Replacing legacy systems

In the private sector, a legacy system might be a few years too old, a couple versions out of date, or just not serving the needs of the company anymore. Investment in a new system is often just an issue of making the case about cost savings to the right people.

Not so in government.

These legacy systems are another beast entirely. Because it can be so challenging to get systems updated or replaced, they often remain in use far longer than they should, postponing the problem for as long as possible, and making it worse in the meantime.

The biggest barriers to replacing government legacy systems are often:

• Cost. Modern technology is expensive, and agencies often would (understandably) prefer to spend their budget on what they need to keep providing services. And of course, priorities change with new administrations or appointees, so even a project funded by one may be defunded by the next.
• Bureaucracy. It takes a lot to get new technology approved for use, budgets approved to purchase it, and a team onboarded to do the work. Some of the processes for these approvals can take months, even years, especially if you’re navigating them for the first time.
• Dependencies. Systems rarely exist in a vacuum, and updating or replacing one can require changes to a whole set of systems or processes. Data is often not stored in just one system, but across multiple systems that have been customized to work together. Break one link in that chain and you risk losing the data forever. Any changes to a system (and its ecosystem) could also require new processes, training, and change management.
• Risk. While there is a wide spectrum of government services and systems that support them, it’s simply true that some of them have the power to save or ruin lives. No one wants to be responsible for a system failure that results in someone losing their home, not receiving medical care, or other serious consequences.

Despite these challenges, replacing legacy systems was a huge part of the work that 18F did before we were eliminated on March 1. In fact, we were really good at it.

We were successful because of our experience and expertise. Many of us had done similar work in the private sector and brought that expertise to government. We also had folks with deep knowledge of government regulations who had worked in agencies across all branches of government. That expertise, combined with our core principles and approach to this work, made us uniquely well-suited to tackle the barriers that usually block these types of transformations.

Keeping costs low

Negotiating the cost of a technology project can be intimidating – especially if the people doing it aren’t familiar with the tools, processes, and industry standards involved. 18F was able to help our agency partners be confident they were getting the best possible deal with the vendors they hired because we brought all our expertise to negotiating those contracts. Kind of like bringing a friend who works with cars with you when you take yours to the mechanic.

We also made sure that new systems and contracts were set up for long-term success, not just short-term goals. The most important aspect of keeping costs low was to use strategies that avoid locking in with a single vendor for services. If you replace a legacy system with a new system that is completely dependent on proprietary technology, you’re just creating the next costly legacy system. The government needs to understand every facet of the product strategy so they can avoid the costly pitfalls of being held hostage by their vendors — this can only be done by having a long-term internal staffing strategy.

There are no silver bullets that make this an easy fix, but 18F developed methods for de-risking technology contracts and replacing legacy systems successfully and responsibly.

Untangling dependencies

In each agency, there isn’t a simple list of what systems and software does what — it’s more like an entire ecosystem. Things pop up to meet specific needs at specific times, and then evolve in place depending on factors like budgets, priorities, and institutional knowledge. There are redundancies, yes, but also keystones that if removed could bring the entire ecosystem tumbling down.

When starting a new project, 18F took the time to understand the entire system — the technology and the people — and all the other parts of government that it impacts. This is where design thinking is vital.

We didn’t just ask: What does this thing do? We’d ask a whole slew of questions, talking to internal stakeholders and end users, and back it up with research.

• Goals and jobs to be done: What does this system do? What are the people using it trying to achieve?
• Processes and understanding why something is in place: Why does it work the way that it does? What were the conditions when the system was put into place? What, if anything, is different about those conditions today?
• Reliances: What data or other systems feed into it? Or rely on it?
• Pain points and experiences: What does the whole experience feel like for the end users? What parts of it are hard and what parts are easy? Why?
• Backstage: For the people who manage this system, what information do they have access to? What pieces of it can they change? What do they like about it? What makes their job more difficult?
• Legal, security, and privacy: What policies or laws guide the functions of this system? How is the data protected, stored, and secured? What are non-negotiables?

Once we mapped the whole ecosystem, we’d have a full understanding of what the needs are, where they are and are not being met, and perhaps most importantly, what the consequences would be if we changed any piece of the process.

Navigating bureaucracy

Yes, red tape is always part of the process in government. But if you want to cut red tape, you need to engage with the current process.

Agencies create systems to help them track and complete their work. The system will set up the information so it can be processed and acted on based on the policies, rules, and laws of that process. No matter how good the technology, a system cannot be better than the underlying rules and processes that govern the system.

With every project 18F completed, we learned more about how to tailor an agency’s approach to, for instance, getting required security approvals known as Authority to Operate (ATO). 18F worked hand in hand with agency experts to understand the “why” behind the current process. Once there was a base level of understanding, we worked with the agency to make the system more efficient while still following the regulations that protect citizen data, privacy, and other requirements.

When replacing large systems, improving the process it has to follow can yield greater gains than just technical improvements on their own. But process improvements require thoughtfulness, trust, and collaboration — a “destroy and rebuild” approach simply won’t achieve the desired goal. It’s also important to know what pieces of red tape to keep. Some hurdles in the process are there for important, valid reasons. They do things like protect the privacy of Americans’ data, allow Americans with disabilities to engage with their government, and reduce the amount of paperwork the public has to do to access their government benefits. Cutting out pieces of a process simply to make it shorter won’t protect the American public.

Managing risk

“Move fast and break things” might have its place. But in government, breaking things can ruin lives. There are some things that can’t be fixed once they’re broken. And in many cases, the government is the only provider of that service. It’s vital to understand where the risks are and address them proactively.

Part of managing risk is not trying to create and launch an entire finished product all at once, only to find out the problems after all that work is done. Instead, 18F took an agile, iterative approach, starting small and getting user feedback to make sure we get each piece right before we add on more. This type of approach means you begin adding value quickly and responsibly. The alternative is a big, flashy, headline-grabbing release that takes years to produce, incurs cost overruns, and that doesn’t end up meeting user needs.

Yes, it takes time and effort to work responsibly and effectively. But government systems affect people’s lives. The American people deserve government technologists who are committed to responsible, proven practices that do no harm. Breaking technology without caring about its impact can break people.