When Smaller Is Better: A New Way to Modernize IT (Contributed)

A growing number of governments are adopting a Silicon Valley, user-centered design methodology for creating better software programs, known as Minimum Viable Product. The results include huge time savings.

by / April 23, 2019
Shutterstock

Government digital services are complex and overwhelming to the program managers and procurement officers charged with modernizing software. Whether it is a citizen-facing portal; legacy software used to evaluate, approve and manage requests; reporting requirements to maintain compliance; or the decades-old databases that support each of the aforementioned services, the existing web of systems is overwhelming and difficult to navigate.

Efforts to modernize these digital services often involve outside IT contractors that conduct “requirements gathering” to resolve the problem. IT agencies can spend millions of dollars for a vendor to go away and “just implement” all the requirements and then later be held accountable if something goes wrong. Too often, these strategies end in failure.

To address this ongoing problem, a growing movement has emerged in federal, state and local governments to make better, modern software, and re-envision digital services as a promise to deliver on a mission that serves users. The U.S. Digital Service and 18F are federal organizations inside government dedicated to bringing agile software development, human-centered design and acquisitions reform to agencies that need or are ready to adopt modern ways of rebuilding digital services. Code for America, Agile Government Leadership and the Digital Services Coalition are organizations supporting this from outside the government, in the nonprofit and private sectors.

A key ingredient to this effort’s success is the Minimum Viable Product (often just called “MVP”), which has its roots in Silicon Valley-type software practices. In the private sector, MVP is a way of building the smallest thing to deliver value to gauge customer interest and product-market fit. In government it can be used to replace time-consuming requirements-gathering for concrete and experiential solutions that start delivering value to users right away. With a “government MVP” or, say, a pilot, this small project accomplishes four important things:

  1. Begins to deliver value to a small subset of users right away, often in months and not years.
  2. Empirically reveals technical constraints and a way forward to address them at scale.
  3. Starts to surface outdated business processes as well as policy and compliance constraints.
  4. Begins to teach government workers how to understand and own their own digital service strategy.

Take the case of MilMove, a recent project in which the Defense Digital Service and the U.S. Transportation Command (USTRANSCOM) modernized an outdated system that coordinates how service members and their families move themselves, their families and their belongings. 

The project adopted an MVP-first approach by first choosing a smaller subset of its user base: military members who do their own moves — using their car, or renting a small truck — leaving untouched, for now, people who rely on the government to coordinate with private moving companies to do the move for them. 

By purposefully addressing the smaller user group, USTRANSCOM was able to mitigate the amount of technical integration needed at this scale and learn modern software development or “product strategy” thinking. For some military service members this meant they could use a modern, user-friendly IT system that allowed them to organize a move using their smartphone.

What began as an MVP pilot, using an initial subset of data, has since grown. In just over a year, this project has become a road map using empirical user and technical data, and has regularly increased the number of military personnel benefitting from the service.

Here’s the takeaway from the MVP methodology. Instead of traditional requirements-gathering processes that can take months or even years to produce results, which, in turn, may end up as nothing more than thousands of pages of spreadsheets and documents, a government agency could try this:

  1. Run a three day “MVP workshop” to understand the space and choose a “first user group slice,” and let your design and development team interview stakeholders to draft a plan.
  2. Allow researchers and designers four to six weeks to interview users to understand pain points and find opportunities to truly meet their needs. Then start designing an actual user interface.
  3. Have software developers start after researchers, to build upon that research and begin to test how they would deploy software in a modern way and integrate with or replace legacy software.
  4. Focus on executing on the MVP. With agile development methods, work is organized by short week sprints so work can adapt to learning along the way. 

What’s the timeframe for producing results using an MVP-driven strategy? Often within six months to a year, MVP developers will be delivering value directly to a set of users. Perhaps more importantly, not only will an agency have a plan for scaling, they will be executing on it daily.

Jesse Taggert

Jesse Taggert is the director of Research and Design at Truss.works, a software engineering firm. Before that she led design at California’s Child Welfare Digital Services and design and product strategy at 18F Consulting.