Hacking for Defense - Department of Defense

Project Details

Design Sprint

 

Hacking for Defense seeks to accelerate the speed at which national security organizations solve their hardest problems. As part of a 90-day collaboration between the DoD and BMNT, I worked with government officials from the Joint-Improvised Defense Agency (JIDA) to create a solution for Analysts' challenges using the Lean Startup methodology from Steve Blank. Analysts play a critical role in keeping the United States and the rest of the world safe from threats. The results from our design sprint were presented to the Department of Defense senior leadership and will be used to inform an Integrated Regional Threat Model. 

 

Define the Problem

 

From interviews with stakeholders and Analysts I identified 3 pain points in the collaboration process : Discovery, Analysis and Presentation.

When a commander in the field needs additional tactical or strategic information, he routes the question to an Analyst at JIDA. For example: Identify the primary ignition trigger for vBIDs used in Iraq during 2016 near Kirkuk? An Analyst sorts through many databases and uses different software to prepare a report (Discovery). Sometimes they spend more time cleaning datasets or transferring data from different software programs. By the time they start the analysis process, they only have a few hours to synthesize the information and understand the patterns (Analysis). Given the time crunch to deliver reports, Analysts don't spend much time referencing the sources--a "must" for commanders who want reliable sources so they can make time - sensitive decisions (Presentation). 

 
 

I used a business model canvas (Mission Model Canvas) and iterated on it with feedback from stakeholders to define the problem more clearly before leading the team into the prototyping phase. After several assessments, we created a clear value proposition and identified the beneficiaries, J2 Analyst (Generalists) of our solution. 

 

Refine Hypothesis

 

The problem statement we started with aim to address all 3 pain points. As we progressed we moved away from a strictly data visualization-focused effort towards creating solutions  that improve the way data is utilized and manipulated.  

Early Problem Statement: The variety of data and tools undermine the Analyst's ability to generate insights that can be trusted, verified and replicated

New Problem Statement: The majority of analyst's time is not spent doing analysis.

Critical Hypothesis: If analysts have the ability to customize their workspace using unbundled basic functions, they will create their own custom tools that reduce the amount of time spent doing data "janitorial" work.

Diverge

Learning from Experts & Analogous Applications

We expanded our scope by interviewing Stanford PhD students to understand the technical and process challenges associated with data visualization and analysis. We also reviewed 12 data driven consumer products to understand the landscape of efficient analysis and presentation of data. The best tools enable different individuals to approach challenges in the way that best matches their thought process and preferences. For example, Yelp enables customers to sort through a variety of restaurants to find the choice that matches their needs. If Analysts can filter options and repackage data sets, intelligence organizations should be able to increase the quality and rate of insight production and thus identifying more threats too. 

 
 

After drawing inspiration from other experts and industries, we revisited our problem statement and revised our Mission Model Canvas several times before diving into concepts and sketching. 

 

Experiment & Prototype

 

We set out to develop the most important features that will support the future of intelligence analysis through an iterative process of brainstorming, drawing and then voting on the best concepts. Our teams proposed 3 ideas: a landing page, a survey and a collaboration workspace. The first 2 aimed to get buy-in from stakeholders while the last is a Graphical User Interface that allows real-time manipulation of data. We saw collaboration as being key to help all Analysts stay up to date on the most effective tools. 

 
 

Personalization Options 

 "Don't force them into a box, they need to maneuver"

Every analyst has different preferences. In this first round, we explored a central location where data and tools can be dynamically manipulated. 

 
 

We drilled down on the interaction by exploring a drag and drop that can help Analysts filter and merge data using a visual layout (Lego Bricks).

 
 

In another interview, we validated the solutions with Analysts to decide what to keep or refine. The landing page and survey quickly helped us validate an idea similar to Product Hunt, where Analysts can vote what tools they prefer. Though it was an innovative approach,  we discovered the workspace collaboration resonated more with them in terms of efficiency and ease of use. 

Evaluate

We presented our final MVP to the Defense Innovation Unit Experimental (DUIx) and active military officers studying at Stanford to get an additional layer of feedback. We walked them through the workspace where Analysts will be able to construct their own tools using a visual interface. Our solutions were well received. The audience asked us to also think about transition plans since a lot of the software used by Analysts are in legacy systems that can't be transformed overnight. 

 
 

Reflection: Fail Fast

We spent less time prototyping and invested more time refining the hypothesis. This experience taught me the importance of thinking before creating, something that's tough to fight when as a designer I want to create. Someone suggested to just do a survey, which I first thought wasn't really valuable because it was not a potential solution. But I realized during the validation interviews with the Analysts, a survey is more helpful to validate an MVP, than a prototype, because it saves time from building something that may not even be in line with what our audience needs. 

 

 

Other Projects

Backbone Labs

Helping people become aware of their posture to prevent injuries and back pain

Stroll Health

Empowering people to make better health care decisions by providing price transparency for complex procedures

Toastmasters

Reinventing a learning experience for people seeking to become better communicators and leaders

Vint

Reducing the friction with technology to help admin staff create better relationships with customers at gyms

Cisco

Migrating training content and redesigning new platform to help 1000+ Partner companies discover Cisco products