I’ve spent the last few weeks working closely with an amazing team in Sacramento across numerous state government offices to do something new.
What was going to be a business-as-usual procurement (a long, thousand-plus page contract for a complete solution, driven by requirements and a likely waterfall delivery) of a new Child Welfare System will now be a series of procurements for long-term services, not solutions, driven by understanding and meeting user needs, delivered iteratively.
The result will not only be a better system that helps the state of California support and care for vulnerable children, but also a fundamental transformation in California’s approach to understanding and controlling the technology required for services that work for the people who need them.
From Monolithic to Agile
Earlier this year, we were invited by Will Lightbourne, Director of California’s Department of Social Services, to provide feedback on a draft request for proposal (RFP) for a new Child Welfare System. Unfortunately, this RFP followed the traditional monolithic model, and we were concerned that its chances of success were very low. We’ve known for a while that large, monolithic contracts for government services don’t work. They’re frequently late, over-budget, and don’t deliver what their users need. There are lots of reasons for this, enough for an unhelpful and unproductive exercise of finger-pointing in any number of directions (vendors, customers, regulations). We recommended a dramatically different approach.
Fast forward to a meeting this fall with Michael Wilkening, the Undersecretary for Health and Human Services, and Marybel Batjer, the Secretary for Government Operations. Invited to present general findings about how the state of California procures technology (based on the detailed work for the Department of Social Services) and how it might work better, I pulled in Jennifer Pahlka, Founder and Executive Director of Code for America (and my boss). Worried that this procurement entailed relatively high risks and that the consequences to the children of California could be dire, she in turn pulled in Todd Park, former Chief Technology Officer for the United States. During their time in Washington, D.C., Jen and Todd, inspired by the United Kingdom’s success with the Government Digital Service, sowed the seeds for the U.S. Digital Service and 18F, two new parts of the federal government created to preserve and build on the practices of the Healthcare.gov recovery effort and improve how the government builds and buys technology. Jen and Todd’s feedback and advice to Batjer and Wilkening was clear: there’s a new model for government technology that works better, and if it can work in the federal government, it can work in California.
Soon after that, we got direction from Batjer, Wilkening, and Lightbourne. Even with the hard work of the Department of Social Services and Office of Systems Integration over the past two years, the business-as-usual procurement for a new Child Welfare System would be risky. Based on experience, it would probably be late, over budget, and lack functionality or optimal usability. These are the reasons why a new Child Welfare System would be the candidate for a new modular, agile approach to delivering government technology. It would be the start of thinking about developing and delivering services that are always improving, not buying products with maintenance and operation.
Too Important to Fail
But why would a Child Welfare System in particular be a good choice for this new approach? Child welfare services personnel in California investigate nearly half a million reports of severe maltreatment and life-threatening neglect to children a year. Of those half a million, around 80,000 reports are confirmed annually, 30,000 children must be removed from their homes, and at any time almost 100,000 children are living in foster care for their protection or live with their parents under close county protective supervision. The Child Welfare System was the perfect choice for a new approach because it’s too important to fail.
Meanwhile, at Park’s introduction, I started working with two groups of people in Washington, D.C. Rafael Lopez, the Commissioner of the Administration for Children, Youth and Families (ACYF) and his team were briefed and immediately got on board. ACYF provides matching federal funding for child welfare services under a program called SACWIS (soon to be replaced by a new program called CWIS), and Lopez and his team were eager to help states improve how they were using technology. Second, we were glad to have the support of Aaron Snow, the Director of 18F, the new division of the General Services Administration that provides development and consulting services to federal agencies and federal-funded state programs. These two supporters, among many others, would prove to be incredibly important to starting the Child Welfare System project down a better path.
Agile Government Technology
This fall, I started working with a small, cross-agency team at the California state capital in Sacramento to see how practical and realistic it would be to move from a traditional, large, and typically risky RFP for an entire system to a set of smaller RFPs, explicitly requiring user-centered, agile, iterative development. With this would come a fundamental change: from procuring or buying a set solution to building California’s ability to understand, direct, and control the critical technology needed for government to do its job well.
Our multi-disciplinary, cross-agency group included staff from the Department of Social Services, the Office of Systems Integration, the Department of Technology, and the Department of General Services, all crammed into the Leonard Carter Conference Room at the GovOps (Government Operations) Agency, all assigned to work the problem.
The entire group was focused on getting to yes. We asked ourselves: Was there anything in the regulatory environment preventing agile or iterative development and delivery of government technology? What do California’s IT policies look like, and how flexible are those policies? What are the risks that are trying to be controlled or prevented? What was the budget of the original RFP, and how might we deliver better services faster? Based on the original RFP, how might the overall service be broken up into smaller RFPs to reduce risk? What sort of capabilities would the state need to develop to assure successful delivery, in the short, medium, and long-term?
The Team Behind the New Direction
I’m proud to say that we got to yes. There are many reasons why we succeeded, but most importantly, each member of our team has been incredibly dedicated to the project and the outcome. The professionalism, passion, and sheer hard work of this team can’t be underestimated. Instead of coming up with reasons why we couldn’t do things differently, they all came back with ways that would work, and sometimes even ways that were better.
All of this came to a head quickly. After about a week’s worth of work, we reported back with a go/no-go recommendation to our executive team. After that, it was a tour of the Department of Finance, the Governor’s Office, the Legislature, and most important of all: the Child Welfare Director’s Association (CWDA), a key partner whose members represent end users. The CWDA responded admirably and with an open mind to the unbearably short timeline we presented them.
Lastly, we needed to talk to the project team. Around 60 people across multiple offices, agencies, and departments are involved in procuring the new Child Welfare System. On the Thursday before Thanksgiving, after the new approach had been announced to the project team the afternoon before, I took part in the first of regular and frequent open houses for the team to ask questions and talk.
I left that meeting with even greater respect for the people who work in government than I had going in. At Code for America, we understand that people who work in government are motivated by the mission of serving the public. In social services, I’ve had the privilege of seeing that motivation applied to the welfare of our country’s children. Everyone in the room on Thursday morning was excited about being part of a new process that would deliver the tools caseworkers need to do a great job of safeguarding and supporting children. Everyone in the project team asked smart questions. The atmosphere in the room was optimistic (albeit nervous) about the new direction. And the sense of relief, and later, excitement, was palpable.
For the public servants in the process we’ve started, this is a time of great uncertainty. A new approach means that the team could be asked to learn new skills and take on new or different roles and responsibilities. And yet the people in the room that morning weren’t just optimistic. They were visibly moved by the chance to do something different. Steve, who runs operations for the existing mainframe-based system, stood up and told his colleagues that he’d been working in government for 37 years, and that from his point of view, what he’d just heard was a once-in-a-lifetime opportunity to be a part of something big, something amazing. For him, the support and encouragement for doing something new, different, and better, from leadership at the state and federal levels to the team on the ground, was off-the-charts.
What’s Happening Next?
We’re producing and publishing the first RFPs to come out next month and preparing for a vendor forum on December 4th. California has shown that it can turn on a dime, decide that it’s time to change the way it buys and treats technology, and start massively empowering the team responsible for delivering.
If you’re an agile software development company that wants to work on something that matters, now’s the time to pay attention to California.
Come to the vendor forum
When: Friday December 4, 2015, 1pm to 3pm, doors at 12pm
Where: Department for Social Services, 744 P Street Auditorium, Sacramento, California 95814
The Department for Social Services, the Office for Systems Integration and the Department of Technology will be holding a vendor forum on Friday, December 4, 2015 from 1pm to 3pm, with doors opening at 12pm. If you’re an agile, user-centered software development vendor interested in hearing about how California will be procuring a new Child Welfare System, then please do come along.
Please RSVP to firstname.lastname@example.org or phone: (916) 654-0616 by 5 p.m. on Wednesday, December 2, 2015. You’ll need to check in at the Department for Social Services lobby with photo ID. Please don’t send more than 3 participants.
Take part remotely
If you can’t come to Sacramento, you can:
- View the presentation using Lync Meeting; here’s how you can install Lync Meeting
- Take part by phone by calling (877)336-1828, using access code 844631#.
With thanks to:
Marybel Batjer, Secretary
Stuart Drown, Deputy Secretary
Carlos Ramos, Director, Department of Technology
Alex Chin, Project Management at Department of Technology
Rick Goldberg, Attorney at Department for General Services
Mike Wilkening, Undersecretary
Will Lightbourne, Director of Social Services
John Boule, Director at Office of Systems Integration
Pat Leary, Chief Deputy Director at Social Services
Pete Cervinka, Program Deputy Director at Social Services
Kevin Gaines, Program Chief at Social Services
Amy Tong, Chief Deputy Director at Office of Systems Integration
At the Federal Government:
Todd Park, former U.S. Chief Technology Officer
Rafael Lopez, Commissioner at Administration for Children, Youth and Families
Jenny Wood, Chief Deputy, Administration for Children, Youth and Families
Aaron Snow, Director, 18F
Lynn Overman, Senior Policy Advisor, White House Office for Science and Technology Policy
At Code for America:
Jen Pahlka, Executive Director
Mike Migurski, Chief Technology Officer
Rebecca Coelius, former Director of Health