's front-end product and features. They want Liquidnet's features to be easier to find and use, and will avoid features with ambiguous value to them.

How did we study the product?

RESEARCH METHODS

Parallel to our generative research on users, our team gathered preliminary research about Liquidnet and their competitor's offerings. This involved a competitive analysis, as well as studying Liquidnet's sequence of operations and adherence to design heuristics.

Liquidnet's advantage

There is a historical preference for Liquidnet, the first dark-pool
BIDS and ITG are cheaper software, but give less data to traders
Liquidnet has the best liquidity as the first dark-pool

Liquidnet's disadvantage

Competitors ask for a lower commission (cheaper trades)
Liquidnet's software is significantly more expensive
Competitors excel at algorithms and quick matches

Mapping workflows

We mapped each sequence of actions in workflow
Many workflows interact or connect throughout a day
E.g. passive match flow contains the active match flow

Patterns in workflows

Many workflows use the same features in slightly different ways
Mapping showed which workflows were most common and important
Active and passive trade workflows are most routine for all traders

Industry standards

Fintech software often emulates spreadsheets
Common colors are bright red, green and yellow
Industry standards haven't changed in the last decade

Design standards

Many visual heuristics were not met by Liquidnet 5
Color and iconography will make their software more intuitive
Visual hierarchy must guide trader through workflow

SYNTHESIS

Through the methods just shown and from our onsite interviews, our team found the following insights. Overall, many traders have cognitive and emotional overload from their job and their software. Traders avoid new features and experience difficult UI interactions on a daily basis.

Workflow overload

Traders typically master a workflow and stick to it
Connected workflows are not highlighted visually in Liquidnet
Thus, traders are reluctant to learn new workflows and features

Finding features

Due to of visual clutter, new features are hard to find
No in-app help means traders must call sales for help
Hidden features in Liquidnet's blotter are unused and unseen

Juggling workflows

Traders have to juggle many workflows and data
Traders have to shift workflows based on market
Allowing the trader to change action should be easy

Finding liquidity

Traders inherently will use competing applications
Liquidity offerings can change trader's app usage
Liquidnet needs to convince traders to stay with them

Information overload

Poor visual design and hierarchy can confuse traders
If traders can't execute trades, they'll switch to a competitor
Showing liquidity opportunities with data keeps trades

Scarcity of space

Screen real estate is extremely valuable and limited
What isn't seen isn't used or known to traders
When and what info is shown can convince trader to act

Users + Product

After user and product research, we understood where Liquidnet stood with its users. The trader's needs that were not met and the usability issues in Liquidnet 5 showed our client where they could improve.

Different workflows

Liquidnet's users have vastly different workflows and trading styles. The product should be tailored to these user types instead of forcing all users to trade the same way.

Analytics

Data and information is at the heart of trading. Traders want proprietary information from Liquidnet to decide if they or their team should trade on Liquidnet or at all.

Usability

Liquidnet 5 has many hidden features that do not match current design heuristics or usability standards. This makes their product and new features less friendly to traders.

How do users currently interact with the product?

SYNTHESIS

To collect our user and product related findings in one location, we developed a service blueprint. Each swimlane designated a different user-type, all of which use Liquidnet through unique workflows. Our clients reviewed and confirmed our understanding of the design space.

Overall trends

Six user types were apparent from our research
Algo usage was a big difference between user types
Three phases of the day were similar across all types

A day's events

Phases of the day indicate actions executed
Open and close of the day are most stressful
Mid-day monitoring looks different for all user-types

Liquidnet's dilemma

Liquidnet struggles to connect traders to their front-end
Traders want what Liquidnet provides but cannot find it
Traders are creatures of habit and avoid change

Opportunity through alerts

Current successful interactions are through pop-up alerts
Liquidnet could use alerts to connect traders to front-end
Liquidnet is designing an 'alert stream' in the front-end

MOVING FORWARD

After meeting with our clients, our team narrowed down on our problem scope to focus on developing alert design guidelines and prototypes. We began to exploring how to turn our insights into the backbone of our design to ensure the user was central to our work at all times.

What we did

We gathered our product model, stakeholder map, affinity diagrams, and user interview notes together and posted them on a board to look through all at once. Then, we took some time to create new insights and re-emphasized the most important user needs.

What we learned

Quality of Professional Network: I want my leads to come from an exclusive source.
Data Analytics & Visualization: I need to be able to customize my work station.
Workflow: To do research, I need to be on multiple platforms which is taxing.

What we did

We generated questions that tried to answer the problems we found through walking the wall. The main goal here was to try to be specific enough to answer the user needs we identified, but also broad enough to create a whole design solution around it.

Example questions

How can we keep traders mentally alert to optimize their productivity?
How can traders be willing to revise their routines without sacrificing efficiency?
How can we design something visually articulate but familiar to traders?
How can we help traders better tackle work-life balance/avoid burnout?

What we did

We made assumptions based on our users’ needs and reversed each one. Then, we used this reversal to brainstorm a new service or feature offering. This set up the stage for preliminary solution exploration.

What we learned

Quality of Professional Network: I want my leads to come from an exclusive source.
Data Analytics & Visualization: I need to be able to customize my work station.
Workflow: To do research, I need to be on multiple platforms which is taxing.

What we did

We created 15 "How Might We" statements to sum up our most important problem areas and user needs. We plotted these onto a matrix and considered Liquidnet’s feasibility of building the potential product and also how much value this product could bring to our client. We ended up choosing 5 major questions we wanted to solve below.

How might we...

... get traders to participate in making the product better?
... keep traders mentally alert to optimize their productivity?
... automate the right analytics fields to visualize data better?
... reorient traders with where they left off at work after a hiatus?
... get traders to avoid snowballing when they’re at a disadvantage?

Takeaways

Overall, this phase of research was able to turn our team from financial novices to traders-in-training. We deeply understood both the trader's needs and our client's goals in designing for the next generation of institutional traders. Our charge was now to utilize our research insights in our design phase to develop front-end style prototypes.

Progress

Our team narrowed our project focus based on our client's feedback and the research gathered. Future phases will focus on designing for Liquidnet's front-end and alerts system.

Learnings

Based on the data provided by our clients and interviews onsite, we discovered that front-end interactions and Liquidnet's new 'alert stream' may be the best avenue to design for.

Next step in our journey: