How can data + design protect organizations from network crashes and security threats?

A. IBM Security

Opportunity: Co-lead design research for IBM's division of 75 cyber security software products. Work with Fortune 500 clients to understand their security professionals’ workflows and pain points. Support and collaborate with multiple design and product teams to create human-centered products that prevent cyber security attacks against clients.

My impact: Established design research best practices, resources, and artifacts for IBM’s Security division. Helped division executives and product team members gain empathy with communities of use. Collaborated with multiple product teams to translate actionable insights into implemented product designs.

Process

Key Research Questions

Cyberattacks are increasingly sophisticated and costly. Organized crime rings orchestrate 80% of all cybersecurity attacks (United Nations Office on Drugs and Crime). These attacks collectively cost organizations $400 billion a year (Fortune). To help security teams protect their organizations and enterprises, we sought to understand their enablers and inhibitors.

Questions about communities of use

• Who uses IBM Security products and what do they do? 

• What do they value and what motivates them? What challenges and frustrates them?

• How might we support communities in preventing, identifying, and mitigating cybersecurity threats?

I created portfolio-wide outputs including personas, journey maps, and stakeholder and ecosystem maps that were informed by qualitative research with communities of use and synthesis techniques such as affinity mapping and spectra analysis. By socializing these artifacts with division executives and product teams and aligning them with existing product development processes, I helped to increase empathy for and establish a baseline focus on end-users across all product teams in the division.

Note: I can’t show much of this work online due to confidentiality. Happy to share my human-centered, collaborative process in person or on a call though!

My impact

B. IBM Predictive Analytics

Opportunity: Help IT professionals at data centers prevent application problems on enterprise networks. Redesign the product interface to present information generated by complex analytics algorithms to communities of use in understandable, visual, and actionable ways.

Impact: Most recent product release reflects implemented designs informed by user insights and has received positive feedback. Distributed 3-country product team that previously did not work with designers has embraced and actively practices design thinking.

Understand the Space

Enterprise organizations like banks and telecommunications companies must run and maintain hundreds of business-critical applications. These applications range in function from payroll management to customer-facing websites. A single failure on a business-critical application can cost hundreds of thousands of dollars. As a result, many organizations employ large teams to monitor their systems and quickly resolve application errors. However, traditional network monitoring systems alert IT professionals after a problem has already occurred.  

By applying data analytics to examine network activity, IBM Predictive Insights detects and preemptively alerts users about abnormal application patterns. Given advanced warning, users are able to prevent crashes and problems before they occur.

Unfortunately, the original Predictive Insights interface was unusable. Our communities of use did not know what information was conveyed or what resulting actions they should take. As the first designers to work with the product team, we challenged ourselves and the product team to consider:

• How might our distributed, 3-country product team remotely practice design thinking?

• How might we empower IT professionals to accurately identify and act on application errors?

• How might we help network operators embrace an error-prevention mindset?

Identify problems and existing solutions

To understand the target users’ workflows and pain points, I coordinated and conducted 13 contextual inquiries at two data centers and multiple generative interviews and usability tests. Next, I facilitated empathy and affinity mapping exercises to synthesize the data.

Insights from research:

• Users prioritize application issues based on potential business impact. Issues with great potential business impact (ex: website crash of a bank’s customer-facing website) cost a company the most money.

• Users do not trust predictive alerts. Because users are accustomed to system alarms that signal actively occurring problems, they doubt the validity of early warnings.

• Anomalies often indicate system errors. Users repeatedly compare current application metric values with historical metric values along cyclical time points (ex: an hour ago, the same time last week). If the metric values don’t match up, something might be wrong.

• Too much noise, not enough guidance. Network operators often use more than 30 applications to do their jobs. They spend most of their time being bombarded by and deciphering opaque false alarms.

Through journey maps, personas, and other visual design research artifacts, I communicated user insights and design recommendations to my product team, executives, and technical sales division.

Facilitate Participatory and Iterative Design

To align our whole product team, I organized and facilitated a series of design thinking workshops for our product managers, development lead, architect, technical sales leads, and portfolio executives. Through collaborative ideation, prioritization, and business model canvas exercises, we leveraged our collective experience to brainstorm new ideas.

Implementation and Impact

After refining our designs based on user feedback, we collaborated with our development team to implement them. The most recent product release allows users to understand and act on alerts much more effectively. Benchmark user test task completion rates skyrocketed from under 25% with the old product interface to 100% with the redesigned interface.

Implemented changes that reflect design research insights:

1. Actionable summary information, including potential impact and historical trends, is prominently displayed.

2. Improved navigation and visualizations empower users to more quickly isolate potential issues.

3. Discrepancies between expected and actual metric values are highlighted, helping users know where to take action.