In December 2020, I became the director of the U.S. Department of Treasury’s nearly $2 billion RESTORE Act1 programs to revitalize the economy and ecosystems of the U.S. Gulf Coast. It was a nonpolitical appointee position and a fascinating opportunity to oversee two grant programs to address the damage done by the Deepwater Horizon oil spill. Funded by civil and administrative penalties imposed because of the spill, our work spanned economic development and workforce development, infrastructure, ecosystem restoration, scientific research, and other activities in the 47 states and counties named in the statute as the most affected by the oil spill.
The very first thing I did was ask our grantees for feedback on working with us — and I got an earful. Their most common complaint? “It just takes too darn long.”
When I presented the customer feedback to my staff, to my surprise, they heartily agreed: grant processing times were too long; our grants management system was clunky and not user friendly; our processes and requirements were confusing and not transparent to the grantees we served.
What began with one candid conversation about why we weren’t as efficient as the grantees wanted us to be — or as we wanted us to be — morphed into a two-year employee-led war on inefficiency.
In this effort, we identified and addressed seven interrelated challenges, as detailed in this paper:
- We did not adequately understand the root causes of delays.
- We lacked data to validate or refute customer feedback.
- Staff felt unempowered to address the root causes of delays.
- The typical hiring process was ill-suited to finding candidates who could turbocharge our efficiency drive.
- Employee performance plans were ill-suited to recognizing or rewarding employees who were driving efficiency gains — or holding inefficient employees accountable.
- Our grants management system was ill-suited to our needs.
- Legacy contracts put our budget on an unsustainable trajectory, even as we needed to find money for a new grants management system.
Over the course of two years, with the support of our leadership, we overhauled every aspect of our operations. Our new employee organizational structure, employee performance plans, and case-study approach to hiring meant that we had the right people in the right positions with the right customer service skills, growth mindset, and “can-do” attitude. We rebid our contracts to reduce costs, improve vendor accountability, and reduce vendor capture. We jettisoned our old grants-management IT infrastructure and replaced it with a new system that turbocharged our efficiency and reduced grantee burden. Most importantly, we evolved from a top-down culture focused on compliance to a bottom-up culture focused on accountability, customer service, and mission fulfillment and results.
There was no silver bullet, no single approach that ensured success. Because all of these challenges were interrelated across culture, human resources, information technology, and procurement, we had to address them together. It required enormous amounts of time and energy to overcome the myriad process and culture roadblocks in the way.
It shouldn’t have been this hard. But our results show that dramatic improvements are possible even within current systems when the conditions are just right, and it underlined the importance of empowering civil servants to effect change.
Challenge #1: We poorly understood the root cause of delays in serving grantees
We already knew our grantees were frustrated. But in the past, most feedback had come in the form of angry missives about one particular grant application that had been held up. From what I understood about the process before I arrived, feedback had generally been more tactical than strategic — What do we need to do to get this particular application over the finish line? — and it had not been a two-way dialogue about the root causes of the delays.
Solution: Engage in detailed one-on-one listening sessions with grantees
We changed this dynamic by conducting one-on-one listening sessions focused on collaboratively identifying and addressing the root causes of these delays. After letting a grantee share concerns and experiences for the first five minutes, instead of defending ourselves, we asked the grantees what they thought were the root causes of the delays. Their answers were illuminating.
First, some rural grantees highlighted the challenges they faced in completing our application package correctly and in a timely manner, citing their own high staff turnover, limited resources to manage the program, and limited knowledge of federal grant requirements. Many praised our free class on the government-wide federal grants requirements codified in 2 CFR 200, but noted that it was offered infrequently. As a result, many new staff at their organizations had to wait six months or more to attend a training. Nor was the training class offered to subrecipients who, in many cases, were responsible for the day-to-day work, and it did not cover Treasury RESTORE-specific grant requirements.
Second, many grantees expressed confusion about Treasury’s requirements, review process, and timelines, resulting in lengthy exchanges with Treasury in order to submit a complete, compliant package.
Third, nearly all highlighted frustration with our grants management IT system, which lacked data validations, required numerous time-consuming manual workarounds, and was generally not user friendly.
Why we were able to do this — and why it doesn’t happen more often
We were fortunate that our grant program, by statute, had only 47 eligible state and local government recipients, which made holding one-on-one listening sessions with each grantee feasible. While other agencies with a larger pool of customers could accomplish much the same thing by holding listening sessions with a cross-section of customers or by holding focus groups with customers, I suspect many do not over concerns of violating the Paperwork Reduction Act, which requires a multimonth process involving Federal Register notices and formal approval by the Office of Management and Budget. The PRA requires approval for even voluntary collections, which OMB interprets to include “customer service and behavioral surveys” and “focus groups with a set of the same questions.” Since it was a verbal conversation, and we did not ask a “set of the same questions” but simply had open dialogues about their views of the root causes of delays, the PRA did not apply.
Challenge #2: We lacked data to validate or refute customer feedback
While everyone agreed that it “took too long,” no one at Treasury actually knew exactly how long.
Solution: Get the data
A team member downloaded the data from our current system, which showed the time stamp for each action taken for each application. She then embarked on a difficult and laborious process to map these time stamps to key steps in our workflow, calculate the time elapsed between key actions, and slice and dice the data by application type.
She then conducted a comprehensive review of our processing times to identify potential bottlenecks. She found that the number of grant actions we processed increased 38-fold between 2015 and 2020, while staffing had only doubled, and that the median grant processing time was 233 days. Echoing grantees’ concerns about lengthy exchanges, the analysis found that we devoted two-thirds of processing time to working with applicants to produce a complete, compliant application and that many applications were returned three or more times along the way. Processing times also varied significantly by grants specialist, suggesting potential employee performance issues.
Challenge #3: Staff felt unempowered to address the root causes of delays
At an all-hands meeting, we discussed the results of the grantee listening sessions and data analysis.
To my surprise, staff agreed that processing times were too long and that many of the key problems were those that grantees and our data analysis highlighted. The staff, however, felt powerless to address the causes. They agreed that our grants training was subpar, but noted that we couldn’t change it because of the contract with the vendor. They also said they felt constrained by our procedures, which they described as entailing excessive levels of management approvals for routine actions such as no-cost grant deadline extensions, and an excessive level of detail in the lengthy grant approval memoranda they were required to write. They also said they felt constrained by our legacy grants management system, which did not include any conditional fields or data validations and required numerous manual workarounds.
Solution: Flip the script, secure small wins to build trust
We flipped the script and asked, “Pretend we could remove these constraints. Then what would it take to reduce our processing times from 233 to 120 days?”
That question led to biweekly strategy and brainstorming sessions devoted to dramatically reducing processing times, which generated numerous ideas that we subsequently adopted. Management approved replacing the lengthy funding memoranda with a streamlined, staff-developed version. At the suggestion of staff, I gained authority to sign off on many actions that previously required the approval of a Deputy Assistant Secretary.
Staff also developed a consolidated checklist to identify all application deficiencies, with the intention of eliminating the constant back-and-forth that grantees found so frustrating. We committed to providing this list within 30 days of application submission, and we committed to issuing a grant award within 120 days of the submission of a complete, compliant application.
While no panacea, these actions built trust between grantees and employees that real change was possible.
Challenge #4: Our hiring process was ill-suited to finding candidates who could turbocharge our efficiency drive
Culture only permanently shifts when you empower, attract, retain, and reward employees who are change-makers. Within a few months, staff felt empowered to recommend and implement solutions, and most of them were enthusiastic about our efficiency initiative. When less enthusiastic staff left, we needed to replace them with new employees with outstanding grants management and customer service skills — not to mention a growth mindset and a can-do attitude!
We had been using the traditional federal government hiring model:
- HR posted a vacancy announcement on USAJOBS, asking candidates to rate themselves against questions that HR derived from the position description, which in fact were not always germane.
- HR then allowed only the hiring manager to interview candidates who gave themselves the highest marks, without an external validation of whether they were actually telling the truth.2
We found through experience that this process typically yielded terrible results, with many interviewees lacking basic grants knowledge despite having rated themselves as experts in every aspect of grants, and that some terrific candidates were eliminated even before the interviews.
Solution: Switch to skills-based hiring
We decided to try shifting from the traditional recruiting model to a skills-based model, which the Office of Personnel Management technically allowed but that was rarely used. In fact, my HR contact claimed it had never been done in Treasury’s Departmental Offices, and we had to go through three levels of HR approval just to start the process.
We proposed a skills-based hiring model that involved a practical online exercise requiring a candidate to review a fake grant application riddled with errors, identify the errors, and outline the technical assistance they would provide the applicant. HR insisted it would be too burdensome for candidates to complete this exercise up front but said we could post a two-part, hurdled announcement in which those candidates who self-assessed their skills as at least qualified would be directed to complete the exercise. We then needed to create a detailed, objective scoring rubric, appoint a panel of three subject-matter experts, and have each subject-matter expert rate each answer for each applicant. Finally, we would be able to interview those candidates above a particular cutoff score.
The results were surprising — and completely validated our approach. Many candidates who had rated themselves as grants experts and said they had 20 years of federal grants experience could not pass our test. In contrast, a more junior external candidate from a state government agency aced our test. We hired her and, within weeks, she was adding significant value and demonstrating the customer service skills and can-do attitude we had been seeking.
Why we were able to do this, and why it doesn’t happen more often
We were fortunate that HR was ultimately willing to try a completely new hiring model outside its comfort zone. Based on my discussions with other former federal officials, very few HR teams have implemented skills-based hiring, and most federal managers are not even aware this is an option. Other barriers to widespread implementation include the fact that USAJOBS, the federal government’s hiring portal, is not well-suited to handle case studies, and many small teams do not have the bandwidth, capacity, or expertise to deploy three team members to spend hours reading and scoring responses according to a detailed scoring rubric.3
Challenge #5: Employee performance plans were ill-suited to recognizing and rewarding employees driving efficiency gains — or holding inefficient employees accountable
The performance plans I inherited contained vague language about “responding to grantee emails” and “processing applications,” but no concrete metrics. Therefore, under the current plan, if employees could show they “processed applications,” even if it took an average of 300 days, and answered emails, even if it took five days to send an unhelpful response, my management team and I had to give the employee at least a “satisfactory” rating.
Solution: Create new performance plans aligned with our focus on process improvement and customer service
The new plans at each level in the grants review process required the individual responsible for the review to work against quantitative performance targets for speed and accuracy; if each individual at each level of review met the “satisfactory” performance target, the office as a whole would meet its goal of processing grants within 120 days. And to demonstrate that I was holding myself accountable, too, I limited my time to review and approve grants to just three business days.
Most staff members were cautiously optimistic about their ability to meet these more stringent standards. Those who felt the new standards were unreasonable soon found other jobs.
Challenge #6: Our grants management system was ill-suited to our needs
Since the vendor of our legacy grants management system was either unable or unwilling to fix the issues with the system, we mutually decided not to exercise the next option year. That meant we had just seven months to procure, develop, and launch a brand-new grants management system.
Solution: Procure, develop, and launch a custom solution on a Salesforce platform in just 7 months
To meet the aggressive timeline for procuring, developing, and launching a new system in seven months, we assembled a team of Treasury system users to work on parallel tracks over the next six weeks. This consisted of a mix of my team members and members of a Treasury compliance team, with the as-needed involvement of accounting and other staff who occasionally used the system. A sympathetic senior executive also offered the services of one of her staff members who had extensive experience documenting business requirements for IT projects.
One workstream was to develop our idealized new workflow to maximize efficiencies and automate manual processes. This included converting all previously nonmachine-readable PDF forms into electronic application forms with conditional mandatory fields and the necessary data validations to reduce the rate of application errors. This team met twice a week for 90 minutes, and most team members put in an additional 5–15 hours a week creating field-by-field requirements for each form. The team also took the opportunity to recommend the deletion of any redundant or unnecessary application questions; if approved by our attorney, we removed it from the form given to the vendor even though we would need official Paperwork Reduction Act approval from OMB to actually do so.
The second workstream was to investigate other possible solutions that would fit both our business needs and our tight procurement timeline. We scheduled a demonstration of another shared service solution and determined it could not meet our functional needs or timeline. We also researched developing a Salesforce-based system that would better align us with other Treasury grant programs and provide more flexibility to update the system as new statutory requirements arose or our needs evolved. Because of the tight time constraint, we analyzed the fit between our needs and Salesforce’s capability while we identified all contract vehicles at Treasury through which we could access a contractor to customize Salesforce for us.
Within six weeks, we completed our vendor research and recommended the selection of a vendor who had successfully implemented Salesforce for other Treasury grant programs and with whom we could easily contract using an existing Blanket Purchase Agreement. Within a few days, Procurement executed the task order.
Four weeks after that, we hosted a project kickoff with the vendor, during which we provided incredibly detailed documentation of business requirements for a brand-new, user-friendly system that would address all of the pain points grantees and Treasury staff had identified. At the same time, we submitted a request to OMB for emergency approval of our new, electronic, streamlined application forms under the PRA, which requires agencies to conduct an extensive public comment process and secure formal OMB approval to change even one field on one form. We were running the risk that we would have to start all over again — and delay the launch — if we did not meet the criteria for emergency approval. And even if we did receive emergency approval on the legal grounds that “the use of normal clearance procedures is reasonably likely to prevent or disrupt the collection of information,” we ran the risk that OMB would subsequently reject our new forms during the regular PRA clearance process, since emergency clearance is only granted for a period of 180 days. If that occurred, we would have to take the system offline and spend a significant amount of time and money to revert to the old forms. Fortunately, we received an emergency approval, and assurances from OMB that normal approval would likely follow as well, which allowed us to proceed with confidence.
That left just four and a half months to develop, test, and launch a minimum viable product by October 1, 2023, when our legacy system would become unavailable.
Over the summer, our team and the vendor worked hand in glove during the agile development process, dedicating the time needed. The vendor adopted three overlapping two-week sprints: requirements, development, and testing. Staff met with the vendor three times per week, for 90 minutes each time, to review user stories that the vendor drafted based on the highly detailed business requirements we provided. At these meetings, we live-edited the user stories, and then I formally approved each user story as system owner. We also wrote our own test scripts so that our team could perform role-based user-acceptance testing of each module using real grant applications with real data.
And when the legacy vendor had the audacity to demand compensation to return our own data to us for migration, our team stepped up and volunteered to work nights and weekends to manually download and clean the data from the legacy system. They asked for no extra pay, only compensatory time off. As a result of these efforts, our data was successfully offloaded in the two months prior to launch — and seamlessly uploaded by the vendor into the new system.
Against all odds, we successfully launched our minimum viable product on October 1, 2023, on time and on budget — and just four and a half months after development began.
The new system replaced all of the PDF forms with machine-readable electronic forms that Treasury could easily query. The system was able to guide applicants seamlessly through the process so that applicants were only directed to complete the necessary forms and fields based on their prior responses. The new system did not permit the submission of incomplete applications or budgets that did not add up; if a required form or field was left blank or there was a mathematical error, it alerted the applicant. The new system also removed all the duplicative steps that the prior system required because the system had been tailored to another federal agency with more levels of review and approval. In other words, we built a modern, user-centered software system.
Because it was built on a Salesforce platform and Treasury had Salesforce licenses, we avoided the vendor capture that previously had bedeviled us. We owned both our data and the customization code, and if we ever parted ways with the vendor, we could easily hire another to handle updates.
The only core functionality that was in the prior system but not the minimum viable product was the grant reporting module, which we launched in December 2023. By April 2024, we had completed Phase II development, which included a module to receive and review statutorily required Multiyear Implementation Plans as well as the first agency module to receive and process Build America, Buy America waivers, which we demonstrated to the OMB Working Group on Build America, Buy America. Neither of these critically important modules could have been created in the legacy system.
The system not only turbocharged our efficiency, but also saved money over the long run. The cost of developing the minimum viable product was only 30 percent more than the legacy system provider was going to charge us to add a single dropdown filter. The new system, coupled with the contractual savings and HR changes outlined above, allowed us to reduce processing times far beyond our initial target, from 233 to 75 days, while hiring only one additional staff member.
Why we were able to do this, and why it doesn’t happen more often
We benefited from a great team of dedicated civil servants and a great vendor team.
Perhaps even more importantly for our aggressive timeline, we were able to leverage an existing Blanket Purchase Agreement that another office in Treasury had with a qualified vendor to customize Salesforce for grant programs. This allowed us to sign a task order just six weeks after deciding to move off the legacy system, and to start development just four weeks after that.
Most crucially, it allowed us to circumvent the usual procurement processes, which from my experience at other agencies can range from six months to execute a contract with an 8(a) small and disadvantaged business, to 18 months for a full and open competition.
Most agencies cannot replicate our experience, simply because they do not have the flexible contract vehicles in place with the right vendor for the right type of product — and because the current “normal” procurement process is anything but nimble and fast.4
Challenge #7: Legacy contracts put our budget on an unsustainable trajectory, even as we needed to find money for a new grants-management system
One of the perks of being RESTORE program director was that I had two-year money. In fiscal year 2021, for example, I had access to two pots of funding: fiscal year 2020–2021 funds and fiscal year 2021–2022 funds. However, when I came on board, I could see that we were dipping into that second pot earlier each year. After consulting with our budget director, I determined that, given the current annual rate of growth in payroll and contractual costs, we were on a path to run out of operating funds before the end of FY27.
This was clearly unsustainable, and we would need to tighten our belts. To make matters worse, we didn’t just need to reduce our costs to prevent a shortfall in FY27. We needed to actually find money in our budget to replace the legacy grants-management system that was the source of so many of our inefficiencies.
Adequate staffing was essential, so cutting staff was not an option. That meant we would need to pare back our two existing non-IT contracts.
Solution: Revamp and rebid grants training to reduce cost and increase utility
Revamping our grants training was a priority for both staff and grantees. We worked with our procurement team to cancel the remaining years on our training contract and to rebid the contract. We then rewrote the RFPs to address all of the deficiencies of the prior training: to switch from in-person to virtual trainings; increase per-class enrollment caps to 60 participants; allow contractors and subrecipients to enroll; include RESTORE-specific grants requirements; and require all sessions to be recorded. We awarded a new contract and co-designed the training with the vendor to ensure that it emphasized RESTORE-specific requirements. To save money, instead of exercising the next option year, we ended the contract and replaced live training with the recorded training, augmented with optional live tutoring sessions with our staff.
Solution: Bring some contracted functions in-house
Staff conducted a thorough review of the utility of our other existing contract. This contract was for environmental and engineering services; contracted engineers conducted site visits to ensure projects were completed according to specifications and reviewed statutorily required Best Available Science narratives to ensure applicants properly cited peer-reviewed scientific articles to support their proposed approach. We also tagged along on several site visits with the contractors to see the value they added.
Our verdict?
We would be much better served by taking this function in-house for significant cost savings. We could hire one in-house engineer for about one-eighth the cost of the annual contract. We already had several Ph.D. scientists on staff who, while working as grants specialists, were more than qualified to review Best Available Science narratives. And conducting our own scientific reviews and site visits would allow us to provide better technical assistance to grantees.
Why we were able to do this, and why it doesn’t happen more often
Most agencies are constrained by one-year funding and, in some cases, a lack of diversified skills in their talent pool. That wasn’t true in our case.
We were able to achieve significant cost savings by working proactively with the procurement team to end the cycle of automatic renewals and by attracting and retaining a team of diverse academic and professional backgrounds. Even though nearly all of our employees were classified as grants specialists, they had other valuable skills — including Ph.D.s in scientific fields! — that we could leverage to replace contracted-out functions.
Two-year funding also gave us far more flexibility to spend more during the initial phase of development for our new grants-management system, and then significantly dial back spending in the following year as we realized cost savings from cutting these two contracts.
How streamlining allowed us to refocus on the mission and results
So often, we talk about government efficiency as if it is separate and apart from effectiveness. But they are actually linked; when you reduce the time spent on low-value activities, you can increase the time spent on high-value activities.
Our team spent the time we freed up from endless application reviews and manual workarounds to
- Improve customer service and technical assistance. Grants specialists held weekly calls with each grantee to troubleshoot issues. The team ramped up site visits, and we held quarterly virtual town halls with grantees.
- Create a community of practice. We launched a webinar series to highlight best practices and allow grantees to brief their peers on their RESTORE-funded ecosystem research.
- Embark on a yearlong effort to identify and implement outcome metrics to enable us to track and report the ecological and economic impacts of RESTORE-funded grants.
Maureen Klovers has 14 years of federal experience across four agencies, most recently serving as Chief Enterprise Officer and Deputy Chief Operating Officer of USDA Rural Development. She is the Co-Founder of We the Doers (wethedoers.org), a group of former civil servants dedicated to making the federal government more effective, efficient, and trustworthy.
- Officially the Resources and Ecosystems Sustainability, Tourist Opportunities, and Revived Economies of the Gulf Coast States Act. ↩︎
- For a more detailed explanation of how a typical federal hiring process works, see We the Doers’ recent report, Former Civil Servants Speak: How to Achieve Real Government Reform in a Post-DOGE World. The section “The Federal Hiring Process is Convoluted” gives a good overview. ↩︎
- The Chance to Compete Act supports more skills-based hiring, but requires smart implementation. The Niskanen paper “Beyond the checkbox: Unlocking skills-based hiring under the Chance to Compete Act” goes into detail. ↩︎
- I do wish to note, however, that the procurement rules are currently being revised as part of the Revolutionationary FAR Overhaul. ↩︎
